General Betting
There is currently 1 person viewing this thread.
These 249 comments are related to the topic:
Question for the Pros.
Question for the Pros.
sort by:
Show
per page
Replies: 249
|
By:
I must have been desperate for a distraction from working on my 17% ROI picks
 |
|
By:
17%! you need your own thread :)
|
|
By:
h-bomb
RE: Picking selections with no regard for price, price doesnt matter etc etc The only time price doesnt matter is it you have a 100% strike rate. Then you can simply take any price you want since all your bets win. As this is not the case, the only reason you are in profit after all the bets made in your sample is solely because of the excellent prices you have been getting (providing of course we consider the sample large enough). Your numbers are hard for most of us to believe as has already been pointed out on here, they would make you an overnight millionnaire and one of the best gamblers to ever lived! If I go back through my records, i'll find similar periods where I've had a ROI as high and even higher than that and then there will be periods where I've had a disgustingly negative ROI. Nearly all gamblers will have at one time or another, had a run of months where everything they touched turns into gold, only for the variance to catch up and bite them in the ass, finding themselves back at the deposit page once again. I do believe your figures, but for me you've gotta go atleast full season for them to become anywhere near meaningful, and even then its not a forecast as you can expect that number to change from year to year. In terms of staking, I would just stick with something very small (like tobe has pointed out) whilest you're in the back building stage. Plus it can help guard against unrealsistic expectations an early ROI may have given you. |
|
By:
I am showing a net profit of just under 4% and I thought
that was good. Apparently not.  |
|
By:
h-bomb, kelly is a formula where you bet a certain percentage of your bankroll based on how great the divergence is between the odds you are being offered and your estimate of the true probabilities.
It is the best method provided that your assessment of the odds is equally accurate in each case. Usually it won't be, and you'll be more inaccurate the more confident your are (because the market has weighted some factor better than you.) The formula is much easier to understand than you suppose--it's only 1) taking first taking the notional returns and adding one to designate a win; 2) multiplying by the actual chance; 3) subtracting one to designate the chance of a loss, then 4) taking all that and dividing by the odds you're taking. I've made it sound opaque but there's a clear page on wiikipedia amongst others. As for your margin, I can't quite believe it. I would kill for a long-term edge of 15% on day-to-day horses. I've had seasons where on heritage hcaps (not then so-called) I've made 48%; also seasons where I've spent lots of time and made -26% overall. |
|
By:
Any kind of single-bet, Nostradamus-style approach to reading the markets and always getting the best price on horses tends to return 20% on level stakes. The problem is the difficulty reading the markets, then getting on at sufficient liquidity at the top.
In fact, I'd say that getting on even at advertised prices is 95% of the battle. |
|
By:
Would you mind giving a quick kelly example please askari1 ?
|
|
By:
calculator for it here Lex
http://www.seriousbet.co.uk/calculators/kelly_staking |
|
By:
Kelly staking near useless in the real world.....
One of its major flaws (and it has a few) is its accuracy is dependent on the accuracy of the 'edge' calculation which is just a guesstimate most of the time. The object of a staking plan is to manage the risk of ruin of a betting bank against seeking maximum returns for that bank. In the real world, most of the people seeking out staking plans are trying out a 'system' that they don't even know if it has an edge let alone what figure it might be. Kelly is over complicated tosh, if your brainy enough to realise the importance of a staking plan to try out a system then all you need to do is set aside a bank for it and work out how many consecutive losses you want to allow the system before you admit its failure. Divide the bank by this figure and you have your bet size. Re-assess and adjust if the bank drops to 50% or doubles, repeat until bust or rich. One of the so called advantages of Kelly is that it adjusts the bet size according to the near fictional edge calculation but in reality this is poppycock! If you're running a system with odds or an 'edge' that requires a stake that varies enough to make any difference then you are actually running 2 (or more) different systems under the same bank and using Kelly will cloud the view of how each system is performing. |
|
By:
top2 - Can't see how Kelly can be called over complicated for starters and further the alternatives of fixed stake or point staking are really quite primitive in relation.
