[TheCalculator]

Would love to hear your money management strategy.

Do you use full kelly? Half kelly? Quarter kelly? Or something else?

[varkolek]

I looked into that Kelly thing, and it seems to involve betting small amounts on heavy favourites, and huge amounts on dogs who have a high probability of winning.

This isn't a scientifical explanation, but I think the probability of who is going to win is unknowable, so I prefer just having consistent and low exposure over many bets than big exposure on the bets you're confident with.

Like I might believe the probability of a fighter winning is 90%, and the actual probability is 2%. So staking a huge fraction of my bankroll is detrimental in the long run because some of my picks won't be accurate.

This last card I thought Dustin Poirier, Demarques Johnson and Kid Yamamoto were all going to win. On heavy favourites like this were I think there is negligible value I bet 2.5%. I think I had exposure of 3% on Kid and I lost that, because my belief in the probability Kid would win was wrong. I didn't have time to research the others, but if I were exposed to 3% on all those three it would have nearly covered my loss.

I bet around 5% were I think the odds have some value. Like the odds might be 1.5 for 1 fighter and 2.5 for another. The probability the fighter wins with odds of 1.5 might be 10%, and the probability the fighter wins with odds of 2.5 might be 90%. I just go for roughly the same exposure on both and don't concern myself with the odds too much. I try to get the highest odds possible, but the odds don't reflect an actual probability of who is going to win. And no one include myself knows the probability of who is going to win. I don't trust my own judgement, because half the time the bets I'm confident with fail, and my second guesses are correct.

Like for Marquez-Pacquiao the odds were 2.5 for the fight going over 10.5 rounds, meaning the odds reflect over 50% probability that Marquez won't get knocked out. The only time Marquez got rocked by Pac in the second fight, was when Marquez pursued Pacquiao, and I didn't think he'd do that this time. When I think the probability is better than 50% I bet around 5% of my bankroll. So by my own rules I should have bet about 5%, but I think I only had exposure of around 3% which included some on Marquez. I do this often and it hurts my profit I think.

If the odds for Pac-Marquez over 10.5 was 1.2 then I wouldn't bet, but I'm not going to analyse whether I should bet (or how much I should bet) if the odds are 2.3 or 2.5. I just limit my exposure so that when inevitably I get some of my picks wrong there isn't a catastrophic loss.

Some of my biggest losses in the past have come from hedging bets, where I think I have all the outcomes figured out, and I'm exposed to over 7.5 or 10%, and then an outcome happens I didn't expect. That's big exposure for low profitability. I would rather have low exposure for big profitability. And I got hedges and stuff wrong often even when I'm confident I have all probable outcomes covered. These 'certain' bets at low odds made the big losses to my bankroll, not small bets at high odds. I still do it occasionally but for total exposure less than 6%.

The Kelly formula or other formulas depend on your estimation of probability of the winner, and that's a huge variable considering you can't actually know that. I'm sure there's some mathematical way of analysing stats and coming out ahead, but I just try to bet relatively consistent amounts and I've been coming out ahead for a while now.

[illmatick]

Depending on how efficient I think the market I'm betting into is determines how aggressive I am with Kelly.

Sports like MLB,NHL, and Football are where I feel a lot more comfortable using pinnacles no-vig line as an indicator of actual probability. In those cases I use a lot closer to Full Kelly.

A lot of it comes down to how accurately you feel you've been able to define your edge.

[NunyaBidness]

I looked into that Kelly thing, and it seems to involve betting small amounts on heavy favourites, and huge amounts on dogs who have a high probability of winning.

That's not at all what kelly is.

[NunyaBidness]

Would love to hear your money management strategy.

Do you use full kelly? Half kelly? Quarter kelly? Or something else?

Kelly isn't "money management" it's bankroll growth optimization. 1/2 kelly ideally, but it's easy to bump into limits in mma/boxing.

[NunyaBidness]

Depending on how efficient I think the market I'm betting into is determines how aggressive I am with Kelly.

Sports like MLB,NHL, and Football are where I feel a lot more comfortable using pinnacles no-vig line as an indicator of actual probability. In those cases I use a lot closer to Full Kelly.

A lot of it comes down to how accurately you feel you've been able to define your edge.

Agree.

I use the exchanges vig-free lines over Pinny, I think they seem a tiny bit sharper, especially since the word is that most of the big offers on the exchanges are actually pinny anyway.

[Vitooch]

Can anyone provide links to the basics of Kelly Criterion and other money management strategies?

