General Betting
There is currently 1 person viewing this thread.
A thread about staking - without doubt the weakest aspect of my own game.
Setting aside a bankroll of £1,000 for this experiment, let's see where a staking policy based broadly on Kelly principles takes us.
Kelly, in a nutshell, is maths to find the right stake for any given bet - and that depends on (1) the size of the bankroll, and (2) the "edge" (or value) that you think you've found in the market.
I'm aware that Kelly, applied strictly, is seen by some as too risky, and so prefer instead to divide the Kelly Stake by 2 or more. But I'll just see how the idea develops, drawing conclusions along the way.
Setting aside a bankroll of £1,000 for this experiment, let's see where a staking policy based broadly on Kelly principles takes us.
Kelly, in a nutshell, is maths to find the right stake for any given bet - and that depends on (1) the size of the bankroll, and (2) the "edge" (or value) that you think you've found in the market.
I'm aware that Kelly, applied strictly, is seen by some as too risky, and so prefer instead to divide the Kelly Stake by 2 or more. But I'll just see how the idea develops, drawing conclusions along the way.
sort by:
Show
per page
Replies: 314
|
By:
Good luck with the haters. Bear in mind when they come along that most people who say Kelly is bad have no idea how or why it works.
Modifying Kelly for your own personal situation and ability is the way to maximize the return from your edge so all the best. Remember to use a calculator that gives EG as well as EV as this is pretty much the point of the exercise. |
|
By:
With the possible exception of a listed race at Longchamp this afternoon, the only price that stands out for me today is the 7/1 about Rasaman (H i l l) in the 3.40 at Redcar.
Bearing in mind my tendency to overrate sprint handicappers in their bids to follow-up off a higher mark, let's be cautious and assume that a price of 11/2 better reflects its chances. Bank = 1000 Available odds = 7/1 My odds = 11/2 Kelly stake = £35.20 (or £35) |
|
By:
Cheers, Lori.
I am using a calculator, unfortunately, and hope it works... |
|
By:
I get 15.38% for 11/2 which gives $32.90 or so fwiw.
|
|
By:
Hi bacchanal - long time no see.
Good luck with this, I'll follow it with interest. |
|
By:
Cheers, Gin.
Although backed at 7/1 & sent off 9/2, an uninspiring effort from Rasaman there. The 5f conditions race at Beverley tomorrow night has an interesting shape because Waffle may be overrated on his juvenile form and there's little strength in depth elsewhere in the field. For a horse that has not yet won as a 3yo, maybe I like Noble Storm a bit too much but I rate his chance as no worse than 2/1 (33%) in this event. Should perhaps wait for the markets tomorrow, but the early price of 10/3 at P o w e r is something I'd rather take now. Bank = 965 Available odds: 10/3 My odds: 2/1 Kelly stake = £128.43 (or £130) This raises an early question: at over 13% of the available bank, is the suggested stake too high? As a punter who seldom commits more than 3% on a single bet, 13% seems ludicrously high. The bank could be wiped out before we've really started... |
|
By:
I'm trying this with tennis and anything higher than quarter kelly is too high in my opinion.
You also need to make sure your pricing is accurate and has an actual edge. Trouble is, in the markets we operate the real price is extremely difficult to determine and very much open to opinion. When you lose 10 value bets on the spin will you still believe you have an edge? What I'm finding more helpful is choosing a winner first then calculating stakes based on my price and this works better as you don't get disheartened by the seemingly pointless value bets that you never believed were going to materialise in the first place. This coupled with trading value that is too low/too high (but not sure of end result) is what is working best for me. |
|
By:
I have said on ere many times before, Kelly is good but you need to know the "fuzziness" in your estimation.
Also take commission into account to get the true odds (with the PC this is even more complicated - and its my main complaint about the charge) Also if you're not confident about estimating your odds, try using the Kelly in reverse. |
|
By:
Thanks for the input. This business of estimating true odds is interesting. My approach tends to be a lot less sophisticated. The basic premise, of course, is always to wait for something you believe has a better chance than the market suggests. But instead of maths, I'll tend to substitute phrases like "fair value", "good value", or "excellent value" and then apply what seems an appropriate stake.
My long-term punter's instincts tell me that 13% on a 10/3 shot is much too high. A 3pt bet to win 10pts (where 1pt = 1% of the starting bank) looks more balanced. |
|
By:
I rarely go over 10% for any bet - and thats only if I'm feeling more confident than one of those skydiving bodyform women.
Try reversing the Kelly it may make life asier if you're not being too clinical about it. |
|
By:
If your estimate is correct, the bet sizes will be correct.