Kelly at least compels you to price up a selection and therefore the race to some degree and makes you aware that each selection has some sort of chance. This will hopefully force you to be realistic about "your own perceived odds versus the odds available". Too many times I hear flippant comments 3 point bet, 5 point bet, Can't be more than an x/1 shot etc. What fundamentals to people have behind these statements? Have they analysed the race to price up each selection? Can they really justify it if questioned? What Kelly does if used properly is to take your analysis to a different level and gives you a better understanding of the overall market for the race. Many times I've gone into a race with an immediate thought, that I was waiting for this horse to run next time and wanted to be on but until you analysed the race and given it a respective chance of winning you are simply following it blind. Recently for example I saw a race where I was waiting for a horse to run for some time but the market priced it at what I thought its real odds should be so I had to pass and I bet something else in the race which was overpriced IMO. This sort of discipline is good practice and is what Kelly can give you. I'd venture to say that 80% of punters don't even look at all the runners in the race properly before doing their bet. |
|
By:
Hello sirs,
my names is J. D. Rockefeller and I'd like to invest money in this terrific scheme! I'll be forwarding my Cayman Island account details to you straight away. Godspeed, h-bomb. $$$ |
|
By:
top2rated - I understand the point you are making, however say you have a system that rates a horse race and you have a STRATEGY that is to back any horse in the top 2 rated ( whow ...what a coincidence T2R
  ) that is 2-1 SP to win.If you trust your system ( as all the edges are built into it) to rate races to high standard, and you accept the "true odds" for a horse in top 2 rated to be EVENS ....ie 50% probability to win, then that would be one way to arrive at the probability. You can then apply KELLY...... If your system is cr ap ....then you will soon find that out ! If your system is truly built on "edges"......the probability indicated by a horse being in top 2 rated will stand the test of bets. |
|
By:
BTW - I am after going through all the races my system has rated and produced a list of stats where the horse is in the top 2 rated and its SP odds are EVENS to 2-1 SP. The figures below show SP/BF SP/IR lowest odds for each horse in this range of odds in top 2 rated :_
LOWEST SP BF SP IR ODDS ODDS ODDS 177.345 198.53 48.45 1.88 2.12 0.25 1.38 1.6 0.31 1.5 1.21 0.95 1.5 1.59 0.62 2 1.96 0 1.75 1.7 0.55 1.62 1.79 0.16 1 1.1 0.81 1.5 1.36 0 1.2 1.31 0 1.62 1.84 0 2 2.3 0.61 1.88 2.1 0.66 1.5 1.98 1.78 1.75 1.92 0.19 1.5 1.76 1 1.88 1.98 1.8 2 2.11 0.7 1.5 1.8 1.64 1.88 2.05 0 1.75 1.98 0.62 2 2.31 1 1.88 1.72 0 2 2.6 1.9 1.38 1.46 0.5 2 2.55 0 1.5 1.58 0.69 1.5 1.68 0.01 2 2.35 0 1.5 1.6 0 1.2 1.36 1 1.5 1.67 0 1.38 1.47 0 1.1 1.14 0.56 1.38 1.46 0.01 1.75 1.96 0 2 2.48 0 1.5 1.66 0 1.875 2.65 0 1.75 2.35 1.1 1.88 1.99 2 1.88 1.91 1.52 1.75 1.88 0 2 2.15 0 1.88 2 0.66 1.88 2.32 0.11 1.5 1.7 0 2 2.45 0.73 2 2.2 1.1 1.88 2.3 0 1.5 1.92 1.12 1.62 1.81 0.6 2 2 1.6 1.88 1.98 0.7 1.88 2.2 0 1.88 1.97 0.1 1.38 1.6 0 1.5 1.99 0 1.5 1.56 1.14 1.25 1.57 1.3 1.38 1.5 0 1.75 1.94 0.56 1.5 1.67 1 1.75 1.86 0.6 1.88 1.98 0 1.38 1.5 0.52 1.88 1.81 0 1.88 2.05 1.76 2 2.06 0.34 1.5 1.8 0.52 2 2.25 0 1.5 1.56 0 1.38 1.99 0.34 1.75 2.08 0 1.88 2.05 1.52 1.25 1.43 0.5 2 2.54 2.3 1.88 2.39 0 1.75 1.86 1.2 1.5 1.54 1.14 1 1.09 0 1.62 1.62 0 1.38 1.45 0.61 2 2.09 0.51 1.38 1.51 0.05 1.88 2.13 0 2 2.65 0 2 2.05 0 1.75 1.64 0 1.75 1.94 0 1.62 1.94 0.1 1.88 1.91 0.48 1.38 1.56 0 1.2 1.31 0 1.62 2.3 0.42 1.38 1.4 0 1.25 1.54 0.24 1.25 1.59 0 2.25 2.61 0.3 2 2 0 1.88 1.96 0 1.75 2 0.7 1.38 1.66 0.44 1.75 1.9 0 1.38 1.41 0 2 2.3 0.2 Now....in your opinion , do you consider the system is providing selections where true odds are EVENS? |