I personally have a tough time judging how much money I should be shelling out for each fight. The spread system in football and other sports makes betting amounts easier to judge, but the steep odds in MMA makes this difficult. Futhermore, I am on a low budget and am only able to spend from 100-200 dollars per event, and that is usually a big percentage of my bankroll. Plus I have a really tough time laying down the juice on certain fighters when I only have a limited amount of money to risk

[NunyaBidness]

http://en.wikipedia.org/wiki/Kelly_criterion

[Vaughany]

I use this Kelly strategy calculator, gives a nice summary after you enter the figures... http://www.albionresearch.com/kelly/default.php

[warriorfan707]

The Kelly Criterion is quite possibly the biggest fukking joke in the history of mankind

[TheCalculator]

Kelly isn't "money management" it's bankroll growth optimization. 1/2 kelly ideally, but it's easy to bump into limits in mma/boxing.

Hi Nunya -- Excuse my ignorance here - isn't growth optimization a part of money management? Care to elaborate on the differences?

[NunyaBidness]

Hi Nunya -- Excuse my ignorance here - isn't growth optimization a part of money management? Care to elaborate on the differences?

Yeah, in a sense it is. But money management more often than not gets used to explain all sorts of things that don't belong in a professional sports bettors toolbox. Chase systems, martingales, all sorts of systems and rules of thumb that might sound good.

Kelly is the only 'money management' system to use. Quants use it on wall street, every top pro bettor uses it, pro blackjack players use it. It's maths, not what some guido's uncle told them to do at the track.

Kelly will optimize your growth and reduce your ROR to 0.

[warriorfan707]

Pro blackjack players use Kelly betting?

WTF are you talking about? The whole idea of kelly betting is coming up with an estimated % for winning each bet

In blackjack that number remains pretty constant

[Vaughany]

Pro blackjack players use Kelly betting?

WTF are you talking about? The whole idea of kelly betting is coming up with an estimated % for winning each bet

In blackjack that number remains pretty constant

Dude, the kelly criterion was created by Thorp FOR blackjack! Suggest u read this... http://www.bjmath.com/bjmath/thorp/ch9.pdf

[warriorfan707]

I'll read it later

I wonder if I should try using it at the tables sometime

[TheCalculator]

Kelly is the only 'money management' system to use. Quants use it on wall street, every top pro bettor uses it, pro blackjack players use it. It's maths, not what some guido's uncle told them to do at the track.

Kelly will optimize your growth and reduce your ROR to 0.

ROR? You mean risk of ruin?

I have to admit -- I ignored Kelly on Saturday night and bet big on Cain. And LOST BIG. I feel fawkin stupid for at least not hedging.

[TheCalculator]

I'll read it later

I wonder if I should try using it at the tables sometime

Here's more stuff related to BJ: http://bj21.com/bj_reference/pages/k...lization.shtml

Here's another good kelly calculator: http://www.sportsbookreview.com/betting-tools/kelly-calculator/

[varkolek]

I found a Kelly Strategy calculator that allows for various odd formats, including decimals: http://www.sportsbookreview.com/betting-tools/kelly-calculator/ I didn't understand the odds in that thing Vaughany posted.

Edit: TheCalculator already posted a link to this.

[varkolek]

That's not at all what kelly is.

Well, the part of the formula: (b + 1) - 1 / b
where b is the odds, means that the larger the odds are in relation to 1 the larger your stake is. So for the same probability you bet bigger on dogs.

Even at fractional stakes, the amounts it gives for betting on serious dogs have unacceptable exposure. But this article: http://www.investopedia.com/articles...#axzz1diohD8hv says regardless of the Kelly percentages, you shouldn't commit more than 20-25% of your capital.

I suspected Nunya was lying with his comment all top pro gamblers use it, so I quickly googled 'kelly criterion do all gamblers use it'. I didn't get any real pertinent results to the search, but I scanned a few results and found the following things:

http://www.madduxsports.com/kelly-criterion.html
"Kelly Criterion normally fails for most bettors because they cannot accurately estimate their true winning percentage. Almost all gamblers overestimate the edge they truly have and this is normally the reason sports gamblers lose betting with Kelly Criterion."

That's what I was saying. I flat bet according to bracketed probabilities.
If I think the probability is about 50:50 I bet 2.5%.
If I think the probability is better than 50%, meaning one fighter has an edge, I stake 5%.
And if I think it's a lock, meaning a probability of 75-85% I stake around 7.5%.

