Use Win & Place stats Posted on January 12, 2018 at 05:13:06 PM by StraymarJA
Good message Tiger.
Due to the small sample size, I now tend to use Win & Place combined to get a Quinella percentage to use in a computer simulation. My software is set-up for GP-W-P-S records for front and back player (if doubles), however, have found over time that using show records too can make the simulation output less reliable.
If a player is average, enter 1000 GP, 125 Wins, 0 Place, 0 Show. The difficulty lies in when the early game players are in late games, vice versa, and there is limited data on that player against players of that level. Then I need to estimate the input to use for that particular player in a game. Usually that involves taking their actual record (singles or doubles record) and factoring it up or down by up to 20% max.
i.e. if Arta was playing in late game and would have used 1000 GP, 100 wins, then using 1000 GP, 80 wins.
I wish my software was Excel based as it would make running my ancient simulation so much easier. I still use old MS-dos based programs written in Turbo-Pascal that are automatically sent to the printer (no option to just see output results on a screen). I hang onto a couple of old computers I have that still have DOS capability and fire one of them up when I do use my old software (only use on Jai-5 these days). This simulation output does give me WPS %'s, Quin/Exacta %'s and ranks them from easiest to hardest, along with trifecta and superfecta (sup if 9-pt ga) # wins and rankz those too from easiest to hardest for the number of games simulated.
Would be interested in hearing capabilities of software other Chalkies use.
My first Jai-Alai simulation was on a Commodore 64 (5 1/4 disk) in fall of 1986 for simulating the superfecta game. Since it was so slow, it took 1-hour to simulate 1000 9-point games. My brother was programmer and I got him to write the code to where you would run it for X games and then save the results to disk, then next time you powered it up and ran it for more games, it added to the saved previous # of wins for each super combo. After 200 hours, I had 200,000 games of simulation stats (assuming all teams equal).
Really crushed it in the early days of superfecta wagering as nobody else in south Florida was doing this. Made a killing at WPB, Dania, Tampa and profited most weeks. Should have done Miami too.....but I think Julio had that covered already. Only had to play the same $1 based superfecta list each super game. Advance it and go home if you like.....the money was so easy then. Pay a runner $250 a week to go to a fronton with the list, give list to teller to punch (tip teller too of course).
The infamous Randy Gallo approached me about being superfecta partners in the early days. He did not have any simulation and wanted in on my superfecta bets. Told me if I didn't take him on as partner, it wouldn't be long and he would get his own programmer/simulation stats and play against me. Thus, was partners with him on that..... Replies: