SYSTEMATIC CONTROL AND PROCESSING TO MONITOR AND MANAGE CONTESTANT ENTRY DISPERSION OF SINGLE AND MULTIPLE SESSION INTERNET CONTESTS OVER THE ESTIMATIONS AND PREDICTIONS OF FUTURE EVENTS

Abstract

An information and synchronous communications computer system and process displays (i) contingent and non-contingent promotional offers or contest rewards, and (ii) data relating to future events with uncertain outcomes, where the realization of contingent promotional offers and rewards is based upon a contestant's successful predictive balloting (through interactive computer interface of the system) over future events with uncertain outcomes. Balloting involves the submission of one or more predicted choices or values relating to single or multiple future events with uncertain outcomes. Through real-time processes, the system creates, validates and modifies balloting interface and related processes to control dispersion of contestant entry selections in order to maintain a forecasted aggregate amount of awards within a statistically predetermined range. The system and process makes extensive use of sample population data, survey data, social graph data, and contestant entry data for management of contestant pool and contest space within which contestants ballot.

Description

FIELD OF THE DISCLOSED EMBODIMENTS

The disclosed embodiment relates to information management systems, synchronous communications systems, and related specialized computer systems for creating and dynamically modifying internet contest environments which run over one or more sessions. Specifically, the system and process for the system's creation and serial modification of the contest space make extensive use of real time statistical measurements for the purpose of creating the contestant pool, creating contest ballot selections, and monitoring contestant entries and triggering real-time modifications to the contestant interface and the underlying system when realized contestant entries deviate materially from sample population or survey data.

BACKGROUND OF THE DISCLOSED EMBODIMENT

Product oriented businesses use the internet in various forms including: (i) the introduction of product launches, (ii) the delivery of information on existing and prospective products lines, (iii) the delivery of customer support, and (iv) promotions of special pricing or enhanced offerings. Promotions and offers may include discounted goods, or prizes or services to selected prospective customers. To drive traffic to a web site, businesses create discount, giveaway and contest offerings. The larger and more exciting the promotion, the greater the user traffic.

Service oriented companies use the internet for product promotions and support similar to product oriented companies. Service oriented companies also use promotions for brand awareness using contests and giveaways for visitors who identify themselves and provide personal data and contact information on a particular site.

Further, service oriented companies may engage in contests of chance or skill on their internet platforms. Where permitted by law, such contests of chance or skill may or may not include cash consideration for entry by the contestant.

Through the use of contests, an internet site owner can increase exposure to a site's commercial content. In the context of a news related web site and in particular relating to financial news, the very nature of a web site user relationship involves minute-to-minute, hourly, or daily visits, and repeated access to a multi-session or continuous event contest further enhances loyalty to a given website.

Yet there still remains a need in the art for processes and controls of a single or multiple session internet contest with the structural and functional architecture for: (i) ensuring the aggregate award amount of the promotion or prize remains substantially within a pre-determined target amount, (ii) enabling relatively larger rewards in absolute size, and (iii) tailoring the contest space to individual contestants based on the contestant's entries. The disclosed embodiments below accomplish these objectives through systems and processes which measure and control the dispersion of contestant selections throughout a contest.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic representation of a system specialized window display 20 on a user's internet access device (computer, tablet, mobile telephone, or other handheld display device), with windows indicating one of more product promotions or contests 22, a detailed specification of a future event and the event space 26, and a window for inputting the user predictions or selections over the uncertain event 28;

FIG. 2 is another embodiment of the schematic representation in FIG. 1 where the system specialized window display is presented on a handheld mobile device 20 where the product promotion 22 and detailed specification 26 are presented, and where the window or area for inputting the contestant selection 28 is an application window or texting area;

FIG. 3A is a schematic representation of the user predictive input window 28 illustrating a contest space with 3 discrete and mutually exclusive choices, each selected as a radio button in the window;

FIG. 3B is another embodiment of FIG. 3A indicating a contest space with many mutually exclusive choices, each selected as a radio button;

FIG. 4 is a schematic representation of the user predictive input window 28 illustrating an embodiment in which a continuous variable event space is articulated into a contest space with (n) point estimator inputs requiring the user to input a numeric value;

FIG. 5 is a schematic representation of the user predictive input window 28 illustrating a rank ordering contest space in which the category items are rank ordered in a mutually exclusive manner;

FIG. 6A and FIG. 6B are schematic diagrams illustrating a specialized window display 20 on a handheld mobile device and the connectivity with the contest processing system 11 and the related central storage device 16 in which contestant balloting is time dependent and the system stores, as contestant selections, the time and any related value attributes at the selected time as indicated by the preferred embodiment computer record 29;

FIG. 7A is a graphical diagram illustrating a preferred embodiment where the system converts a continuous variable into ranges (A, B, C, and D) occupying equal probability weighting.

FIG. 7B is another embodiment of FIG. 7A and it illustrates a contest space of four mutually exclusive range choices 28 corresponding to ranges A through D in FIG. 7A;

FIG. 8A is a picture diagram illustrating system translation of a discrete variable event space into probability weighted outcomes.

FIG. 8B is another embodiment of FIG. 8A and it illustrates a contest space of four mutually exclusive range choices 28 corresponding to the probability ranges in FIG. 8A.

FIG. 9 is a binomial tree diagram illustrating a 3 round contest where “S” indicates success and “L” indicates loss for a given round;

FIG. 10 is another embodiment of FIG. 9 where the 3 round contest has awards of $10, $3, and $1 for each of 3, 2, and 1 success, where the probability of success (or S) is 25% in each round, and the expected value of the contest per contestant is $1.00 indicated at Start;

FIG. 11 is a graphical illustration of the system's selection limitation process, where the majority selection (Range D) is incrementally increased on a pro-form a basis, and the remaining choices (A, B, and C) are decreased in a manner consistent with their sample population proportions;

FIG. 12 is a graphical illustration and an extension of FIG. 11 where it illustrates the sequential pro-form a increase in the sample population majority selection, Range D, and where the system calculates corresponding concentration statistical confidence levels using a non-parametric statistical test;

FIG. 13 is a binomial tree diagram contrasted with FIG. 10 in that a system selection limitation is employed at a level of 90%;

FIG. 14 is another embodiment of FIG. 14 and it illustrates an 80% selection limitation;

FIG. 15 is a graphical illustration of contestant survival, where the maximum number of contestants is assumed to populate the successful choice (Range D) over a 3 round contest;

FIG. 16 is a schematic representation of the user predictive input window 28 illustrating a contest space with 4 discrete and mutually exclusive choices, each selected as a radio button in the window, and FIG. 16 is contrasted with FIG. 3 in that a system process has bisected the “UP” selection from FIG. 3 into “UP1” and “UP2”;

FIG. 17A is an graphical illustration and extension of FIG. 3 and where the delineated range selections under the probability space map to the input area of FIG. 3;

FIG. 17B is graphical illustration and an extension of FIG. 17A wherein the change from FIG. 17A to FIG. 17B illustrates the system modification in response to a range concentration limit also indicated in the change between FIG. 3 and FIG. 16;

FIG. 18A is graphical illustration of a time series variable with specifically identified values (y-axis) at specifically identified times (x-axis);

FIG. 18B is another embodiment of FIG. 18A where four contestants (A, B, C, and D) each ballot for times T1, T2, T3, and T4 (and the associated variable values) respectively;

FIG. 18C is another embodiment of FIG. 18B where contestant ballots are modified by the system based on a queuing process;

FIG. 19 is a graphical illustration of a contest in which the unknown future event is endogenous to the contest and contestant selections and where the contest is partially based on the concentration of investor selections and where the system modifies the game space in response to a deviation in expected versus realized dispersion of contestant selections;

FIG. 20 is a schematic diagram illustrating the system connectivity;

FIG. 21 is a time line diagram which illustrates the timing relationship in a preferred embodiment of balloting or entries to the event space;

FIG. 22A is a process diagram from contest start to the beginning of contest rounds;

FIG. 22B is a process diagram from the start of contest rounds to contest end;

FIGS. 23A, 23B, 23C, 23D, and 23E are five step diagrams which illustrate each step of an embodiment of the process of contestant pool construction;

FIG. 24A is a graphical illustration of a range selection bifurcation relating to the contest example of FIGS. 22A and 22B; and

FIG. 24B is a graphical illustration of a range selection bifurcation relating to the contest example of FIGS. 22A and 22B.

DETAILED DESCRIPTION

The disclosed embodiment processes and controls a single or multiple session internet contest for the purposes of: (i) ensuring the aggregate award amount of the promotion or prize remains substantially within a pre-determined target amount, (ii) enabling relatively larger rewards in absolute size, and (iii) tailoring the contest space to individual contestants based on the contestant's entries. The disclosed embodiment accomplishes its objective through processes which measure and control the dispersion of contestant selections throughout a contest.

The system creates, monitors, and modifies internet display interfaces and their underlying systems for internet and other computer network based contests. The system also creates optimal contestant pools utilizing sample populations, survey data, social graph data, and reality mining in the process of creating a contestant pool with a measurable selection choice dispersion; measurable dispersion varies with implementation, but examples of measurable dispersion are (i) a less than 80 percent concentration in a two choice contest, and (ii) less than a 51 percent majority in a three choice contest—in both examples an expectation of at least 20 percent concentration in a non-majority selection. Further, the system will link the processes relating to creating contestant pools and the processes relating to contest game play to effect an expectation of dispersion of contestant choices among two or more mutually exclusive contest selections.\

In the contests, contestants will be presented with promotional offers or contest rewards whose realization is contingent upon the contestant's successful prediction of the outcome of future events over one or more interne sessions. In particular, the system creates and validates the parameters of the contest on a real time basis from initial launch through final tabulations.

The term “contest selections” is used herein to denote the choices presented to an incremental contestant at a point in time. The term “contestant selection” is that contest selection indicated by contestant as her desired choice. The term “contestant entry” is a contestant selection which has been both received and processed by the system and that value which will be used by the system in rewards processing. The distinction between contestant selection and contestant entry can be characterized as raw input (contestant selection) versus system processed data item (contestant entry).