There's no way I could possible determine if someone has 67.8%, or to be realistic, even 60% chance of winning. With NFL and stuff maybe, but for a fighter I can only say I think maybe they won't lose.

For example, if I think the probability is a bit above 50%, I shouldn't bet unless the odds are at least 2.0 or +100 or whatever in American odds. But I'll often bet at odds of 1.5, because I think the actual probability is such a big variable it negates discrepancies in the odds like that. I often think there's no practical difference in odds of 1.5 to 3. You win some and lose some, eventually it evens out.

Also, I found a Kelly hater:
http://www.professionalgambler.com/debunking.html
"Right here is precisely where Kelly promoters always screw up. Let's say they have 100 actual bets wherein they win, say, 58 and lose 42. That's a great winning percentage of 58%, of course. Now, they'll explain that their basic bet is, say, $100, but if their expectation is higher than such-and-such percentage they risk $120, or $130, or whatever, and if their expectation is even higher than such-and-such they might risk $200 or more.

Then they compare what they won by using the Kelly system to what they would have won had they been risking only $100 on each of the 100 bets.

.....Duhhh!.....They risked more money with the Kelly system and they made more money after going 58-42. I hate to burst their balloon, but when you go 58-42, the more money you risk the more money you figure to make. Sorry, boys, but that is not news.

The only way to fairly compare the Kelly system (or any other progressive betting scheme) to flat betting is to use a flat bet the same size as the average size of all the Kelly bets. That way, you’re risking the same total amount against the same overall won-lost results. No fair risking more money overall with one system than the other. That obviously skewers the results.

Or, another way to fairly compare betting systems is to keep track of the winnings as a percent of the total amount risked. If Betting System A wins 8 percent of all monies risked while Betting System B wins 12 percent of all the monies risked, Betting System B is obviously better than Betting System A.

This is precisely what Kelly-promoters choose to ignore. Like religionists, they want so very hard to believe their fantasy they won't be made to face facts, no matter what. Comparing flat betting against a "1-star, 2-star, 3-star" system, for example, and going 58-42, if all your flat bets are only as big as your "1-star" bets, of course you will win more with the star system. You're risking more and you're winning 58% of your bets. "

and:

"The cold hard fact is that all progressive betting systems are nothing more than modified versions of the Martingale system. In the Martingale, you risk one unit, and if you win you keep risking one unit. If you lose, you double your bet, and if you lose again you re-double and keep re-doubling until you finally do win. Then you go back to risking one unit.

As any fool can plainly see, the Martingale can’t miss, so long as you win one more bet before you die you’re going to be a winner.

Well, yeah, if you lose 12 bets in a row you’d have to risk $409,600 to win $100, but how often is that going to happen?

As it turns out, plenty."


It's a long article but he hates it for Blackjack also.

Seriously, I mean no disrespect, but I think it's easy for newb gamblers like myself to come here and see people with proven success and this method used widely, and accept stuff like the Kelly strategy as an appeal to authority.

I can't dismiss it outright without trying it, so when I get time I might put aside an experimental bankroll, like $100, and use fractional Kelly stakes and see how it pans out. At the moment I'll continue flat betting though.

[NunyaBidness]

I suspected Nunya was lying with his comment all top pro gamblers use it, so I quickly googled 'kelly criterion do all gamblers use it'. I didn't get any real pertinent results to the search, but I scanned a few results and found the following things:

Lying? That's a pretty strong word, asshole.

I've never met a pro sports bettor who doesn't use it, nor heard of one who has proven results. Glad you could find a website that someone making some claim doesn't use it.

I guess you're right, I'm lying, there's probably someone out there who is minimizing his results by ignoring an extremely ELEMENTARY aspect.

[NunyaBidness]

Not to mention, the two sites you brought up are ******* TOUTS! Exactly the type of people who benefit from selling THEIR system.

Hmmm, let's try doing research your way.

I think the sky might not be blue. See, look, this guy agrees!
http://nayarweb.com/blog/2011/the-sky-is-not-blue/

Guess what, humans are NOT evolved from apes:
http://www.firstthings.com/blogs/sec...lve-from-apes/

[NunyaBidness]

Pro blackjack players use Kelly betting?

WTF are you talking about? The whole idea of kelly betting is coming up with an estimated % for winning each bet

In blackjack that number remains pretty constant

You know nothing.

Kelly doesn't estimate your % chances of winning, it calculates your ideal bet size from a known edge.

You can use it on slots machines, blackjack, flipping a coin, etc, etc, etc.

