Forecasting contest

Details Forecasting contest Forecasting contest

1999-09-08

2003-08-12

Follansbee, John (Department: 2121)

Data processing: artificial intelligence

Knowledge processing system

Reexamination Certificate

active

06606615

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally concerns contests involving the prediction of future values of certain variables whose values change over time, and is particularly applicable to contests involving forecasting of financial and/or economic variables.
2. Description of the Related Art
Forecasting Contests
A number of forecasting contests have been conducted in the past. Such contests range from various wagering events, such as Superbowl pools, to various financial forecasting contests. Typically, such conventional contests seek to identify the best predictor for the outcome of a single event. For example, the website at www.investorsforecast.com allows participants to predict where the Dow Jones Industrial Average (DJIA) will be and what the prices of certain stocks will be at the end of next week. The person submitting the most accurate prediction for the DJIA and the person submitting the most accurate prediction for an individual stock are each given a fixed monetary award, such as $300. Other contests in the financial arena typically allow participants to invest an imaginary amount of money, with the winner being the person whose portfolio is the largest at the end of the contest. One example of such a contest can be seen at www.fantasystockmarket.com.
However, the present inventors have discovered that such conventional contests are inadequate in the following respects. First, the rankings generated by such contests typically do not provide useful information for truly identifying the best forecasters. This is a particularly significant shortcoming with respect to financial and economic forecasting, in which it is very useful for third parties to have that information. In addition, these conventional contests often reward short-term or single-event thinking, and such qualities may not be the most desirable in many cases. Finally, partly because of such short-term and single-event thinking, partly because of the specific events for which predictions are solicited in such conventional contests, and partly because of the manner in which such conventional contests are typically structured, the utility of the data produced by such conventional contests for purposes such as combination forecasting often is sub-optimal.
In the financial and economic arenas, the result is that traditionally there has been insufficient data upon which investors could rely in order to select investment advisors. As a result, many investors are left to select advisors based largely on arbitrary criteria or, in the best case, to rely on recommendations from friends. At the same time, many actual and potential investment advisors who are very capable at reading the market conventionally have had very little opportunity to demonstrate their expertise to the public, and thereby attract new clients. Similar concerns exist for other financial and economic experts who wish to demonstrate their expertise or the validity of their prediction techniques.
What is needed therefore, is a contest in which the rankings and/or rewards are tied more closely to the forecasting characteristics that are most desirable and that yields a large database of information which can serve as the basis for comparing the predictions of different forecasters. It is also desirable that the contest provide data that are statistically significant and can provide the basis for a wide variety of combination forecasts and other statistical analyses as well as being highly useful for marketing purposes.
Prediction Input
In conventional forecasting contests, participants typically submit their predictions by writing, typing or speaking their predictions. Most frequently, such predictions consist of a numerical estimate of what the value of the predicted variable will be at a specified point in time. Thus, for instance, in the www.investorsforecast.com website contest mentioned above, participants type in the values of their estimates and then submit those estimates by clicking a button on the website.
While such prediction submission techniques are adequate for their intended purpose, they suffer from many shortcomings. The following examples of such shortcomings have been identified by the present inventors.
First, such conventional prediction submission techniques frequently are not very intuitive from the participant's point of view. In particular, they often require the participants to digest a significant amount of information in order to translate their rough feelings about the way the prediction variable is likely to move into a hard number. This is a significant disadvantage for those participants who are very intuitive oriented. Moreover, to the extent such persons are prone to errors in processing such data when converting their rough perceptions into a hard number, their submitted predictions may vary from what they actually believe about the subject variable.
Second, having to enter numerical estimates for each prediction variable can be cumbersome and time-consuming. This may have the effect of limiting the number of variables for which participants are willing to submit predictions.
While other prediction submission techniques have been utilized, they typically have had very limited applicability. For example, the website at www.cyberskipper.com permits participants to compete in predicting certain sports-related events. One of the prediction submission techniques utilized by this site is to display a grid of possible events. The participants can then click on a cell within the grid to designate their prediction that a particular event will occur. Thus, a different grid is displayed for each baseball game, with each row of the grid corresponding to a different baseball player and each column corresponding to a different event (e.g., “runs”, “hits”, home run”). If a participant believes that a certain player will get a home run in a game, he simply clicks on the appropriate cell to enter that prediction. As can be readily appreciated, this technique generally is limited to predicting binary events (i.e., will/will-not occur). In many cases, this deficiency will limit the applicability of such techniques to collection of very coarse predictions.
What is needed, therefore, is a more efficient and intuitive way to enter or submit prediction data that is applicable across a wide range of prediction events and that can permit participants to submit predictions with more specificity than has been available with conventional techniques.
Provision of On-Line Resources
Use of the Internet has become more and more common over the past few years. Similarly, the number of websites on the Internet has grown exponentially and is expected to continue to grow at a fast pace. As a result, the amount of information available on the Internet can be staggering. However, there is often little done to insure that the information provided to end users is the most relevant to those users.
A typical website might contain advertising, as well as a certain amount of content. Both types of information are typically controlled exclusively by the owner of the website, possibly based loosely on some indications as to what visitors would like to see, or based on what advertisers might believe will be most effective. However, the present inventors question how good such strategies are at actually providing website visitors with the information that they actually want and, in any event, have concluded that the effectiveness of such conventional strategies must necessarily vary based on the website owner's individual skill in gauging his audiences desires.
Accordingly, the present inventors have discovered that what is needed is a more systematic technique for providing appropriate resources to users over an electronic network, such as the Internet, that more accurately reflects the users' desires.
Financial and Economic Forecasting
The American economy is made up of the simultaneous activities of hundreds of millions of participants, simultaneously buying and selling goods and services in the

Affiliated with

Also associated with

Rating

Forecasting contest does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Forecasting contest, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Forecasting contest will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3114795