Using real time validation and monitoring, the system ensures that the expected aggregate award amount remains below a preset value without the use of pari-mutuel or related prize sharing methods. Upon the breach of system statistical measurement triggers, the system alters the game space (including the user interface) in order to restore process results consistent with preset limits, including statistical confidence levels and the original expected aggregate award amount.

In comparison with known methods, through real-time measurement and alteration of the game space, the disclosed embodiment produces a material reduction in possible aggregate award amounts. This occurs with little or no perturbation to the contest participants.

In the creation, modification and management of the contest environment, the system utilizes sample population data, survey data, social graph data, and reality mining data. Sample population data are polling data drawn from subsets of contestants, prospective contestants, and internet sources including social media websites, polling organizations, and news websites. Survey data are comprised of informal and summary estimates relating to the event space (defined below), and survey data are distinguished from sample population data in that survey data may refer to multiple sources, may be more highly aggregated, and may require translations or estimations for mapping to contest event spaces. Social graph data are (i) granular data relating to individual contestants which indicates their personal, family, product, professional, commercial, transactional, and community preferences, and linkages and activity on the internet and social networking sites (the “individual social graphs”); and (ii) aggregated data relating to a contestant pool indicating commonality in one or more Individual social graph categories (the “aggregate social graph”). In a disclosed embodiment, individual social graph data are obtained by the system through an opt-in, linking, or other website access mechanic facilitated by the contestant through the “consideration process” (defined below) and the system acquires the data through a social graph API or essentially similar computer protocol. Data sources may include general internet content or specific content from social media sites such as Facebook, Google, Foursquare and Twitter. The system derives aggregate social graph data from the available individual social graph data. Reality mining data consists of one or more datasets accessed by the system where the data identifies potential and actual contestants at all stages of system activity with respect to: (i) the absolute and relative location of an individual's internet or network device (where absolute location means a level of specific address or GPS coordinates detail, and where relative location means proximity relative to other system identified internet or network devices associated with other individuals), (ii) the time of day and frequency which an individual interacts with the related internet or network device, and (iii) the wired or wireless systems used by an internet or network device to connect either directly or indirectly to the system.

The disclosed embodiment processes and controls five aspects of an internet contest:

• • 1. The contestant pool: Where a preferred embodiment includes system contestant control, drawing from a wide pool of potential contestants, the system optimizes the construction of the contestant pool through statistical measurement and screening processes for projected contestant selection dispersion. Where the preferred embodiment does not include system contestant control, the system will rely on system active selection control only.
  • 2. The contest space: the interface displayed on an individual contestant's internet device, which displays (i) balloting choice on a contestant's device; (ii) an availability of inputs and choices based on aggregate contestant activity; (iii) a description and instructions for participation; and (iv) control relating to the opening and closing of contest rounds and system access. The system's active and automated selection control process has the effect of managing contestant dispersion in the event that realized contestant choice selections exhibit higher concentrations than predicted at the onset of a contest. Unless otherwise specified, contest selections or ranges are the system presented alternatives to an incremental contestant, a contestant choice or contestant selection is the raw selection indicated by a contestant on a linked device, and a contestant entry is a system accepted and system processed finalized contestant ballot for consideration in the rewards processing.
  • 3. The reporting space: the interface and reporting system for administration of the program by the contest sponsor or sponsor's agent, which includes all granular and summary data relating to participant balloting, outcomes, and rewards.
  • 4. The event space: the feasible set of the future event or events with uncertain outcomes which the probabilistic prediction is over. The event space is characterized by either (i) a discrete variable (e.g. will the stock market rise or fall?), or (ii) a continuous variable (e.g. how much will the stock market rise or fall?). For example, in the context of a deck ordinary playing cards, the event space of a single unbiased draw is the 52 individual cards. In a preferred embodiment, the event space may relate to an uncertain event exogenous to the contestants selections such as financial markets or meteorological data. In another preferred embodiment, the uncertain variable impacting the event space may be endogenous to the contest and a direct consequence of contestant selections; an example of an endogenous uncertainty is a survivor contest in which contestants compete to list the 100 largest countries of the world where a country is removed from the available selections once it has already been identified by more than fifty percent of the contestants. For the purposes herein, reference to a continuous variable is meant to have conventional meaning; further, for the purposes herein, a near-continuous variable or near-continuous range of selection is meant to refer to range consisting of a finite number of choices where the number of choices is more than would typically be presented in a selection list, checklist or radio button display (e.g. more than 20 discrete selections or choices).
  • 5. The consideration and reward: the system processes consideration elements (content and value coming from contestants) and rewards (value delivered to the contestant). In a preferred embodiment, consideration consists of a contestant's individual social graph data, and in certain other preferred embodiments, cash amounts which may or may not be related to product purchases. Reward amount processing encompasses a tally of all contestant inputs, a comparison with the event space realization, and processing rewards consistent with the contest space rules.

Furthermore, the present disclosure enables realtime control using a specialized computer system to control dispersion of contestant selection. With control of contest selection, the contestants can not all pile into one category or have the same selection. Therefore in a hypothetical contest, multiple contestants sharing a limited prize pot is avoided and the likelihood of a concentrated outcome is remote. Depending on the embodiment, this objective is implemented by using at least one of the following processes that may be utilized in conjunction with each other, or individually implemented. These processes are system range control and contestant control.

In system range control the contest space changes. Depending on the implementation, the contest space may be altered by changing the relative contest interface such that each contest may have more or less or different selections available than other contestants. This adjustment to the contest space is done realtime through the computer system and will adjust the contest space to reduce concentration issues.

In contestant control, the computer system selects contestants useful for a particular contest. Self selection of contestants by the computer system permits a evenly distributed contest pool such that statistically the contestants will not have the same proclivity. There are various ways this selectivity is accomplished. For example, depending on the embodiment, selectivity may be based on and include, but not limited to, various demographics, purchasing trends, contestant entry history, financial information, and the like.

Adverting to the drawings, FIG. 1 shows a windows display 20 on a contestant's internet access device (desktop computer, tablet, or other internet device) indicating (i) one of more product promotions or contests 22, (ii) a description of the contest space and event space 26, and (iii) a window for inputting the user predictions of the uncertain event 28.

Examples for promotions at 22 include: (1) an all expenses paid vacation sponsored by an airline; (2) discounts on the purchase of an automobile sponsored by a car company; and (3) a $1 million cash prize offered by a gaming industry company.

The user display 20 may be presented in a multiple offering format as depicted in FIG. 1 or may be presented in a single offering format. FIG. 2 is an extension of FIG. 1, differing in that FIG. 2 displays a single offer 22 and a single related contest space 26 and an area for inputting the contestant's entry as either an application window or a texting area 28.

In the embodiments depicted in both FIG. 1 and FIG. 2, the promotional space occupies the upper half of the display and the contest particulars occupy the lower half of the screen. Other embodiments may alter the positioning of the screen areas or use windowing to overlay the areas.

FIG. 2 depicts an embodiment in which the contest is presented in two regions of the display. The contest description is contained in a sub-area of the display 26, wherein examples include: (i) “predict the direction of a widely published stock market index”; (ii) “rank order model investment portfolios by realized returns”; (iii) “predict the outcome of a machine process designed to produce random outcomes”; and (iv) “select the winner of a sporting contest”. In the FIG. 2 embodiment, the area for contestant input 28 is that region of the display in which the contestant enters a predictive ballot for system processing.

Embodiments of contest inputs 28 are depicted in FIGS. 3, 4, 5, and 6. FIG. 3 illustrates a contest input with three mutually exclusive choices selected through a radio button interface. FIG. 4 is an extension of FIG. 3 where the number of mutually exclusive choices, selected through a radio button, may be a large number. FIG. 5 illustrates a user interface with one or more point estimator choices in which the contestant enters one or more values rather than choosing from a direction or range. FIG. 6 illustrates a rank order input interface in which the contestant is presented with a number of categories, and a contest success is indicated by a successful unique rank ordering of the categories or presented items. Examples of such categories are horse racing results, investment portfolio performance, and political candidates percentage share of votes in an election.

The system controlled contestant interface includes: (i) the information displayed at display segment 26 over which the contestant will submit a selection or ballot is referred to as the event space, and event space 26 will refer to one of more future events with uncertain outcomes (e.g. “over how many of the next 5 trading days will the market finish with a gain?”), and (ii) the contestant input region 28 in addition to the event space 26 is referred to collectively as the contest space.

Alternative forms of the displays illustrated in FIG. 1 and FIG. 2 include, a rearrangement of display screen segments, multiple screen presentations with links or tabs to switch screens, scrolling screens displays, or pop-up windows, each an alternative embodiment.

FIG. 6 illustrates a preferred embodiment in which the contest space and event space relate to a time sequence contest in which, over the course of a stock market trading day, contestants will attempt to indicate both the high point for the day and the low point for the day as they occur. For example, if a contestant expected the stock market to decline continuously over the trading day, the contestant will be prepared to submit a “HIGH” indication at the 9:30 am market open and a “LOW” near the afternoon market close; of note, in this preferred embodiment, each contestant will submit “HIGH” and “LOW” selections real-time and each such selection submission will be uniquely linked to a time of day, and specifically linked to a stock market value. As illustrated in FIG. 6, the contestant submissions will be transmitted to the [system processor] 11 via a cellular or other wireless connection and the system will store a time entry data record 29 containing a unique contest identifier, the selection type (“HIGH” or “LOW”), the time of day the selection is received, and the related market value.

Prior to arriving at the screens depicted in FIG. 1 and FIG. 2, a user identifies himself through a combination of one or more unique identifiers including an email address, a phone number, a user name, a user code, an internet IP address, a mobile device ID, and a password for unique contestant identification by the system. The identification interface and processing is consistent with prevailing internet website practices.

