Constructing Of The Toto Sports Lottery Initial Odds Model Based On Discriminant Analysis

ToTo sports lottery ushered in an important development opportunity in the context ofthe steady and rapid development of economic and social and lottery market growingmaturity in present China. As the core competitiveness products of the sports lottery,the methods of playing ToTo sports lottery in China is constantly being innovative tomeet public demand, then the “Odds” methods is also slightly to taking shape. Thekey prerequisite for the implementation of odds methods is to be able to effectivelycontrol the payment risk of ToTo Lottery, so to reasonably estimate the probability ofresults occurrence of the sports game for calculating the corresponding initial odds isthe core technology of the odds methods. However the calculation methods of oddsalmost have not yet begun to study from theory to practice in China, this greatlyaffected the development process of the TOTO sports lottery. Using the methods ofcombination of abstract and concrete analysis, qualitative and quantitative analysis,macro and micro analysis and theory research and empirical test, this papersystematically and completely discusses the subject of prediction of the athleticfootball match results and calculation of the odds through analysis of a large numberof academic literatures and consulting experts in related fields. On this basis, thepaper built a relatively rational athletic football match results occurrence probabilityestimation model on category discriminant analysis, thus deduce the conversionformula of the probability of football match results occurrence to the odds of ToTosports lottery. At the same time, through empirical test to the2012-2013PremierLeague’s4to21round, a total of168games, this paper has verified the feasibility ofthe approach to deduce ToTo sports lottery initial odds model.This paper is divided into four parts. Part I summarized and screened the factors thataffect the results of athletic football matches in accordance with literature reviewingand interviews to build the athletic football match results impact factors index system;Part II developed the index operationalization scheme, the index value strategy andthe data standardization rules then converted the football match results impact factorsindex to the football match results prediction index; Part III is to build the athleticfootball match results occurrence probability estimation model based on the footballmatch results prediction index, and counted the data collected in the2012-2013Premier League by category discriminant analysis in SPSS19.0to test the effect of themodel. Part IV is aims at deducing the conversion formula from the match resultsoccurrence probability to the initial odds in by the reverse explore to average oddsdata in the empirical samples opened by more than200sports betting agenciesworldwide, then analog the odds opening process by virtual betting, calculate profitand loss according to the real game sample results to evaluation of the effect of theinitial odds model.Athletic football match results occurrence probability estimation model contains threecategorical variables (Home team wins, Draw, Visiting team wins) and twodimensions of a total of12explanatory variables. Result have shown that the amount of information included in each football match results prediction index finally selectedhaving the difference to judge and interpret the final result of the match, and thecontribution to the prediction accuracy are different between each index. According tothe results of the stepwise discriminant analysis, the discriminant effectiveness of theindex “the points difference between home team and visiting team” and the index “thedifference of top18players’ total PPI values between home team and visiting team” isbetter than other index, and they were able to accurate predict about61.3%of theresults of the match in empirical sample. Therefore, this paper thought it is feasible toestimate the probability of athletic football match results occurrence by categorydiscriminant analysis model. Ultimately, this paper proposed the formula to calculatethe initial odds based on the above athletic football match results occurrenceprobability estimation model, and the profitability of the analog odds opening hasverified the practicality of the ToTo sports lottery initial odds model.