Expected Value Question

**shantystar** · 10-26-09, 03:54 AM

find any mathematicina!!

**ferrari355spider** · 10-26-09, 09:42 PM

Huh? lol

**Peeig** · 10-26-09, 09:44 PM

I think the answer is either Infinity or empty set

**suicidekings** · 10-27-09, 03:48 PM

Originally posted by ferrari355spider

This might be a stupid question as I don't know how capping is done but I am just wondering: If you were to bet let's say your best 1500 games a year in the top leagues, how many bets would be 60%, 57%, 55% or 53% roughly?

Not enough information. I assume you mean what are the probabilities of winning 60/57/55/53% of wagers, but you need to know what the bettor's expected hit rate is to do it. Random selection would assume a 50% expected hit rate, but presumably this bettor has defined what his edge is, and it's higher than that. After that, it's just a Binomial distribution, which you can do quickly in Excel.

**ferrari355spider** · 10-27-09, 03:58 PM

Ok I obviously never capped anything before but let me try to re-explain. I have heard ppl say this play has 54% chance of hitting. So what I am asking is if a legitimate capper plays 1500 of his best plays a year, how many would be 60% plays, 55% plays, 53% plays? Is this a valid question or am I totally out of the loop?

**suicidekings** · 10-27-09, 04:45 PM

Umm... depends on his analysis of the games, his betting strategy, and other variables. I can't really give a meaningful answer to that question. A more meaningful stat for a bettor to keep track of is having an idea of what his edge is when betting sports (ie: what percentage will he win over the long run). This gives him an idea of how successful he can be over time.

If you place all your bets at standard vig (-110), you need to hit 52.38% of wagers to break even. (100/210 = 47.62 -> 100% - 47.62% = 52.38%) You can look back into the Thinktank forum to see much more comprehensive of this math to explain it. Look for posts from Ganchrow.

That being said, any bettor that can win better than 52.38% of his wagers over time (in your example, that's winning 768 of the 1500, or 72 more wins than losses) at standard vig will be making money over time. Every time a bettor pays more than the -110, he needs to win at a higher rate, and wagering at lower vig (plus money dogs) means he can hit at a lower rate and still make money over the long term.

Required win percentages:
Faves: W% = Bet amount / (Bet amount + amount to win)
Dogs: W% = Amount to win / (Bet amount + amount to win)

ie: 1000 bets made at an average of (-120) requires a 54.55% success rate to be profitable, while 1000 bets made at an average rate of (+108) would only require a strike rate of 48.08%. I say average because the price fluctuation makes a big difference as the market adjusts to bets coming in on both sides, so you're very unlikely to always be playing at standard vig.

This all assumes flat betting (the same amount wagered on each play), but the bet sizes would probably vary depending on the bettor's confidence in each play (but again, that depends on a lot of things and varies for each game).

I hope that helps set you in the right direction.

**ferrari355spider** · 10-27-09, 05:04 PM

Thanks for the reply. What I don't understand is how a legitimate capper determines if a play is a 60% play or a 53% play, is there an easy answer to this? or is it more complicated then I am thinking.

**MrX** · 10-27-09, 05:08 PM

Anyone trying to maximize profits will have a lot more plays at the lower end of the +EV spectrum (<55% in your example) than the higher end, simply because there are a lot more opportunities in that range. The linesmakers and the market just don't miss the mark by a ton very often in major sports.

**suicidekings** · 10-27-09, 05:12 PM

Depending on the capper's technique, he'll be developing stats from game data to generate probabilities for a particular team winning in a particular matchup. The most common methods are not too complicated, making use of the correlation between expected win percentage and the number of points/goals/runs a team scores and allows in a game on average. Look up Pythagorean expectation as it relates to sports. There are a lot of websites that can give you a crash course in it.

You need a little bit of stats knowledge and average Excel skills to generate your own probabilities, provided you have good data. Again, you can read back in the Think Tank to find a few threads that go into creating models.

**ferrari355spider** · 10-27-09, 05:14 PM

Ok so to answer my question what you guys think would fall over those %s over 1500 games, for example, how many out of 1500 would be 60% or better plays?