Kelly criterion

[valueman]

Does anyone have a link for a free download of the kelly criterion staking plan for windows XP?

The Kelly Criterion is a dream, rather than a reality, system. Most big-time bettors never risk more than 5% of their bankroll on plays, and most risk 1% to 3%.

[ParanoidAndroid]

Originally posted by: reno
The Kelly Criterion is a dream, rather than a reality, system. Most big-time bettors never risk more than 5% of their bankroll on plays, and most risk 1% to 3%.

The irony of your post is that 1-3% is EXACTLY what the Kelly Criterion recommends for sports betting for sharp players [img]i/expressions/face-icon-small-wink.gif[/img]

At the most basic level, the Kelly Criterion states that you should never bet more than your edge at any given game. For most sharp players (the guys who can legitimately beat the vig with a min of a 52.5% winning average) should be betting their advantage, or about 2%.

I wouldn't be so quick to dismiss the Kelly Criterion. There is a TON of legitimate math behind it and there are very few professional gamblers (sports & BJ primarily) that don't incorporate the Kelly Criterion in one way or another.

[valueman]

hey - i have made over $30K for the last 9 years using it so don't knock it

[SlipperyPete]

I think there are three main reasons why many people fail when using Kelly.
One is that they cant assess their real edge. Everyone is thinking of themselves as a 55% percent handicapper right off the bat. But in reality that figure should be based on past results (large samlple SHOULD be required) not on future expectation
Second what most people dont do, is that when they lose they should bet less. Usually the opposite is true. Percentage of bankroll bets should be recalculated almost daily.
Third, there should be a limit on how much of your bankroll they should at risk in a session. In my opinion, not counting seasoned pros, that figure should never exceed 20% percent

[So Cal Baseball]

Kelly Criterion is great for -110 lays AND if you have a real grip on what your bankroll actually is. If you are someone who knows without a doubt that your bankroll for football season is 10 dimes (no more no less) you can use it quite easily. However, if you are like most gamblers you know good and well that either you will come up with more money to play with during the season or actually have to borrow against your bankroll for one reason or another.

Find a guy who has that kind of discipline I will show you a guy who could be using his money more wisely in either real estate or the stock market. The Kelly Criterion doesn't allow for fluctuating when you actually do have an inside edge on a game or games do to information, and it is inside plays banged a little harder that make the people who actually do make money betting games successful (and there aren't many).

The KC is also mathematically a little whacked when it comes to ML plays where either you are laying big chalk or betting large dogs.

[count zero]

Re: SmallDaddy's point about how much to risk in a session, the only published material I know of that addresses this is a paper by blackjack mathematician Peter Griffin on making simultaneous bets with Kelly. He deals specifically with NFL betting. It's available in his book Extra Stuff, and probably available on the net as part of the proceedings of the University Of Nevada Reno's Conference On Gambling

[count zero]

Re: SmallDaddy's point about how much to risk in a session, the only published material I know of that addresses this is a paper by blackjack mathematician Peter Griffin on making simultaneous bets with Kelly. He deals specifically with NFL betting. It's available in his book Extra Stuff, and probably available on the net as part of the proceedings of the University Of Nevada Reno's Conference On Gambling

[Iowa Child]

Great points, Socal Baseball, that's precisely why I don't use it. It's too rigid when there is truly inside info on a game, a bad line at a certain book, etc., etc.

[ParanoidAndroid]

Originally posted by: So Cal Baseball
Kelly Criterion is great for -110 lays AND if you have a real grip on what your bankroll actually is. If you are someone who knows without a doubt that your bankroll for football season is 10 dimes (no more no less) you can use it quite easily. However, if you are like most gamblers you know good and well that either you will come up with more money to play with during the season or actually have to borrow against your bankroll for one reason or another.

Find a guy who has that kind of discipline I will show you a guy who could be using his money more wisely in either real estate or the stock market. The Kelly Criterion doesn't allow for fluctuating when you actually do have an inside edge on a game or games do to information, and it is inside plays banged a little harder that make the people who actually do make money betting games successful (and there aren't many).

The KC is also mathematically a little whacked when it comes to ML plays where either you are laying big chalk or betting large dogs.

Let me start by saying that your post indicates that you only have half-assed and mal-informed knowledge of the Kelly Criterion. In other words, you're talking out of your ass.

KC absolutley DOES allow for fluctuations in wagers. Again, you have to address each situation individually - each wager (and your bankroll at that time) will dictate the amount that you should wager AT THAT TIME. You could have a "betting bankroll" (not just sports, but overall) and know that certain wagers (ie sports) dictate an X wager, while some table games that you play with an advantage (ie blackjack) dictate another betting pattern. At the end of the day, or long sessions, you tally your new bankroll and work from there the next day.

As for laying big chalk - you have to be able to calculate the house edge when you're laying everything from -110 to -800 or even higher. That house edge, and then your edge beyond that should give you the numbers you need to work with KC.

Again, I dont think that there are ANY serious, long-term pros in any form of gambling (with poker as an exception) that doesnt use some form of KC proportional betting in their daily activities.

The KC takes WORK to both understand and use it - you've done neither.

[So Cal Baseball]

PA
I do understand the criterion. You say I'm talking out of my ass for not knowing that criterion allows for changes. Where the hell does it say anywhere in the KC to double or even triple your bet if you found out from a source say close to a college basketball team that 2 or 3 of the guys that start and contibute to a team won't be playing because of team violations, well before the books get it (just an example)? These situations don't occur that often, as a matter of fact very rarely, but they DO occur.