Event space exogenous uncertain events include financial markets, sporting contests, and other forward-looking events. The system can electronically access information or data for such events from news media, market exchanges, governmental internet portals, or corporate websites.

Under system balloting, a contestant accesses the system through an internet or network interface, over one or more sessions. During a session, a user indicates or submits a value for one or more future events with uncertain outcomes. The system identifies each unique user by reference to interface inputs and a database residing on a storage device within the system such as a hard drive, RAM on a server or host computer or alternate rapid access cloud storage and retrieval medium. Based on system stored contestant entries and system retrieval of realized outcomes of events, the system utilizes computer processes of statistical permutations, statistical combinatorics, tests of statistical variance, and measures of statistical dispersion deviations to deliver system processed promotional offers in fields 22.

Outcomes in the event space 26 may be based on a continuous variable or a discrete variable. A continuous variable in an event space 26 is a variable which may take any value within a range: financial market indicies, time of day, and snowfall amounts are examples of continuous variables. A discrete variable in an event space 26 is a variable with a limited set of outcomes: coins tosses, playing cards, and horse races are examples of discrete variables.

FIG. 7A is a graphical illustration of an event space displaying a continuous variable and the manner in which the system transforms the continuous variable into a discrete choice. FIG. 7B is the corresponding contest space display 28. The symmetrical bell curve of the continuous variable in FIG. 7A, has a mean value of 1500, a normal distribution, and a standard deviation of 25% or 375 (1500×0.25). In adapting a continuous variable event space to a contest space with discrete user choices (see radio selection, user input space in FIG. 7B) the system creates ranges over the continuous space, by equally distributing the probability area across two or more ranges, which are then mapped into discrete range-based, mutually exclusive choices.

In FIG. 7A, the continuous variable of a market index level is mapped by the system into four discrete range choices labeled A through D, which are selected by a contestant through the radio buttons indicated in user input in FIG. 7B.

The boundary lines for each region (A through D) are delineated by the system such that the probability space is substantially equally represented in each region. The delineations may be performed by the system using either numerical or analytic routines. In certain financial markets the system may take into account related instrument pricing along different areas of the probability space and the system results may deviate slightly from a pure statistical result.

FIG. 8A is an illustration of an event space involving a discrete variable. In FIG. 8A, the event space is suits from a deck of standard playing cards (hearts, clubs, diamonds, and spades). Dice would also be representative of a discrete event space. Discrete variable event spaces either have a known probability distribution (in the case of fair dice or cards) or a probability which may be determined empirically by the system. The creation of a contest space for a discrete variable is illustrated in FIG. 8B in which, through the mutually exclusive radio button choices in user input segment provides a predictive choice.

A numerical example of a contest using both the disclosed embodiment and known methods is as follows:

Specifications for an example of a contest:

Contest space rounds: 3

Independent success probability: 25%

Event space: see FIG. 7A

Awards: any 1 of 3=$1.00; any 2 of 3=$3; 3 of 3=$10

TABLE 1
probability (p) 0.25
					combination		Expected Value
successes	trials	combinations	p{circumflex over ( )}s	(1 − p){circumflex over ( )}(t − s)	probability	$ Award	$Award
0	3	1	100.00%	42.19%	42.19%	0	0.00
1	3	3	25.00%	56.25%	42.19%	1	0.42
2	3	3	6.25%	75.00%	14.06%	3	0.42
3	3	1	1.56%	100.00%	1.56%	10	0.16
Totals		8			100.00%		1.00

The first four rows in Table 1 display the complete set of possible success/loss combinations over three rounds for a single contestant, where, within each round, the outcome is a mutually exclusive success or loss. The number of total successes over three rounds will take on an amount between 0 and 3, as indicted under the column heading “successes”. The column entitled “combinations” indicates the number of ways or combinations each total number of successes can occur. For example, there is only one permutation or path to three consecutive success, but three paths to a single success. For example, the “round 3” row in FIG. 9 has eight possible outcomes utilizing “L” and “S” (a single success “S”, indicated in the last row as “LLS”, “LSL”, and “SLL”) as explained below. The system applies a combinatorics equation:

$C_s^t=\frac{t!}{s!\left(t-s\right)!},$

where t equals the number of trials and s equals the number of successes.

Applying the above equation, using three rounds, and examining the number of combinations of 0, 1, 2, and 3 successes (table 1, column 1), indicates a number of outcome combinations equal to 8 (table 1, column 3, row/entry 5), where the number of combinations for each of the possible number of successes is 1, 3, 3, and 1 (table 1, column 3, entries 1-4) respectively for 0 through 3 successes.

FIG. 9 is a binomial tree illustration of the permutations of the Table 1 combinations. In FIG. 9, a movement right is indicated as a success or “S”, a movement left is indicated as a loss, or “L”, and a letter sequence of “SSL” indicates two consecutive successes followed by a loss and is depicted in FIG. 9 in the 7th node (left to right).

In Table 1, the column titled “p̂s”—read “p” or probability raised to the number of successes “s” indicated, the column titled “(1−p)̂(t−s)”—read the probability of loss or “1−p”, raised to the number of losses indicated or “t−s”, indicate the cumulative probability of the respective number of successes or losses. For example, the probability of three losses is (1−0.25)̂(3−0) or 42.1875% and is indicated in the top entry of column (1−p)̂(t−s).

The column entitled “combination probability” is the product of the immediately preceding three columns. The value for each row indicates the probability of the total number of success in the first column. For example, two of three success can be achieved through any of the following three permutations—SSL, SLS, LSS, and as indicated, the probability of two successes is 14.06% (from (3)×(6.25%)×(75%)).

The final numerical specification to the example contest is the value of the award for specified results. Under the heading “$ Award”, Table 1 indicates award amounts of $0, $1, $3, and $10 for 0, 1, 2, and 3 successes, respectively. With a specification of the award amounts for each success combination, the aggregate expected value of the contest can be calculated through the sum product of (i) the combination probability value, and (ii) the respective dollar award. The expected value for the entire contest illustrated in Table 1, using known methods, is $1.00 as indicated in the row total under the Table 1 heading “Expected Value $Award”.

FIG. 10 illustrates how the contest specified in Table 1 progresses through the three rounds based on known methods relating to a single contestant. The starting point is indicated at top center (“Start”). Each successive row down is a round. A movement to the right indicates a success in the round (25% probability for “S”) and a movement to the left indicates a loss in the round (75% probability for “L”). Note that there are eight nodes in the last row, consistent with the total number of combinations in Table 1. The nodes are indicated with a three letter sequence indicating the applicable permutation, where “S” equals success and “L” equals loss, “LLL” indicates three successive losses and “LLS” indicates two consecutive losses followed by a success. The value indicated at each point in the diagram is the expected value per contestant, based on the indicated path. Each value is the probability weighted expected value at that point; post-round 3 all values are known since all rounds have resolved, so the values indicated at round 3 are realized values as opposed to expected values. As a calculation example, the rightmost number in round 2 ($4.75) is the sum of (i) 0.25×$10, and (ii) 0.75×$3; all other values in FIG. 10 are calculated in a similar manner.

Unlike known methods, which proceed on the basis of the values in Table 1 and FIG. 10, in the disclosed embodiment, the system begins with an evaluation of a sample population or survey data indicating prospective (or sample) contestant tendencies relating to the contest selections. The disclosed embodiment executes a process which: (i) uses sample population data, survey data, and social graph data for constructing the contest space and sets concentration limits relating to selections or values choices within the contest space prior to the first round, (ii) evaluates the contest on a real-time basis to validate the original contest space construction and its conformity with projected aggregate award amounts, (iii) runs real-time monitoring over concentration limits of contest space selections, ranges or values, (iv) alters the contest space interface and related underlying systems upon indication of an anticipated or actual breach of a selection concentration limit, and (v) aggregates and validates results.

Before the contest, the system processes data relating to a population estimate over the event space which indicate selection tendencies of prospective contestants. Tests of concentration tendencies in the sample distribution and tests of modality or multi-modality are performed by the system. The concentration tendency tests are used by the system to determine the number of selections in the contest space, the relative positioning of the selections or ranges, and the contestant limit (or concentration limit) for each selection. The modality tests are utilized to position the contest space choices or ranges across the possible event space outcomes to split indications of projected peak concentrations.

Prior to the first round of a contest, the operator of the system in the disclosed embodiment creates a pro-form a contest space consistent with FIG. 3 through FIG. 6, where a number of preliminary ranges is identified. The system uses the sample population estimate and populates the pro-form a contest space. Table 2, below, is an example of the process in a preferred embodiment.

	TABLE 2
		Population	Range	Maximum
	Ranges	Statistics	Limit	Chi-Squared
	Range A	9%	na	3.75%
	Range B	12%	na	5.00%
	Range C	27%	na	11.25%
	Range D	52%	80.00%	80.00%

Table 2—columns 1 and 2 display the results of a sample population across four choice selections. The sample population poll figures are displayed in the column titled “Population Statistics” and they indicate that a sample population has a rank preference order, from high to low, of D, C, B, and then A. In addition, 52% of the sample population indicates a preference for Range D (that is, a majority of contestants in the sample population chose this option). The values under the remaining columns are explained below.

Using the sample data from Table 2, the system runs a process in which the system performs the following steps:

(1) establishes a statistical confidence level relating to breaching (going beyond) a series of tested selection or range limitation values, beginning with the highest value from the sample population (i.e. 52% in Table 2);

(2) incrementally increases the highest concentration selection or range (Range D) from its sample population value percentage of 52% to 100% (as explained below);

(3) redistributes the resultant amounts attributable to Ranges A through C in proportions consistent with their sample population distribution;

(4) calculates a Pearson's chi-squared test or a similar statistical test of relative distributions over the sample population distribution versus the pro-form a populations of steps (2) and (3);

(5) determines the related confidence level of the statistical value computed at step (4); and

(6) setting the range limitation equal to that value of the lowest concentration range (Range D in this example) for which the statistical confidence level in Step (5) equals the a confidence level established by the contest administrator. Examples of typical, although not limiting, confidence intervals are 80%, 95%, 99%, 99.99%, and 99.9999%.

The sequential process of Steps (2) and (3) above is illustrated in FIG. 11. Beginning with the leftmost bar in the figure, all of the range values start with their sample population percentages (i.e. those values in Table 2, column “Population Statistics”). Moving right in the figure, the Range D value steps from 52(%), to 60, to 70, to 80 where ranges A through C take a value which is equal to the non-range D residual value (i.e., 100-52=48) in a proportion equal to the non-range D sample proportion (i.e., sample A=9, so range A is equal to 9/48=0.1875 of the non-range D value in each trial). A process which maintains the relative proportions, for ranges A to C, will result in the lowest chi-squared value and the lowest confidence level, and therefore the most conservative test result.

FIG. 12 is a graphical representation of the process of Steps (2) through (6) above where FIG. 12 illustrates the relationship between the pro-form a range D value limits (e.g. 52%, 60%, 70% and 80%) and the associated statistical confidence level associated with observing a range D value exceeding such value. The FIG. 12 graph indicates the range D concentration on the x-axis, and the statistical confidence level on the y-axis. The seven values labeled on the curve are the x/y values. FIG. 12 uses a sample size of 100, and different implementations of the embodiment may use different sample sizes.

At each point, the interpretation of the FIG. 12 graph is; “for each value of range D concentration, what is the statistical certainty that a range D value higher than that indicated will not be realized by the system in an actual contest population?”. Referring to the graph, the range D values of 52, 60, and 80 indicate confidence levels of 0%, 72.253%, and 100% respectively. The inference for the system is: (i) we have no certainty that a value above 52 will be not realized; (ii) the system can be 72.253% confident that a range D value larger than 60 will be realized; and (iii) the system can be approximately 100% certain that a value greater than 80 will not be realized. This means that no more than 80% of the contestants will occupy any single selection in any round. To restate, the system indicates that the likelihood of more than 80% of the contestants in the actual contest populating or choosing a single range is effectively 0%.

Referring back to Table 2 above, the values in the two rightmost columns are the result of the six system steps above. The column “Range Limit” is the limitation imposed on the selection with the highest concentration limit, and the value is determined through the system step-wise process which seeks that minimum value (i.e. 80% concentration) which relates to low probability of occurrence (e.g. a probability less than 5%, 1% or lower). The column “Maximum Chi-Squared” is the most likely distribution of across the selections given an 80% occurrence in the selection with the highest concentration; for example, the value associated with Range C is (i) the sample concentration, over (ii) the total non-Range D concentration, all times (iii) 1 minus the Range Limit value or [(0.27/0.48)×(1−0.80)=11.25%]. Other values are computed similarly.

The system sets a range limitation equal to 80% based on the sample population of Table 2. The remaining calculations relating to the 80% limitation are indicated in Table 3 below.

TABLE 3
Chi-Squred Value @ System Maximum 31.41
degrees of freedom               2
One-tailed test probability      0.00002%
Confidence Level                 99.99998%

The Chi-Squared Value of 31.41 reported in Table 3 is based on the Table 2 sample population (under “Population Statistics”) and the system determined maximum distribution (under “Maximum Chi-Squared”), the 4×1 (row/column) sample population possesses two degrees of freedom). Using a computer algorithm to determine the one-tailed test probability indicates a value of 0.00002%. The confidence level associated with not experiencing a range D concentration level in excess of the 80% maximum is almost 100%.

In a preferred embodiment in which contestant selections are based on time stamping or point-in-time identification (e.g. contestants attempt to identify the high and or low points in a financial market on a real-time basis), the system will substitute the chi-squared type of testing with testing based on a poisson or binomial distribution. In such a preferred embodiment, all other aspects of the process will be similar.

In a preferred embodiment, the system performs this range limitation process prior to each round of the contest, and updates the range limitation depending upon results; for example in a preferred embodiment, if a selection other than Range D captures a majority concentration, the steps above would be recast based on the new majority selection. In subsequent rounds, the actual contestant population statistics supplement the sample population and all other parts of the process will be identical.

In contrast to known methods, the system applies a concentration limit to the selections in the contest space. Using the above example, the concentration limit for any selection in any round is 80%. That is, as indicated, no more than 80% of the contestants may occupy any single selection in any round. Based on the selection limitation testing example above, the likelihood that the limitation will be effected is almost 0%; no expected contestant selection is expected to be perturbed by the limitation.

In a preferred embodiment, if the contest is run at a time when contest selection are trending in a general direction over the duration of the contest, the system's range limitation process may produce range limitations which increase or decrease from one round to the next (e.g. 80% to 85% to 90% over three rounds as the above listed steps process is re-run between rounds of a contest).

Event space FIG. 13 is identical to FIG. 10, except that FIG. 14 introduces a constant 90% contestant selection concentration limitation for each range for each of the three rounds. FIG. 14 is also identical to FIG. 10 except that it introduces an 80% range concentration limit for each range for each round; the value indicated by the six step system process above. In each figure, the beginning expected values are indicated at top-center under the heading “Start”.

The expected values indicate the probability weighted expected value per contestant. The value for FIG. 10 under the known method is $1.00, while the values for FIGS. 13 and 14 are $0.73 and $0.51—reductions of 27% and 49% respectively. FIG. 15 contains further detail on the results associated with selection concentration limit. Further, the values for FIGS. 13 and 14 are calculated in the same manner as those in FIG. 10 except for the system imposed concentration limits of 90% and 80% respectively. By way of numerical example, the rightmost value for FIG. 14 at round 3 is calculated as the original $1.00 reward value adjusted for the system imposed concentration or survival percentage applied at each round; that is, only 80% of the contestant can occupy a single selection at each round, so the maximum number of contestants succeeding in round 3 is 0.80-cubed or 0.512, and 0.512 times $1.00 equals $5.12. Other values in FIGS. 13 and 14 are computed in a similar manner. Recall that the system run statistical processes have determined that the probability of the concentration limit actually affecting expected contestant selection is almost 0%.

As indicated in Table 3, above, the likelihood of contestants overpopulating a range is less than 0.0001% in the initial round. Similarly, the likelihood of contestants overpopulating any range during any of the three rounds can be calculated by the complement of the overpopulation probability cubed, or (1−0.00001)̂3=99.997%, which remains statistically immaterial.

Implementation of the disclosed embodiment will permit a sponsor or agent of the sponsor to reduce the liability reserve against the promotion by a considerable amount: 27% or 49% as indicated in the above examples relating to 90% and 80% selection limitations respectively.

In a preferred embodiment, the system creates and tracks two selection concentration limits: a preliminary selection concentration limit and a final selection concentration limit for implementation in intermediate contest rounds. The preliminary selection concentration limit causes the system to modify the contest space, and the final concentration limit indicates the maximum concentration at the end of the related contest round accounting for the application of the contestant selection limitation.

FIG. 15 illustrates the system contestant filtering process which occurs for each round through the selection concentration limit process. Beginning at the top of the diagram, the illustration assumes (i) 100 initial contestants, (ii) a concentration limitation for any range in any round is 90%, (iii) 90 of the 100 contestants (or 90% of contestants) have selected Range D and 10 (or 10%) have chosen Range C, and (iv) the event space realization conforms with Range D such that 90% of the then remaining contestants successfully advance through the first of three rounds.

In a preferred embodiment, the system selection limitation will restrict the contestant selections such that the successful contestant count is always equal to or less than 90% of the successful contestants of the immediately preceding round. The maximum number of successful contestants, indicated as a percentage, is equal the selection concentration limit percentage raised to the number of contest rounds. In the figure, (0.90)/1 or 72.9% as confirmed in the lower right-hand corner of FIG. 15.

During a contest, in a preferred embodiment, the system monitors the realized predictive balloting and compares it to the projected predictive balloting computed immediately prior to the contest start. Using the example from Table 2 above, it should be highly unlikely that observed balloting (analyzed in statistically significant quantities) should approach individual range concentration of 80%. Such an observation during the balloting indicates anomalous outcomes as compared with the sample population or survey data. In a preferred embodiment, upon the indication or actual occurrence of predictive balloting anomalies, the system adopts one of the following seven options to effect real-time control of the selection of contestant dispersal balloting:

• • 1. None: The system does not alter the contest space in response to balloting anomalies, and while highly unexpected based on system computations, contestant selection crowding may exceed selection limitations set at the contest inception.
  • 2. Close Choice or Selection: The system closes a choice or range from further selection by contestants who have not yet balloted.
  • 3. Bifurcation of Choice, Selection or Range: The system bifurcates a choice or selection into one or more additional choices; Bifurcation may optionally result in additional available selections. Pre-existing ranges may optionally be removed or merged such that the selection count does not increase.
  • 4. Adjust Choice or Selection Boundaries: The system shifts the mapping of contest selections to the event space outcomes by reducing the event space probability linked to overpopulated choices and increasing the probability space linked to under-populated choices.
  • 5. Queue and Roll: In a preferred embodiment in which contestant entries are time based and submitted serially (e.g. contestants attempt to identify the high and or low point in a financial market on a real-time basis), the system will queue contestant submissions on a first-come basis. Where concentration limits set by a combination of poisson or binomial distribution testing are breached, the system will alter contestant ballots (rather than adjusting the a priori selections) and push or roll contestant ballots forward to the next unlimited available time designation; FIGS. 18A through 18C illustrate a real world example of Queue and Roll.
  • 6. Pari-Mutuel Adjustment: In a preferred embodiment in which selection control measures have not sufficiently reacted to contestant selection activity, and where contestant selections concentration exceeds preset limits, the system will apply a fractional factor to contestants in a round in a successful, over-populated choice or selection; a pari-mutuel adjustment will be effected by the system in the event of a failure of computer systems, a disruption in communications links, or an overall market disruption The fraction will be equal to the pre-determined range limitation (e.g. 80%) divided by the actual percentage contestants populating the range (e.g. 90%). For example in a two-round, two choice contest where the range limitation for each round is 80%, and the population occupying the successful range in both rounds is 90%, the factor is ( 8/9)² or 0.79. The final reward attributable to each successful contestant in the overpopulated choice for each of the two rounds is 79% of the basic reward.
  • 7. Additional Round or Rounds: When contestants attempt to select an over-populated round, they are given the choice of (i) making an alternate selection, or (ii) by-passing the current round, but participating an additional round, where such additional round only includes those contestants precluded from a choice or selection in an earlier round.

