NFL Forum

This is terribly wrong

Posted By: Kim Lee
Date: 23 Aug 01, 7:30 pm

In Response To: My approach should work for you (Stanford Wong)

"use your own past performance to estimate your edge."

This fundamental error upset me most about your approach to sports. It does a great disservice to your readers. Past performance is not statistically representative of future results.

Realistically most sports bettors have less than 52.4% probability. Otherwise sportsbooks would hold less than their theoretical and historical 4-5%. Heck, some squares probably have a 49% probability. But the standard error of your estimate is roughly .5 divided by the square-root of the number of games. If you hit 53% on 100 weak picks and 55% on 100 strong picks then it is quite possible your picks have equal strength of only 50%. You can easily fool yourself. Perhaps you will collect multiple records for different sports. Meanwhile if you hit 46% then you will change methods and start a new statistical sample. Admit it Stanford, this is exactly what happened to you. You data-mined to find some angles in past data and they were so disastrous in your out-of-sample betting you switched methods. If you really believed your data mining then after finding an angle that was 100-0 you would keep betting it after it deteriorated to 100-10, even though it cost you 10 bets. You subsequently tried to find long-term winners in an internet contest, but found little consistency as the weeks progressed.

If you couldn't find consistency in others then why do you advocate statistics to detect or measure an advantage?
Would you advocate this approach to measure your advantage at blackjack? For example, should I use 100 hours of blackjack to measure my advantage or calculate my optimal bet at different truecounts?
Do you advocate this as an approach to clocking roulette? In other words should I record roulette numbers and then bet on ones that occur with statistically unlikely frequency?
How are these cases any different from sports handicapping? The results all involve uncertain measurement in a game that is very rough to beat.

Messages In This Thread

The Book: Sharp Sports Betting -- Joseph P -- 23 Aug 01, 11:01 am

Short answer -- Don Schlesinger -- 23 Aug 01, 12:43 pm

over-evaluation -- rainmandon -- 23 Aug 01, 1:12 pm

Can't agree with that -- John May -- 23 Aug 01, 4:27 pm

John is dead on. -- Fezzik -- 24 Aug 01, 11:59 pm

Wrong -- StevieY -- 25 Aug 01, 11:42 am

For newcomers to this page, . . . -- Don Schlesinger -- 25 Aug 01, 11:50 am

I'm sorry but -- John May -- 26 Aug 01, 1:45 am

I'm sorry but... -- StevieY -- 26 Aug 01, 10:32 am

The human element -- John May -- 26 Aug 01, 2:34 pm

the crux, IMO -- Colin Caster -- 26 Aug 01, 4:57 pm

More nonsense -- Don Schlesinger -- 27 Aug 01, 8:14 am

A proof -- John May -- 28 Aug 01, 11:10 am

Grown-ups???? -- Don Schlesinger -- 28 Aug 01, 11:49 am

ok ok -- fezzik -- 28 Aug 01, 12:00 am

My approach should work for you -- Stanford Wong -- 23 Aug 01, 5:38 pm

This is terribly wrong -- Kim Lee -- 23 Aug 01, 7:30 pm

Kim Lee, I recognize you from bj21 -- Compador -- 24 Aug 01, 12:40 am

Why sports betting is brutal -- Kim Lee -- 24 Aug 01, 8:10 am

re: poker -- docriver -- 24 Aug 01, 8:30 am

Blackjack, poker, sports -- Kim Lee -- 24 Aug 01, 8:51 am

bj -- docriver -- 24 Aug 01, 9:58 am

Playing level -- Kim Lee -- 24 Aug 01, 12:50 pm

Question to Kim Lee from a novice?? -- big nguyen -- 24 Aug 01, 12:01 pm

Reducing Vigorish and Scalping -- Kim Lee -- 24 Aug 01, 1:17 pm

Your well on your way to winning -- Fezzik -- 24 Aug 01, 11:53 pm

Further explanation requested. -- samo666 -- 25 Aug 01, 2:08 am

1st half betting -- fezzik -- 26 Aug 01, 10:28 pm

NFL Forum is maintained by Pi Yee Press