[NunyaBidness]

ROR? You mean risk of ruin?

I have to admit -- I ignored Kelly on Saturday night and bet big on Cain. And LOST BIG. I feel fawkin stupid for at least not hedging.

Yes, risk of ruin. Kelly will never have you bet 100% of bankroll, you adjust your kelly multiplier to your amount of risk adverseness.

Most people overestimate their edge, and then undersize their bets based on that edge, which I guess saves them a little bit.

[Educ8d Degener8]

F*ck the Kelly Criterion... It's all about the Gabe Betting System -- when you make your bets after the event has finished. Most profitable system. Ever.

[NunyaBidness]

A simple Kelly example:

Your hypothetical friend offers you +110 on the flip of a fair coin for any amount you wish, but you must flip with him 1,000 times. Your bankroll is $100. How much do you wager on each flip?

Flat betting $2, your expectation is to be up $100, with a ROR of near 0%

Flat betting $5, your expectation is to be up $250, with a ROR of around 8%

Flat betting $10, your expectation is to be up $500, with a ROR of around 35%

Betting kelly you would start betting at $4.54 per flip, your expectation will be +$868 with a ROR of 0%

[varkolek]

Lying? That's a pretty strong word, asshole.

I used the strong wording thinking it would be pretty obvious it was tongue-in-cheek, to lessen the blow.

[MMAbetMASTA]

F*ck the Kelly Criterion... It's all about the Gabe Betting System -- when you make your bets after the event has finished. Most profitable system. Ever.

lol...well played, good sir.

Personally, I've looked into the kelly, but I don't use any system really. If you can pick fights, break down styles, know the ins and outs of mma and technique, have good bankroll management and a good idea of betting ratios for each play - then you don't need some formula or strategy. Natural ability in analyzing fights/mma and overall reasoning trumps any betting system in this sport imo...

[varkolek]

Not to mention, the two sites you brought up are ******* TOUTS! Exactly the type of people who benefit from selling THEIR system.

Hmmm, let's try doing research your way.

I think the sky might not be blue. See, look, this guy agrees!
http://nayarweb.com/blog/2011/the-sky-is-not-blue/

Guess what, humans are NOT evolved from apes:
http://www.firstthings.com/blogs/sec...lve-from-apes/

Well, you're probably right. But I didn't mean my responses to be definitive answers, just my honest quick impressions.

Regardless of whether this quote is from a sell out or whatever, selling his system: "Kelly Criterion normally fails for most bettors because they cannot accurately estimate their true winning percentage. Almost all gamblers overestimate the edge they truly have and this is normally the reason sports gamblers lose betting with Kelly Criterion."

I think that quote applies to me. I think about this system and the only thing I can think is that my win rate goes down. My biggest ever loss was exactly half my bankroll and that was awful for me, if I lose like 10% now I get gunshy after that.

When I get time I may try using it, but I still think some of those arguments I linked to have merit even if they're biased or commerical. At the moment I have no way of verifying if some of the points are valid or not. But like you bet 20% and have 4 consecutive losses, your bankrolls down to nearly nothing. I get the Kelly strategy is a means to get quickest growth, but I'm not comfortable with that, I'd get a ulcer.

I've increased my bankroll by trying to lose as little as possible rather than winning big or quick.

Anyway, if I have (probably) erred, I hope no hard feelings, maybe the discussion and incorrect assumptions are informative for other newbs or something.

[NunyaBidness]

I used the strong wording thinking it would be pretty obvious it was tongue-in-cheek, to lessen the blow.

Sorry for my response then, tone is difficult to grasp across the internet.

[NunyaBidness]

Well, you're probably right. But I didn't mean my responses to be definitive answers, just my honest quick impressions.

Regardless of whether this quote is from a sell out or whatever, selling his system: "Kelly Criterion normally fails for most bettors because they cannot accurately estimate their true winning percentage. Almost all gamblers overestimate the edge they truly have and this is normally the reason sports gamblers lose betting with Kelly Criterion."

I think that quote applies to me. I think about this system and the only thing I can think is that my win rate goes down. My biggest ever loss was exactly half my bankroll and that was awful for me, if I lose like 10% now I get gunshy after that.

When I get time I may try using it, but I still think some of those arguments I linked to have merit even if they're biased or commerical. At the moment I have no way of verifying if some of the points are valid or not. But like you bet 20% and have 4 consecutive losses, your bankrolls down to nearly nothing. I get the Kelly strategy is a means to get quickest growth, but I'm not comfortable with that, I'd get a ulcer.