FIGS. 16, 17A and 17B illustrate a real-time system modification to the contest space. The modification displayed in the figures is performed by the system upon the breach or near breach of a selection concentration limit. Through real-time monitoring of the contestant ballot submissions, a breach or near-breach of a selection concentration limit set immediately preceding the related round has occurred.

FIG. 16 displays a radio button selector which has been modified from the radio selector button in FIG. 3. Recall an event space is the feasible set of outcomes relating to the unknown future event, while a contest space includes the system interface within which the contestants will select or ballot over the event space. The system modified the contest space in that a fourth button has been added and the UP range is split into an UP1 and UP2. The original number of contest space ranges indicated by FIG. 3 is three, and the revised number of contest space ranges indicated by FIG. 16 is four. The original event space ranges are indicated by the graph in FIG. 17A as UP, UNCHANGED and DOWN. In contrast the graph in FIG. 17B indicates the system revised event space ranges (increased from 3 to 4) which conform to the revised contest space input buttons.

FIGS. 18A through 18C illustrate a preferred embodiment in which the contestant balloting and selection process is coincidental with the related event space realization. FIG. 18A illustrates an event space variable which begins at 100, and over the course of a single period (with sub-periods indicated as T1, T2, T3, T4, and T5) increases to a value above 115. At identified discrete points in time during the period (the sub-periods), the graph in FIG. 18A indicates the specific values at times T1, T2, T3, T4, and T5 as 110, 111, 111, 113, and 115 respectively. Of particular note, the value of the event space variable has remained static between the time T2 and T3.

FIG. 18B illustrates raw (system unadjusted) selections in a time based balloting contest relating to the data in FIG. 18A. The four contestants are indicated as contestants A, B, C, and D, and they have selected times T1, T2, T3, and T4 respectively. Each time is associated with a value as indicated in FIG. 18B. Of particular note, the event space variable value is identical for the time selections made by contestants B and C. A real-world example of the FIG. 18A through FIG. 18C illustration is a contest in which contestants must identify the high point in the day's stock market as such high point is occurring.

FIG. 18C illustrates the system adjusted contestant ballots. For simplicity of illustration, the FIG. 18C example assumes that only one contestant can occupy a single point in time. Recall that in actual practice, this limit would be determined by the system in a manner consistent with example illustrated in Tables 1, 2, and 3. As indicated in FIG. 18C, contestant C, because it occupies the same event space value as the contestant B ballot, is rolled forward to the next unique time and event space value combination to T4 and 113 respectively. The system tracks the original queue order, and consistent with rolling contestant C forward to T4 and 113, contestant D's ballot is also system adjusted to time T5 and event space value 115 to retain queue consistency with contestant C's ballot. All other contestant ballots would be processed by the system in the same manner.

In the context of the example illustrated in FIGS. 18A, 18B, and 18C, the granularity of time will vary based on embodiment and implementation. A point in time in the context of a financial market will be based on the precision with which time is reportable in that market; typically a second or fraction thereof. In contrast, a point in time in the context of meteorological events over the course of a month would be a single calendar day.

FIG. 19 illustrates a real-time system modification in the context of a contest in which a specified number of contestants will compete over identifying the rank order of the 100 largest countries in the world by population (for illustration purposes, FIG. 19 bounds the list of 100 countries by identifying China at the top and Bulgaria at the bottom). In the contest depicted in FIG. 19, the system would announce that Australia 40, is the middle or 50^th country in the list and that in progressive rounds contestants must identify larger and then smaller countries in alternating order based on the country chosen in the immediately preceding round; for example, following the system announcement of Australia, a sequence of Italy, Belgium, Mexico, and Sweden would denote success in the contest through four rounds. At the onset of round 1 in which contestants are to choose a larger country than that last identified (Australia), in the first column 42, all selections are available. The second column entitled LARGER-1′ 44, indicates the contest space during the processing of round one, and as indicated the contest space is modified in that Mexico is removed from the contest space; Mexico is removed from the contest space in the illustration because a system imposed concentration limit has been exceeded. The system has modified the contest space set of selections to remove one selection. Continuing to the columns entitled SMALLER-1 46 and SMALLER-1′ 48, it is assumed that a system imposed concentration limit relating to Sweden has been exceeded, and similarly, the system will modify the contest space to remove Sweden from the set of selections within the contest space. The example relating to FIG. 19 illustrates a preferred embodiment in which the event space is near-continuous and where the uncertain and statistical aspects of the event space are endogenous with respect to the contestant balloting process.

FIG. 20 is a diagram indicating the connectivity of the system 10 to (i) contestant's display and input devices 20, (ii) the uncertain event data providers at 13 and 14, (iii) the contestant preference processing system 15 and contestant database 16, (iv) the cloud linkage to effect social graph linkages and searches where the system links to systems such as, but not limited to, Facebook, Google+, Twitter, Foursquare, Linkedin, and other community and preference sites 17, and (v) the contest rewards processing system 18 and the contest rewards computer storage device 19 which may be discs, volatile and non-volatile RAM, or other storage media.

The system processor 11 controls all aspects of the contest space and transmits web site pages, email, and messages through the internet or computer network to control contestant user devices. The contestant displays are as indicated in FIG. 1 and FIG. 2. Contestants may have the option to access the contest system through the internet, wireless public and private networks, and cellular systems.

The system processor 11 contains the entire processing infrastructure for running the contest. The system relies on external data, market and news service providers 13 and 14 for data relating to future events (where the future event is exogenous to contestant balloting) and for the realization or results of future events relating to the contest. The system accesses uncertain event data providers through an internet linkage 12. Uncertain event data providers include financial exchanges, news and media providers, government agencies, and gaming industry concerns. Where contestant activity is part of the system processed uncertain future events, data stored at the central storage device 16 will supplement the data obtained from uncertain event data providers 13 and 14.

The system processor 11 identifies individual contestants and logs all contestant inputs and preferences at a central storage device 16. The system stores all contestant choices relating to promotional selections and contest entries in addition to the data uniquely identifying each contestant. The processing system 15 for the contest preference discerns patterns of selection, product or promotion preferences, and visit frequency on the contest events. The system accesses social graph content through the internet or through linkages to existing computer social networks 17 including Facebook, Google+, Foursquare and Twitter, and Linkedin. The system has a synchronous communications linkage with social networking sites enabling contestant social graph data to be updated and modified by the contestant identification and preference processing system

The system processor 11 stores contestant entry data and event data for the purpose of rewards processing on a processing subsystem 18 for the contest rewards. The system, including its storage devices, contains all contestant selections and all event outcomes. The data maintained are used by the system for both award determination and for processing and altering the contest space during a contest.

FIG. 21 is a timeline illustration of a preferred embodiment and how the system processes round balloting in relation to each related phase or realization of the event space. Beginning at step s60, the system receives and processes contestant ballots for the initial round (61). At step s62, balloting for the first round is closed, and the event space for round 1 balloting begins (63). At step s64, balloting for round 2 begins, (65), and is coincidental with event space 1. At step s66, round 2 balloting, (65) ends, event space 1, (63) ends, and event space 2, (67) begins. At step s68, balloting for round 3 begins, (69), and is coincidental with event space 2. At step s70, round 3 balloting, (69), ends, event space 2, (67) ends, and event space 3, (71) begins. Lastly, at step s72, event space 3 ends, and the contest steps end. Optionally a differing number of rounds or time gaps between the components of rounds, however, for a preferred embodiment, balloting for a related event space is nonetheless completed prior to the commencement of the related future event.

FIG. 22A and FIG. 22B are process diagrams which illustrate the sequencing and interactions of the system steps and processes. A specific contest example is followed below. Beginning with FIG. 22A, at step s80, the base contest parameters for system processing are specified, e.g., by a Contest Administrator. The base contest parameters are:

• • (1) the target number of contestants: The discrete number of individual contestants specified by the administrators; the target number will be used in connection with a projected yield percentage (i.e., what percentage of solicited potential contestants will enter the contest due to system screening and contestant participation) to broadcast a solicitation message, webpage, or email over the internet. In the contest example, it is assumed that the target number of contestants is 10,000, and the projected yield percentage is 30%. Accordingly, 3,000 potential contestants will be projected by the system for inclusion.
  • (2) the target contestant audience: The one or more social graph indicator characteristics that identify the contestant pool. For the purpose of the contest example, it is assumed that the target contest audience, based on the primary social graph indicator includes owners of detached single family homes in the northeast and mid-Atlantic states.
  • (3) the event space: The future event with an uncertain outcome over which the contestants will submit predictive ballots. For the purpose of the contest illustration, the event space will be the price of oil.
  • (4) the contest space and balloting: The balloting is both (i) the input area choice selection into which the contestant enters a predictive selection, and (ii) the number of ballot rounds in the contest (e.g., 1, 2, or 3). Examples of different forms of balloting are contained in FIGS. 3 through 6. For the purpose of the contest example, the form of balloting will be a 4 range selection (determined below and based on the system calculated probability of 25%) in the form of FIG. 4, and the number of balloting rounds is 3.
  • (5) the reward: The reward is comprised of (i) the aggregate dollar amount reward (either a projected or an actual limited amount) divided by the number of contestants, and (ii) the reward for each possible combination of successes and losses. With 3 rounds, there are 4 possible result combinations as illustrated above in Table 2. For the purpose of the contest example the aggregate maximum projected reward per contestant is $50, and the rewards are $0, $50, $100, and $1000 for each of 0, any 1, any 2, and any 3 successes.

A summary of the base contest parameters is: (1 & 2) 3,000 to 10,000 homeowners, (3) the contest is over the price of oil, where each round's event space is one month, (4) ballots are as indicated in FIG. 4 and the number of rounds is 3, and (5) the desired aggregate unadjusted award is capped at approximately $50 per contestant, with rewards of $0, $50, $100, and $1000 for each of 0, 1, 2, and 3 successes in the 3 rounds.

Continuing with FIG. 22A at step s82, the system generates the required probabilities for success in each round of the contest. The system calculates the required success probability such that the award per contest is approximately equal to the sum of all reward combinatoric values, where each combinatoric value is equal to:

C(_s^r)×p^s×(1−p)^(r-s)

where:

C(_s^r)=the combinatoric r-choose-s, or the number of ways s successes can result within r rounds

s=the number of successes

r=the number of rounds

p=the system calculated probability

Using the contest example, where C(s−r) is used to denote the number of combinations of s within r and where A₁ is the desired or targeted system award (i.e. $50) and where A₂, A₃, and A₄ are the 1-success award, the 2-success award, and the 3-success award respectively:

A₁˜[A₂×C(1-3)×p¹×(1−p)²]

+[A₃×C(2-3)×p²×(1−p)¹]

+[A₄×C(3-3)×p³×(1−p)⁰]

$50˜[$50×C(1-3)×p¹×(1−p)²]+[$100×C(2-3)×p²×(1−p)¹]+[$1000×C(3-3)×p³×(1−p)⁰]

where C(1-3), C(2-3), and C(3-3) are equal to 3, 3, and 1 respectively, the system calculates a probability or “p” value of 0.25.

$50˜[$50×3×0.25×0.5625]+[$100×3×0.0625×0.75]×[$1000×1×0.015625×1]

$50˜$50.78125; note that the $0.78125 overage is insignificant given that the system dramatically decreases the realized amount as demonstrated in the examples relating to FIGS. 14 and 15.

For completeness, the reward value associated with a probability of 0.20 is $36.80, and the value associated with a probability of 0.30 is $67.95.

Continuing with FIG. 22A at step s84, the system can feature a contestant pool control process. If no contestant pool control procedures are implemented, the system queries the presence of a system selection control process. It can be noted that known or traditional contest methods possesses neither the contestant pool control or contestant selection dispersion control and traditional contests can be traced through the diagram by following “NO” at both control nodes at step s84 and s85.

Continuing with FIG. 22A, at s86 where the system implements contestant control, the system identifies and processes a comprehensive population of potential contestants with a social graph attribute matching the target contest audience in the base contest parameters—in the contest example, homeowners.

The system processes the comprehensive population into groups with identified higher order (e.g., secondary) aspects of their social graph attributes or reality mining attributes for fine-tuning expected contestant selection dispersion. Examples of secondary social graph aspect grouping are professions, educational institutions, subscriptions, club or interest group memberships, political affiliations, participation in blogs, forums and instant messaging networks. Examples of reality mining data for fine-tuning are home location, work location, driving patterns and trip frequency, mass transit use, airline travel, commuting habits, and times of morning rise and sleeping.

For the purpose of our contest example, driving patterns are the higher order grouping indicator. Examples of driving patterns are (i) individuals who drive more than 25 miles per day, (ii) individuals who drive between 5 and 25 miles per day, (iii) and individuals who drive fewer than 25 miles per week.

Temporarily referring to FIG. 23A (FIG. 23A through FIG. 23C is a detailed illustration of s86 and s88) at Step s1, the system identifies groups A, B, and C, three primary groups and for the purposes of the contest example, the groups represent high-mileage drivers (A) 120, medium-mileage drivers (B) 121, and low mileage drivers (C) 122. Moving to FIG. 23B and step s2, the system identifies the sample population or survey data for each of the groups A, B and C. The FIG. 23B histograms indicate the event space balloting of a representative sample of the related group, where the balloting histogram for group A 123 has a left-leaning bias, the balloting histogram for group B 124 has a right-leaning bias, and the balloting histogram for group C 125 has a neutral bias. The sample balloting is derived from an email survey, a web site survey, a blog survey, or data drawn from trade association materials, media or other means.

The contest space designation of the histograms in FIG. 23B is as follows: (i) the left-most column indicates a contestant selection consistent with oil prices down materially, (ii) the second column from the left indicates a contestant selection consistent with oil prices down slightly, (iii) the third column from the left indicates a contestant selection consistent with oil prices up slightly, and (iv) the right-most column indicates a contestant selection consistent with oil prices up materially.

In each instance the histograms are indications of the respective groups projected contest balloting. A visual examination of the histograms, and considering that ranges to the left indicate a lower oil price expectation and ranges to the right indicate a higher oil price expectation, the survey data indicates that Group A 123 has a lower price bias, Group B 124 has a higher price bias, and Group C 125 has a stable price expectation bias.

Moving to FIG. 23C at Step s3 (and step s88 in FIG. 22A, “SYSTEM GENERATES OPTIMAL GROUPS”), the system processes each combination of the contestant groups as illustrated in the table 126. The system combines the survey data distributions in every permutation as indicated in the table and performs a statistical test to determine the likelihood of a single range majority observation during the contest. For the purposes of the contest example, each combination is tested for the likelihood of observing a simple 51% majority in a single range.

Continuing with FIG. 23C at step s3 126, the statistical values for each combination (where combinations include every instance of a single group and its combination with other groups) are reported. For the purposes of the contest example a Pearson's Chi-Squared test (over a 4 by 1 table) was performed and the results are reported in the rightmost column of the table 126.

Values in the FIG. 23C table below 8.0 (e.g. A alone, B alone, C alone, and C+A in combination) indicate a high likelihood that simple range majority balloting will occur. Values between 8 and 12 (A+B+C in combination, and B+C in combination) indicate a lower likelihood that a simple range majority balloting will occur and the related statistical confidence level is approximately 99%. Values of 15 or higher (A+B in combination) indicate a very low likelihood that a simple majority balloting will occur and the related confidence level exceeds 99.95%. As indicated in the system results of Table 126, the system identifies the optimal contestant base grouping to be comprised of Groups A and B, because the combination of Groups A and B possesses the highest estimated statistical dispersion of sample balloting.

Moving to FIG. 23D at step s4 (and moving to step s90 in FIG. 22A—“SYSTEM LOCATES ADDITIONAL CONTESTANTS”), at graphic 128 the system examines commonality in the social graph data of the preliminary contestant population, where the preliminary contestant population is the combination of groups A and B, with other contestant candidate groups. Included in the social graph commonality are educational affiliations, shopping histories, investing behavior similarities, linked friends or social groups and links to public or private concerns or causes. Included in the reality mining data are travel patterns, device locations, and times and frequency of internet or network device use. The figure at center of the graphic 128 indicates the representative contestant and the overall graphic 128 illustrates the system generated social graph and reality mining components. The system will, through a web crawling exercise and through social graph API protocols, compile a third group of contestants, if one exists, which have social graph data or reality mining data matching the social graph data in the preliminary base contestants group determined as the output 127 in step 3. Depending upon the implementation, certain social graph data may be emphasized or over-represented. For the purposes of the contest example relating to oil prices, social graph data relating to interests in news services, global politics, and financial markets may be over-represented. For the purpose of the contest example, a Group Z 129, is indicated in step 4 as an identifiable group possessing social graph data similar to the preliminary contestant group.

Moving to FIG. 23E at step s5 (and step 90 in FIG. 22A—“SYSTEM LOCATES ADDITIONAL CONTESTANTS”), the system compares the projected contest ballot dispersion of (x) the preliminary group of A and B processed as illustrated in the table in FIG. 23C 127, with (y) the preliminary group of A+B plus the additional contestant group of Z 130. The system runs a process identical to that FIG. 23B Step 2 and FIG. 23C step 3 with respect to (i) A+B, (ii) A+B+Z. For the contest example, it is assumed that the pool combination of A+B+Z indicates a higher chi-squared statistic than A+B and further reduces the probability of concentrated balloting or a simple majority ballot selection (i.e. a single selection in excess of 50%)

Some implementation may follow different rules than those immediately above and some embodiments may only alter a preliminary contestant pool.

Moving to FIG. 22A at step s92, the system sets choice or selection limitations on contestant choices. Applying the contest example which includes Groups A, B, and Z (and assuming Group Z possessed a similar distribution to the aggregate A+B Group), the system calculates the applicable range limitation based on the sample population data from the contestant pool.

Table 5, below, is in an identical format to Table 2, and Table 5 uses the A+B+Z distribution under the column heading “Population Statistics”, and the system applies a 70% limitation to contestant selections. As in Table 2, the column entitled “Maximum Chi-Squared” contains the distribution consistent with observance of a 70% single choice population and the remaining ranges are filled in a manner to produce the lowest chi-squared test result and therefore the most conservative statistical measurement.

	TABLE 5
		Population	Range	Maximum
	Ranges	Statistics	Limit	Chi-Squared
	Range A	17.5%	na	7.78%
	Range B	32.5%	70.00%	70.00%
	Range C	22.5%	na	10.00%
	Range D	27.5%	na	12.22%

Continuing with FIG. 22A at step s92, the system calculates the confidence level associated with a 70% limitation in a round to be approximately 100%, and the system sets a 70% contestant choice limitation for any selection in any round That is, the system will not allow more than 70% of the contestants to occupy any single range at any time during the contest, and the probability of the 70% limitation perturbing actual contestant choice is approximately 0% (i.e. approximately 100% confidence level—see Table 6 “confidence level”).

