Purpose#

Umameshi.com Horse Racing Winning Method Refer to the article *** Program the Horse Racing Winning Method * ** **
Optimal solution (** pulp **) with the program I created before the race ( github code ) Use)

Preparation

python --September 20, 2013 Data on winning odds and single horse odds for one race of Nagoya horse racing ――This article will be long, so I will explain the result and what I did without posting the code. (For details, see github code )

data#

--The data for wins and horses are as follows. --index is horse number --columns are win odds and single horse odds (eg 1 is horse number 1 single horse odds)

Comparison

--Output the difference between horse single synthetic odds and win odds ――From here, choose the horse number with the small winning odds and the large difference.

Model selection and results

--From the above results, I thought it would be better to use ** horse numbers '1' and '3' **.

model_0## --In model_0, Mamono '3' was used as an argument. -** Minimize the number of bets (100 yen per bet) and set the conditions so that the difference (yen) between the refund (yen) and the total number of bets * 100 (yen) will be positive regardless of which horse wins. did** ――The difference is negative because the condition is met, but the number of bets has been rounded up at the end, so the value has changed. --Solved indicates that this optimization problem has been solved ―― 48.0 shows the total number of bets

model_1## --In model_1, we used single horse '1' and single horse '3' as arguments. -** The condition was set so that the total refund amount is maximized and the difference (yen) between the refund amount (yen) and the total number of bets * 100 (yen) is positive regardless of which horse wins. Also, put the maximum number of bets in the argument and set a condition so that it does not exceed that. (1000 in this article) ** ――In this case, the betting method is concentrated on the place with the highest odds so that the horse with the lowest odds does not lose much.

Impression

Since the data solved this time (September 20, 2013 Nagoya Horse Racing 1 Race) was used, this winning method seems to be usable, but there are many cases where it cannot actually be solved.
It cannot be said that you will never lose because there are actually negative points (because of rounding up), but if you do *** fully automate *** (such as buying a horse race ticket automatically), it will be regular. Money is likely to come in. (Not yet implemented)
In github code , if you mess with info before the race, you can scrape the data of wins and horses. I have created modules that can be taken, so if you like this method, please use them to create a fully automated system.