The reason your instinct is raising alarms is because it knows from trial and error that your estimate is likely to be incorrect as an edge of this size is incredibly big and unlikely to be found in modern efficient markets. This is also the main reason people don't use full kelly. The better your actual pricing, the closer to full kelly you can use. If, for instance, you were getting bigger than evens about a coin toss that you knew was a fair coin, you'd be silly to use anything other than full Kelly. |
|
By:
listening to irishmen will surely send you broke
|
|
By:
Also, be careful with your percentages. You've taken 33%, which means you bet £124.48 for an Expected Growth of £24.91
2/1 is 33.33% which allows a bet of £128.65 for an EG of £26.57 In reality it's likely you've overestimated your edge on this one anyway, but thought I'd point out on a theoretical level how crucial the little things can be! |
|
By:
Sorry, you've taken 33.33 I believe, which is only a few pence different.
|
|
By:
price the field up at 85% then bet if you egt at least 1 and half your price...and bet accordingly you dont need some irish telling you
|
|
By:
Interesting to see Noble Storm battered into a best price of 7/4 by 11am this morning.
Zola's Back Heel: excuse my ignorance, but what do you mean by reverse Kelly? |
|
By:
Good call :)
Just thought I'd show a little chart of what would happen if you backed it three times at 100/30 and it was really a 2/1 shot, might help you feel better about such a large outlay. Let's say it's £100 for each bet for ease of calculation (it will change depending which order they win in and three losers in a row you'd have reduced a little) 0 winners would be -£300 and happen 29.6% 1 winner would be +£130 and happen 44.4% 2 winners would be +£560 and happen 22.22% 3 winners would be +£990 and happen 3.7% When it's put like that you can see how hard it is to go on a run of catastrophic proportions. You will halve your bankroll 1 time in three with Kelly and double it the other 2 times in 3. It's very aggressive like that, but obviously that works out very nicely in the end |
|
By:
Very honest, no-nonsense performance from Noble Storm there.
A 10p rule 4 (caused by the defection of a horse with a less than 10% chance of winning) produced deducted winnings of £390. New bank = £1,355. The horse was backed all day and on-course - sent off 5/6. So, leaving aside the non-runner, the estimate of its chances at around 2/1 turned out to be notably conservative. (Idle speculation, of course, but it wouldn't surprise me if the Vietch mob got stuck into that one today. The normally secretive Patrick made clear his liking for the horse when interviewed at Chester recently.) |
|
By:
The kelly tells you how much to stake given the odds and the estimated odds.
It is at the very least educational to**the amount and then determine the estimated odds - in relation to the odds offered - required to make this a profitable bet (in the longrun). Compare this to what your independent assesment of the true odds are and you might be surprised. |
|
By:
The Prix Corrida at Saint-Cloud today is an interesting race. Shemima is a worthy fav but likely to be overbet, due to the dream team of Aga Khan, Royer-Dupre & Soumillon.
And the 2nd fav Albisola appears to be a nutcase - she has the talent but was reluctant to load last time and lost 10 lengths out the gates. A couple of horses of interest are: Tubular Bells - unlucky in-running in the Allez-France last time, and Guantana - an completely unknown quantity from respected connections in Germany. Both are likely to be a bit bigger in the market than they should be, but I'm not sure how to "Kelly" a market like this, given that the pari-mutuel price is not known until the off. |
|
By:
OK Guantana is 16s on the PMU 10 mins before the off. I tentatively have her down as a 12/1 shot which gives a Kelly stake of £26.02 (call it £25).
Also Tubular Bells is around 22 on Betfair. I rate her more a 16/1 shot, which (taking into account commission of say 5%) gives a Kelly stake of £15.75 (call it £15). |
|
By:
Haven't seen the race yet, but in spite of getting the top two out the frame, my imagination was unable to find the 19.5/1 winner.
Bank now = £1,315 |
|
By:
I like this thread a lot, best of luck with it bacchanal.
|
|
By:
The National Stakes at Sandown this evening has a nice each-way shape to it: 8 runners and an even money favourite.
I don't see anything to get too excited about here, except that I was pretty impressed with the manner of High Spice's maiden win at Newmarket on 1000 Guineas day. The fav has clocked up some fancy numbers and is no doubt a worthy market leader. The market also says that Raine's Cross is well-fancied. High Spice is only 7s and has copped the worst of the draw. I'm going to rate High Spice's chance at around 15% (ie somewhere between 6/1 & 11/2) and will split the full Kelly stake for each-way purposes. Bank = 1,315 Available odds: 7/1 (B l u e S q) My odds: 15% (5.66/1) Kelly Stake = £37.57 (say £35) = £17.50 e/w |
|
By:
thanks for the kind remarks, do you like owls??