I've increased my bankroll by trying to lose as little as possible rather than winning big or quick.

Anyway, if I have (probably) erred, I hope no hard feelings, maybe the discussion and incorrect assumptions are informative for other newbs or something.

I think the point you're missing and maybe everyone misses is that kelly is based on % of bankroll, and requires recalculation after each bet to optimize. So, like in my coinflip example, if you win the first bet, your new bankroll would be 104.99 or 95.46 if you lost, and your next bets would be $4.77 or $4.34 respectively, continued ad infinitum. Since you have a +expectation bet here your bankroll will tend to grow and you will continue to make risk free bets at higher amounts as your new bankroll can withstand the variance.

In your 4 bets of 20% example, you would be reducing your betsize with your bankroll, so if you had lost 20% on your first bet bringing your bankroll to $800 from $1,000 let's say. Your new bet is 20% of the new bankroll or $160, if you lose again, your new bet is 20% of the $640, etc.

If you had WON the first bet, now your second bet is 20% of the $1200, instead of 20% of $1,000. So your bets decrease as you lose and increase as you win.

If your bets are simultaneous (like most MMA bettors will be) It gets more difficult, and constant recalculation isn't practical, but the general idea holds true.

[TheCalculator]

When I get time I may try using it, but I still think some of those arguments I linked to have merit even if they're biased or commerical. At the moment I have no way of verifying if some of the points are valid or not. But like you bet 20% and have 4 consecutive losses, your bankrolls down to nearly nothing. I get the Kelly strategy is a means to get quickest growth, but I'm not comfortable with that, I'd get a ulcer.

I've increased my bankroll by trying to lose as little as possible rather than winning big or quick.

Anyway, if I have (probably) erred, I hope no hard feelings, maybe the discussion and incorrect assumptions are informative for other newbs or something.

You can bet half-kelly (half of a normal kelly bet) and get 75% of the results of full kelly. This is well detailed inside of:
http://www.amazon.com/Fortunes-Formu.../dp/0809046377

This obviously cuts your risk down quite a bit also.

[varkolek]

In your 4 bets of 20% example, you would be reducing your betsize with your bankroll, so if you had lost 20% on your first bet bringing your bankroll to $800 from $1,000 let's say. Your new bet is 20% of the new bankroll or $160, if you lose again, your new bet is 20% of the $640, etc.
...
If your bets are simultaneous (like most MMA bettors will be) It gets more difficult, and constant recalculation isn't practical, but the general idea holds true.

Yeah I realised after posting that with consecutive bets, the 20% would be a fraction of the new bankroll, not 20% of the original so it was a bad example.

The problem in my mind is I don't think my estimation of the probability of a winner is going to gel with the odds. If you bet more at high odds (but equal probabilities), it's going to be more profitable. I just think I'll get the probabilities wrong too often and my bankroll will dip way down each time.

[varkolek]

I don’t mean to bump the **** out of the thread, but it appears I’ve committed some kind of faux paux in referencing that article by JM Miller, Debunking the Kelly System. A lot of smart people here: http://www.sportsbookreview.com/forum/handicappe...imulation.html disagree with him, and they actually did his test on a spreadsheet and Kelly came out ahead.

But in that thread, Wrecktangle advises against it for laymen, and in this thread: http://www.sportsbookreview.com/forum/handicappe...nt-system.html he says fractional Kelly isn’t optimal, only full Kelly is, and he says you need something called Bayesian Priors.

I read something here: http://beyondthebets.com/2011/07/12/...ports-bettors/ which said the success of the criterion is dependent on the chances of winning and losing being entirely accurate (doubt I can do this).

It’s all very deep, beyond my understanding. Perhaps I keep flat betting like a dumbass and winning a bit until I figure the rest out.

Anyway, I thought it might be of interest to others who don't know these things.

[NunyaBidness]

Good stuff. Glad you're doing the research, it will pay dividends down the line.

Read everything by Ganchrow.

[rocky mattioli]

why doesn`t someone put up a running thread from this point forward applying kelly to their wagers so the rest of us can see it actually work in real time.......

like vaughany,ecco,ill and hoff do....keep it in the same thread so it`s right there as proof positive that it`s solid bankroll mgmt....

it`s possible that the person that undertakes this effort is a lousy mma capper,but i`d enjoy seeing how the bankroll mgmt issue plays out.....


anybody interested in showing the rest of us?...