The criterion fluctuates due to your overall bankroll at the end of each session, when your winning percentage goes up you bet more, when it's down you bet less, etc. etc. etc. I know all this. What would make you think I don't? I also know enough about it to say that the few that do make money playing sports (I personally believe and know that there are a few) take advantage of these situations (inside info) fully and the criterion simply doesn't work for them. IMO the criterion is there only for the straight and narrow and doesn't compensate for people who are "in the know". I know that not very many people out there get inside info to make "gambling" work, but there are people out there that make money based on who they know. They get info from people that wasn't given to the BMs and it makes all the difference. If you think that "all" pros use the KC or "some form" of it then all I can do is laugh. I'm not saying there aren't pros that do. The guys I know of personally that make coin betting sports just don't use it (I've asked).

The criterion (as it applies to sports betting) works just fine for those who are recreational gamblers and don't have too many expectations one way or another of making too much money gambling. It also helps if you don't make too many plays per session. I used it exclusively myself when I started betting on sports regularly years ago. In the end through extensive research I found that it wasn't practical when say 3 or 4 times a year I would get info on games that was huge. The criterion didn't allow for me to take advantage of situations that ended up being sure things. To bet too much simply was to go against the criterion. To quote an old saying "either your pregnant or your not... you are never half pregnant" (ie. you either use the criterion or you don't). If you use it "sometimes" you are truly not following it.

When it comes to card counting in BJ I'll give KC a little more credit. It is much more practical when used in games where humans can't help determine the outcome. I have no opinion on it as it applies to poker, I just haven't researched it.

Also, I know about the mathematical adjustments made when applying them to big dogs/big favs. I simply don't think they are practical and in my personal opinon it is the basic flaw in the system. Again, it is just my opinion and I can debate it with you for hours. It would simply come down to one opinion versus another.

I truly wonder what is coming out of your ass if you believe somewhere/anywhere in that system that it tells you to bet say 20 or even 30 percent (or even more) of your bankroll if you have information that is so solid you can't refute it (i.e. of the inside variety). I've read several books discussing the matter and I've never seen it anywhere. PLEASE point it out to me.

Either you are a BM (which is fine) or you lose gambling. Doesn't mean I have to because I know enough to bet when I get info that is not refutable.

My programmer-partner and I are both quite familiar with Kelly, with each of us having been introduced to it through Huey Mahl's '80s articles in Frontline and High Roller. The big problem with Kelly as it pertains to sportsbetting is, of course, how to deal with multiple plays in a short period of time. There is no simple solution, and Huey Mahl constantly presented strategies--such as round-robin parlays and 1/2 Kelly percentages--to help players deal with this problem. Anyone who bets college football and/or baskets can easily find himself with 20 or more plays in a given day, and pure Kelly doesn't make sense in such a situation because you could have more than 100% of your bankroll at risk on a single Saturday.

What my partner did relative to our program's plays through the course of our database was to run bankroll simulations, betting various percentages of bankrolls to see what transpired. After considering the results, we've opted to bet only 2% to 3% of our bankroll on each play. What many people don't grasp is the downside of Kelly--what happens when you don't hit your predicted (and predicated upon) win percentage. For example, if you think you've got a 60% system and bet, as Kelly prescribes, 16% of bankroll on each play, if you hit, say, 55% winners, you could end up breaking even or perhaps even losing money. And if you hit, say 52.38%, the results, which would be break-even flat-betting, would be downright disastrous. We've decided to go the conservative route, knowing that if we do hit 60% on a high volume of plays, we'll still, rather quickly, make millions of dollars betting only 2.5% of our bankroll.

[So Cal Baseball]

[i]What my partner did relative to our program's plays through the course of our database was to run bankroll simulations, betting various percentages of bankrolls to see what transpired. After considering the results, we've opted to bet only 2% to 3% of our bankroll on each play. What many people don't grasp is the downside of Kelly--what happens when you don't hit your predicted (and predicated upon) win percentage. For example, if you think you've got a 60% system and bet, as Kelly prescribes, 16% of bankroll on each play, if you hit, say, 55% winners, you could end up breaking even or perhaps even losing money. And if you hit, say 52.38%, the results, which would be break-even flat-betting, would be downright disastrous. We've decided to go the conservative route, knowing that if we do hit 60% on a high volume of plays, we'll still, rather quickly, make millions of dollars betting only 2.5% of our bankroll.

Reno, I agree with this totally and it has also been one of my biggest detractions to the system.

What was surprising to both my partner and me regarding our testing on baseball sides, a moneyline wager, is how dramatically profits improved by betting proportionally less on dogs. For example, if we bet, say, a -200 fav for two dimes, we'd risk 4000 to win 2000, but if we were playing the dog at +200, we'd risk 1 dime to win 2 dimes. My partner had read about this before, but dismissed it until he tested it and saw how huge of a difference it made. I don't really understand the rationale behind it, and I haven't spent any time trying to understsand it. I've simply accepted my partner's findings.

[So Cal Baseball]

Reno,
Precisely my point and a very valid one. The system just doesn't hold up when it comes to adjusting big dogs/big favs. I researched this several years ago and came to the same conclusion.

[valueman]

it is great for the horses - especially fixed odds as in the UK