|
|
By:
Bank = £1,280
In the 4.55 at Yarmouth Kheylide could be overpriced at 6/1 (S t a n J a m e s, B et 3 6 5 ) due to his bad run last time out. But that was on very slow ground. On better ground, form lines in 3yo sprints with the likes of Sloop Johnb, Ziggy Lee and even Dispol Grand look quite promising. He's now stepping into competition against his elders, and its not easy to assess the compare the two strands of form. I estimate his chance as no worse than 5/1 (16.67%). Bank = £1,280 Available odds: 6/1 My odds: 5/1 Kelly stake = £35.61 (or £35) Debbie Sanderson has made an excellent start with her small string. The handicapper will catch up with her (maybe he already has) but for the time being she shouldnt be underestimated. She also has All The Nines running for her up at Hamilton today. |
|
By:
For a bit of contrast, and in the hope that this thread doesn't develop into yet another sad tipping sheet, here's something completely different:
The general principles of Kelly make perfect sense, says Simon Rowlands. The recent discussion about staking plans on betting.betfair generated some light as well as a great deal of heat. I was an interested - and at times bemused - spectator, but my ears (metaphorically)**ed up when one particular word was mentioned: Kelly. Some explanation is required for anyone who has not come across this name in this context before. The crux of the issue where staking is concerned is how much of your available money you should risk in given circumstances. Theories of probability are often illustrated by coin-toss experiments, or similar, in which the "true" odds of an outcome are known. In such circumstances it does not take a genius to say what a team of 100 punters with a bank of £1 each and an agreement to share any returns should do when offered 10/1 against heads on independent but fair coin tosses. They should each stake their whole bank in the knowledge that on average 50 of them will be skint at the end of the experiment but the team will be £500 up and the collective payoffs will more than compensate for individual losses. The chance of none of them getting a return is negligible: roughly 1 with 30 zeroes after it to 1. But what if you are just one punter, not part of a team, and your £1 is all you have? Putting it all on an evens shot that is being offered at 10/1 makes plenty of sense in one respect, but there is still a 50% chance that you won't have the price of a cup of tea at the end of the day. It is important, in the real world, not only to seek to capitalise on value when it occurs but to do so in a manner that maximises your chance of being around to benefit from that value. Don't stake the farm every time someone offers you a fraction over the odds, in other words. This conundrum, in a very different context, was what preoccupied a certain John Kelly at the Bell Laboratories in the 1950s. His complex mathematical paper, published in 1956, dealt with telecommunications but was illustrated by way of examples that had a clear and specific application to betting strategy. Nick Mordin devoted a chapter to the Kelly Criterion in his book Winning Without Thinking, which also had the original paper in the appendix, and it is well worth a read. In essence, the amount you stake should be dictated by a combination of the edge you have (10/1 as opposed to evens in the example above) and the likelihood of the event occurring (50% in the example above). Mordin helpfully provided a simplified version of Kelly that mere mortals can understand and apply. |
|
By:
Well done Debbie Sanderson, but the wrong one won!
Such a busy day and yet very little stands out at the early prices. But I don't think Ishe Mac should be 16s in the 4.55 at York (B e t F r e d). Ran in a better race over C&D last time out and got messed about a couple of furlongs from home. 12/1 is maybe a more correct assessment of his chance, but I'm going to give him an even better shot at 10/1. Could go off even bigger on betfair later, given the unfashionable connections, but I think the early price is worth taking. The shape of this race suggests each-way is the way to go. Bank = £1,245 Available odds: 16/1 My odds: 10/1 (9.09%) Kelly stake = £42.43 (call it £20 e/w) |
|
By:
http://www.professionalgambler.com/debunking.html
Takes a different view on Kelly |
|
By:
Apparently maths was not the debunking guys strong suit. I think we can safely ignore his ramblings since he himself was shown to be wrong when this first appeared.
|
|
By:
cpf
Would you care to elucidate |
|
By:
bacchanal, yr thread is so much better than a tipsheet in that you give your view of the horse's actual chances.
My view of Kelly is that it tends to lead you into overstaking just when your view of the market is most skewed. When someone thinks the odds are way out, there's usually some piece of info. they haven't absorbed or something they're not taking into account. Having said that, for someone w/ a very consistent edge (pos. derived from a mechanical, rather than judgment-based system) it's clearly the best way to maximize bankroll growth. I use a fraction of Kelly, book a profit at just under the expected return (to manage comm.) and start asking why I'm wrong when the price starts to drift further. Gl w/ yr bets |
|
By:
good thread Bacchanal :)
|
|
By:
I tried this about about 5 years ago after reading it in one of Mordins books, end result was that I got iin a right mess.
|
|
By:
Yeah that link to the pro gambler guy is a course in how to be dumb.
|
|
By:
as opposed to askari's view which is the best anti-kelly reasoning there is.
|
|
By:
askari is not questioning the kelly criteria, just the use of it when you don't know your edge. thats not an anti-kelly argument.
kelly is a great weapon in the punters arsesnal. just don't point it at your own face. |
|
By:
I'm in total agreement zola, i must have come across wrongly.
I've seen that propunter link before, it's disgusting that stuff like that is allowed on the internet because it's so ill-informed. It's impossible, as you know, to say "Kelly is wrong" because it's not wrong and the proofs that it's not wrong aren't even difficult ones (I have very little real education in the scheme of things and I can follow them). Applying it to your own situation is the key. Anyone who doesn't "believe" Kelly does what it says is not cut out for gambling in the modern efficient environment. It applies to everything, even short term trading, you simply have to work out how to apply this very precisely defined tool to your own methods of gambling. |
|
By:
by anti-kelly reasoning i mean reasons not to use kelly. I suspect i blew my lines there :D
|