TABLE 6
Chi-Squred Value @ System Maximum 64.1
degrees of freedom               2
One-tailed test probability      0.00000%
Confidence Level                 100.00000%

In FIG. 22A moving from step s94 to s87, the system applies contestant dispersion control based input from the contest administrator. As indicated, at step s89, the administrator specifies no control, by indicating “NONE”, or the administrator selects one of the methods of contestant choice control indicated at step s89. For the purposes of the contest example, the administrator has indicated “BIFURCATE” as shown at step s89. A detailed explanation of the contestant choice control selections is outlined above.

Under a contest without contestant dispersion control, contestants may populate a choice, selection or range in excess of the calculated limit, although, as indicated by the contest example and the illustration in Table 2, the likelihood of overpopulation is highly unlikely.

FIG. 22A, step s96 determines that the creation of the contest parameters is sufficiently complete, and that contestant balloting may begin. The next point in the process diagram is FIG. 22B, at step s98.

At FIG. 22B step s98, the system initiates the contest round or rounds through the publication of the contestant interfaces (e.g. FIG. 1 at 20), and the system enables the processes illustrated in FIG. 21 at FIG. 22B step s100.

At FIG. 22B step s102, the system tests to determine if a sufficiently large number of contest entries have been received, such that newly received entries are of a sufficiently large sample size to accurately test for the presence of a real anomalous distribution relating to contestant ballot selections, which indicates likely over-population of one or more contestant selections, choices or ranges (inconsistent with the sample population or survey data); and a triggered imposition of system selection limitations. A brief overview of the process diagram in FIG. 22B will determine that Step S102 is traveled through repeatedly during each round—see logic loop at steps s103, s108 and s112 back to step s100.

In a preferred embodiment, the system and system administrator will select a contestant ballot sample size at step s102 to balance the incidence of statistical type I and type II errors. A type I error is one in which the system incorrectly rejects the null hypothesis being tested, and in the system, the null hypothesis is “is the realized contest sampling consistent with the sample population or survey data.” Under a type I error, too small a sampling can lead to false rejection of the null hypothesis and unnecessary implementation of the contestant dispersion controls in FIG. 22A step s89. Alternatively, a type II error is one in which the system fails to reject the null hypothesis, missing warning signs, and failing to implement the system choice controls (to enforce contestant dispersion) in FIG. 22A step s89. If the costs associated with a type II error are low, the contest administrator will set the sample size to be high (e.g. 200 ballots). If the costs associated with a type I error are low (i.e. contestants expect the immediately implementation of dispersion controls), the sample size will be set to be low (e.g. <50 ballots).

For the purposes of the contest example at FIG. 22B, step S 102, the system sets the level of new entries to be the first to occur of:

• • (a) 200 new entries since the last test, and
  • (b) a number of entries equal to the product of (i) 20, and (ii) the number of degrees of freedom in the test, subject to each selection, choice, or range having a realization of at least 10 ballots.

Before proceeding from FIG. 22B steps s102, at s103, the system makes a determination of whether the current round of the contest has ended (i.e. all entries for the related round have been received or time for contest entry has lapsed). If the related round has ended, the system process will proceed to FIG. 22B step s110 for processing of results, and if the round has not ended the system will proceed to FIG. 22B step s100.

Illustrated in Table 7 below is an example of tests for an anomalous distribution for the current contest example using a tested sample size of 200 new entries. The column entitled “Expected” depicts the expected distribution of 200 ballots, and the other columns indicate increasing stress scenarios. The expectation is for 32.5% of the distribution to be in the largest choice (Range 2 in the example) with the remaining columns indicating a simulated stress test of increasing concentration in Range 2 to 37.5%, 47.5%, a 51% majority, and 60% respectively. The other ranges (1, 3, and 4) are distributed in accordance with the percentages indicated in the expected distribution; in a real implementation, the realized figures for each range would be used.

As indicated in the last row of Table 7, “Dispersion Control Triggered”, choice selection control (to effect contestant choice dispersion) is only effected in the last scenario labeled “Non-Conforming 60%” where the most concentrated range population has almost doubled (32.5% to 60%) in an analysis of 200 newly received ballots. Note that the 60% indicated in Table 7 is less than the 70% indicated in Table 5 above. The 70% limitation set at inception is set as an outer bound and some embodiment implementations would strictly use the initial 70%, and other embodiment implementations would use a lower value.

This contest example will use 60% during contest rounds with the logic being: (a) the statistical distinction between 60% and 70% is almost immeasurable in that each is highly remote, and (b) the 10% buffer during active monitoring at FIG. 22B step S104 would act as a “warning track” and protect the contest results from slippage while contestant selection control is being implemented within a contest round.

	TABLE 7
		Conform-	Conform-	Conform-	Non-Con-
	Expected	ing	ing	ing	forming
	32.50%	37.50%	47.50%	51.00%	60.00%
Range 1	35	32.4	27.22	25.4	20.74
Range 2	65	75	95	102	120
Range 3	45	41.67	35	32.67	26.67
Range 4	55	50.93	42.78	39.93	32.59
Total	200	200	200	200	200
Chi-Sq		2.28	18.5	31.2	68.9
Confidence		68%	99.99%	99.9999%	100%
Alpha		10%	0.01%	0.00001%	0%
Dispersion Control	NO	NO	NO	YES
Triggered

For the purpose of the contest example, the first monitoring of realized balloting occurs after 200 ballots have been received, and for simplicity of illustration, the following analysis will only consider the initial 400 contestants rather than the total expected number of 3,000. Further, for the contest example, it is assumed that realized contestant balloting suffers from maximum concentration and that contestants are unexpectedly clairvoyant with respect to the future event. So at Step s100, all of the first 200 contestants choose Range 2, and at Step s104 the system determines that the distribution is not consistent with sample population and survey data, and that a “NO” is indicated at Step s106. Note that if “YES” results from Step s106 in this example, the system moves to Step s108, queries if contest round is complete, concludes “NO” (because at least 200 more contestant ballots are expected and time has not expired), and the process moves back to Step s100.

Continuing with FIG. 22B, and having processed a “NO” result at Step s106, the system queries the presence of active contestant choice control, and based on Step s105 in FIG. 22B, the logic result is “YES”. The active contestant choice control indicated at FIG. 22A Step s89 is “BIFURCATE”.

Moving temporarily to FIG. 24A and FIG. 24B, the comparison between the figures displays the manner in which the system will execute “BIFURCATE” active contestant choice control. FIG. 24A displays the contest range choice configuration of ranges 1 through 4, where each range contains an equal amount of the event space probability area. FIG. 24A indicates the contestant selection configuration at inception of the contest.

Immediately following the “NO” at Step s106 and “YES” at Step s107, the system employs bifurcation control on a real time basis. Where the system had earlier created contestant screens at display 20 with input area 28 displaying 4 range choices, the system will now modify the screen area 28 to display five ranges; Range 1, Range 2A, Range 2B, Range 3, and Range 4 where Range 2 has been bifurcated into two equally sized ranges. FIG. 24B illustrates the mapping of the range selections into the event space immediately following step s107; Range 2 (R2) has been bifurcated into Range 2A (R2A) and Range 2B (R2B).

Round 1 summary processing of ballots occurs at Step s110. For the purposes of the contest example, 200 contestants selected Range 2, 100 contestant selected Range 2A, and 100 contestants selected Range 2B. Based on the receipt of 400 ballots, the system will determine “YES” at Step s108.

For the purposes of the example, Range 1 (or R1) is consistent with oil prices declining materially, Range 2 (or R2) is consistent with oil prices declining slightly, Range 3 (or R3) is consistent with oil prices rising slightly, and Range 4 (or R4) is consistent with oil prices rising materially. The ranges R2A and R2B split the “slight decline” range with R2B indicating the smallest of declines, and R2A indicating a less than material decline.

At Step s110, the system compares balloting against the realization of the uncertain event for a round; for the purposes of the contest example, the event space (oil prices over the related period) declined slightly. In the contest example, 200 contestants populated Range 2, 100 contestants populated Range 2A, and 100 contestants populated Range 2B. Further, for the contest example we assume that Range 2 and Range 2A (since they are coincidental) are the successful ranges (recall that Range 2A and Range 2B are not coincidental)—all other choices would indicate a loss. As a result, 300 contestants enjoy a round 1 success, and 100 contests are attributed a round 1 loss.

Following Step s110, the system moves to Step s112, to evaluate if the contest has ended and if all rounds of balloting have been completed. A “NO” indication restarts the process at Step s100.

For the purpose of the contest example, rounds 2 and 3 proceed as follows. For each round the bifurcation control remains in effect and the system indicates 5 ranges (R1, R2A, R2B, R3, R4] on a contest display 28. Further the administrator also indicates “CLOSE” for active contestant dispersion control at FIG. 22A step s89 applicable to both rounds.

Round 2: (i) the first 200 ballots select Range 2A, (ii) step s102 indicates “YES”, (iii) step s104 determines that the realized ballots distribution is inconsistent, (iv) Step s106 indicates “NO” and (v) Range 2A is closed and precluded from further selection by contestants at step s107. The system modifies contestant screens at display 20 at input area 28 where one range (that designated as Range 2A) will be precluded from selection and contestants attempting to select Range 2A will be notified to select an alternate range within the input area 28. Continuing with the contest example, the remaining 200 ballots populate Range 2B. In the illustration,

Range 2A is realized for round 2, and 200 contestants enjoy a success in round 2—for clarity the 200 contestants succeeding in round 2 are a subset of the contests who succeeded on round 1.

Round 3 proceeds in a manner identical to round 2.

FIG. 22B, at step s112, the system will conclude that the contest has ended based on a conclusion of the final round; a “YES” will direct the system to proceed to the processing of contest results at s114.

Claims

1. An information management and synchronous communication system, comprising: a processor linked to a specialized computer system running a process of an internet or network based computerized contest over rewards or promotions whose realizations are based on successful predictions or estimations over events with uncertain outcomes; and a memory including instructions, which when executed by the processor, cause the processor to implement at least one subsystem where the computer system modifies contest parameters on a real-time basis to cause contest entries to be dispersed such that contestant entries are controlled from over-populating one or more contest selections such that contestant entries, after system processing, occupy more than one mutually exclusive selection and where the modification of the contest space is triggered by system monitored deviations between realized contestant selections and a priori statistical process estimates or a priori expected values.

2. The system of claim 1, wherein dispersion control is accomplished by either system range control or contestant control.

3. The system of claim 2, wherein dispersion control is accomplished by both system range control and contestant control.

4. The system of claim 1, wherein the processor further runs a process of identifying, selecting, and tracking contestants from a population for an internet or network based computerized contest which utilizes a combination of social graph data and survey data where the system runs statistical processes to measure the selection dispersion tendencies of one or more subset combinations of a population, and where the system constructs one or more contestant pools from one or more population subsets where the system measured statistically expected dispersion of an identified contestant pool's selections exceeds the expected dispersion of selections from both the overall population and other subsets of the population.

5. The system of claim 1, wherein dispersion control mechanically controls contestant selection dispersion with a specialized computer system running processes to select contestants based on social graph attributes and contest selection dispersion expectations where the linking system produces a material reduction in the potential aggregate contest reward amounts compared to reward outcomes without the disclosed system.

6. A process of initiating an information management and synchronous communication system, comprising a specialized computer system linked to the communication system, the computer based system running a process of an internet or network based computerized contest over rewards or promotions whose realizations are based on successful predictions or estimations over events with uncertain outcomes, the process comprising the computer system:

modifying contest parameters on a real-time basis to cause contest entries to be dispersed;

controlling contestant entries from over-populating one or more contest selections such that contestant entries, after computer system processing, occupy more than one mutually exclusive selection, and

modifying contest space that is triggered by system monitored deviations between realized contestant selections and a priori statistical process estimates or a priori expected values.

7. The process of claim 6, further comprising a process which identifies, validates, and stores selection concentration limits, for one or more of the selections presented to an initial contestant, where such contestant selection concentration limit is expressed as a percentage of total contestant count or as an absolute number of contestants where the system has applied statistical procedures for the measurement of dispersion of outcomes of contestant selections utilizing sample population or survey data in addition to an administrator selected confidence level, and where the system stores a contestant selection concentration limit for access by the system to automatically trigger system modifications of the contest space.

8. The process of claim 6, further comprising a process which actively aggregates incoming contestant selections into sequential groups for processing, where the number of contestant selections aggregated in a group is based on a sufficient sample size to control for type I and type II statistical errors relating to a compliance comparison with a preset contestant selection concentration limit or a selection dispersion distribution and where the aggregated group selection distribution is compared for consistency with preset stored and accessed concentration limits or selection distributions and where the system measures group selections using statistical tests from a collection of statistical tests consisting of (i) one or more contest selections indicate concentration in excess of a preset limit, and (ii) the distribution of selections indicated in a group indicates non-conformity with a preset distribution as measured by a parametric or non-parametric statistical test performed at a specified confidence level.

9. The process of claim 8, further comprising a process which identifies and monitors divergences between stored and accessed selection distributions and stored and accessed selection concentration limits with the contest selections received from real-time contestant groups and where a statistical test result or expected value result is a statistically significant divergence from a priori values such a divergence triggers a real-time modification to the contest space for all contestant balloting coincidental and following the imposition of the divergence trigger.

10. The process of claim 6, further comprising automated real-time system modifications to the contest space alter contest space parameters where the parameters modified are selected by the system from a group consisting of (i) selections available for choice by a contestant, and (ii) the selection submitted by a contestant.

11. The process of claim 6, further comprising contestant balloting the contestant balloting precedes the event over which balloting is conducted, using real-time monitoring of contestant balloting by using active system comparison of the a priori system stored and accessed expected distribution of selection choices with the realized selection choices, and where the system automatically and actively modifies the available selection choices for an incremental contestant entry, the system will modify the available selection choices based on preset instructions from an administrator.

12. The process of claim 11, wherein the system contest process is controlled with respect to automatic active modification of the selection choices with respect to a group of preset instructions programmed by the administrator and automatically executed at system determined times where the group of non-mutually exclusive instructions consists of: (i) closing a previously available selection, (ii) splitting or bifurcating a previously presented single selection, (iii) adjusting the boundary between selections, (iv) add a selection not previously presented, and (iv) adding a round of balloting.

13. The process of claim 6, further comprising contestant balloting the contestant balloting is concurrent with the event over which balloting is conducted and where contestant entries are queued by the system as received, using real-time monitoring of contestant balloting by using active system comparison of the a priori expected distribution of selection choices with the realized selection choices and where the system automatically and actively modifies the raw selection made by a contestant by changing the value of the submitted selection to an alternate value to increase the dispersion of system processed contestant entries where the effect of the system modification on a modified contestant selection is to shift a raw selection to an adjacent value or time where the adjusted value or time is less represented in previously received contestant selections.

14. The process of claim 13, further comprising a process which modifies raw contestant selections based on insufficient dispersion with respect to two or more contestant selections and where modifications to a raw contestant selection will retain the original queue sequence of unprocessed contestant submissions.

15. The process of claim 13, wherein contestant balloting is coincidental with the uncertain event over a contest of one or more rounds where contestants ballot through handheld devices or desktop devices and where the uncertain event is based on a group consisting of

i. the level of a traded market where the indication of success for a contestant is the identification of or more a priori unknown outcomes of such traded market from a group consisting of (i) the high-point of the traded market and the coincidental time thereof, (ii) the low point of the traded market and the coincidental time thereof, and (iii) an inflection point in direction of the traded market and the coincidental time thereof where in all instances time is based on the system's internal clock and the system measured time of a contestant entry where such time is raw or system modified.

ii. the points in a sporting contest where the indication of success for a contestant is the identification of one or more a priori unknown outcomes from a group consisting of (i) the aggregate points for one or more participants in the sporting contest, (ii) the differential in points between two or more participants in the sporting contest, where in all instances a ballot processed by the system and subject to system modification is the combination of the contestant's point indication and the time the system received such indication as measured by the system's internal clock.

16. The process of claim 6, further comprising validating the efficacy of the system modifications to the contest space and the effect the modifications have had on contestant selection dispersion, and system modifications to the contest space have been ineffective in controlling contestant choice dispersion; applying a pari-mutuel adjustment to successful contestant entries, where a system calculated factor between zero and one will be applied to contest rewards attributed to successful contestant entries.

17. A process for initiating an information management and synchronous communication system comprising, a specialized computer system linked to the communication system and running a process of identifying, selecting, and tracking contestants from a population for an internet or network based computerized contest which utilizes a combination of social graph data and survey data where the system runs statistical processes to measure the selection dispersion tendencies of one or more subset combinations of a population, and the system constructs one or more contestant pools from one or more population subsets; the system measured statistically expected dispersion of an identified contestant pool's selections exceeds the expected dispersion of selections from both the overall population and other subsets of the population.

18. The process of claim 17, further comprising identifying, selecting, and tracking contestants from a population for an internet or network based computerize contest collects social graph data from prospective and actual contestants by direct entry or by an opting in process or a linking to internet sites or networks containing the contestants social graph data where the social graph data is used by the system to construct the contest space.

19. The process of claim 17, further comprising supplying by prospective and actual contestant social graph data or otherwise facilitates access to social graph data as full or partial consideration for participation and standing in the contest.

20. The process of claim 17, further comprising drawing real-time aggregation of the social graph data of a contestant pool from a target population, the real-time identification of one or more common social graph linkages shared by a subset of the contestant pool, and the subsequent process of a web crawl using a social graph programming interface and utilizing the identified social graph commonalities for the purpose of increasing the number of contestants.

21. The process of claim 17, further comprising real time aggregation of reality mining data in the construction of a contestant pool where one or more patterns in individual behavior or group activity is indicated as correlated with a priori predicted selections within a contest space and where a process, using reality mining data, is used to select contestants from a population such that a desired level or selection dispersion can be expected and where contestant predictive entries are likely to be dispersed among two or more mutually exclusive choices.

22. The process of claim 6 further comprising having context of a contest over a continuous variable event space, and running an analytic or numerical process based on binomial or poisson statistical distributions where the system will apply an administrator identified confidence level and where system will identify a boundary of maximum expected concentration of selection for a particular time or value within the contest space and where such boundary of maximum expected concentration will be used in an automated system contest space modification trigger.

23. The process of claim 6 further comprising having the context of a contest over a discrete variable event space, and running an analytic or numerical process based on parametric or non-parametric statistical distributions where the system will apply an administrator identified confidence level and where the system will identify a boundary of maximum expected concentration of selection for one or more selections within the contest space and where such boundary of maximum expected concentration will be used in an automated system contest space modification trigger.

24. The process of claim 17, further comprising creating an internet contest on a specialized computer system; the system utilizes sample population estimates, survey data, aggregate social graph data and reality mining data prior to a contest to measure and optimize contestant pools such that contestant pools are constructed from a subset of a larger target audience in a systematized manner to control the expected dispersion of contestant predictive choices during contest game play where a larger contestant pool is expected to exhibit a dispersion distribution more consistent with a priori system selection expectations.

25. The process of claim 6, further comprising running the system which supplements initial sample population and survey data with the statistical measures from realized contestant selection balloting for revisions to selection or choice concentration limits and contest range or choice probabilistic positioning in subsequent rounds of the contest.

26. A non-transitory computer readable medium storing computer-executable program instructions which when executed by a processor, perform a process of initiating an information management and synchronous communication system linked to a specialized computer system, the computer based system running a process of an internet or network based computerized contest over rewards or promotions whose realizations are based on successful predictions or estimations over events with uncertain outcomes, comprising the steps implemented by the computer based system of:

modifying contest parameters on a real-time basis to cause contest entries to be dispersed;

controlling contestant entries from over-populating one or more contest selections such that contestant entries, after computer system processing, occupy more than one mutually exclusive selection, and

modifying contest space that is triggered by system monitored deviations between realized contestant selections and a priori statistical process estimates or a priori expected values.