|
By:
Above was one period of results ....here are other period set of results :-
LOWEST SP BF SP IR ODDS ODDS ODDS 277.3 318.2 96.28 2 2.15 0 1.75 1.99 0.12 1.375 1.44 0 2 2.83 2.72 1.875 1.88 0 1.375 1.7 0.03 1.875 1.76 0.57 2 2.15 1.82 1.75 1.9 0 1.625 1.8 0.6 1.75 1.88 0 1.5 1.68 0.71 1.625 1.9 1.82 1.385 1.34 0 1.25 1.25 0 1.25 1.36 0.3 1.5 1.67 0.51 1.25 1.96 0 1.5 1.76 0 2 2.1 0 2 2.09 1.1 1.25 1.3 0 2 2.15 0.44 1.75 1.88 0 1.25 1.36 0.25 1.875 2.1 0 2 2.25 0.55 1.75 1.84 0.9 2 2.2 1.4 1.5 1.68 0 1.625 2.1 0 1.1 1.37 1.3 1.625 1.8 0 1.625 1.74 0.4 2 2.28 0 1.375 1.46 0 1.5 1.65 0 1 1 0 1.75 1.9 1.02 1.1 1.2 1.02 1.2 1.35 1.18 2 2.18 1.9 2 2.73 0 1.375 1.5 0 1.75 1.98 0 1.875 1.92 0.7 1.1 1.26 0.82 1.75 1.9 1.8 2 2.1 0 2 2.52 1 1.75 1.96 0.7 2 2.52 0.25 1.875 2.19 1.98 1.375 1.91 0 1.1 1.26 0 1.875 1.96 0 1.25 1.46 0.2 1.1 1.26 0 1.75 2.9 0 1.375 1.6 0.2 2 2.15 0.53 1.875 2.17 0.23 1.875 2.07 0 2 2.45 0 2 1.92 1.26 1.625 2.1 0 1.25 1.38 0.01 1.25 1.32 0 2 2.3 1 1.875 2 1.86 1.875 1.98 1.4 1 1.01 0 2 2.75 0 2 2.53 0 2 2.18 2.05 1.75 1.79 0 2 2.35 0.15 2 2.2 1.88 1.5 1.6 1.2 2 2.1 2.05 1.75 2.05 0.74 1.5 1.72 0 1.875 2.23 0 1.75 1.84 0.18 1.5 2.06 1.5 2 2.47 2.05 2 2.55 0 2 2.1 1.52 1.1 1.21 0 1.625 1.58 0.87 1.75 2.65 0.42 1.38 1.57 0 2 2.7 1.6 1.5 2.23 0 1.25 1.4 0 1.62 2 1.02 2 2.24 2.1 1.875 1.98 1.68 1.75 1.82 0 1.875 2.24 2.1 1.1 1.18 0 1.875 2.1 1.78 1.38 1.5 0 1.625 1.74 0 1.1 1.24 0.05 1.25 1.49 1.16 1.5 1.7 0 1.62 1.84 0 1.62 1.69 0 2 2.8 0.5 2 2.26 0 1.75 1.97 1.34 1.1 1.26 1.2 1.875 2.25 2.15 1.5 1.63 1 1.875 2.56 1.78 1.75 1.94 0.52 2 2.2 1.78 1.25 1.36 0 1.25 1.36 0.5 1.875 2 0 1.375 1.44 0 1.375 1.67 0 1.1 1.33 0.25 1.1 1.23 0.9 2 2.98 0.5 1.75 1.9 0.15 1.25 1.82 1.3 1.25 1.46 0 1.2 1.4 0 1.875 2 0 1.5 1.68 0 1.375 1.58 1.3 1.2 1.44 0.4 0.73 0.77 0 1.38 1.56 0.6 1.75 1.92 0 1.5 1.7 1.68 2 2.25 0 1.2 1.56 1.58 1.75 2.3 0.7 1.75 1.83 0 1.25 1.45 0.2 2 2.4 0 1.38 1.9 0.1 2 2.29 0 1.62 1.74 0 1.88 2.26 0 1.75 2.19 0 1.5 1.71 0 2 1.88 1.44 1.88 2.05 0 0.91 0.96 0 1.75 2.03 0.25 2 2.03 0.25 1.5 1.66 0.6 1.75 2.15 0 1.5 1.72 0 1.38 1.18 0 2 2.65 2.45 1.62 1.92 1.4 2 2.23 0.53 1.62 1.6 0 1.5 1.69 0.81 1.5 1.54 0.51 1.75 1.9 1.38 1.25 1.4 0.54 1.62 1.95 1.74 1.75 2.03 2.03 1.75 2.5 1.22 You quite like STATS |
|
By:
Comments from TRADERS might be appreciated...good or bad!
|
|
By:
top2rated
Date Joined: 28 Feb 04 Add contact | Send message When: 28 Mar 12 09:24 Joined: Date Joined: 28 Feb 04 | Topic/replies: 2,948 | Blogger: top2rated's blog Kelly staking near useless in the real world..... One of its major flaws (and it has a few) is its accuracy is dependent on the accuracy of the 'edge' calculation which is just a guesstimate most of the time. Which is why you need to take Kelly as a starting point to build a staking system around. It's fairly obvious that if you are a successful gambler, you should be staking more when you back 2.00 on what you think is a 1.8 shot than when you back 2.00 on what you think is a 1.95 shot, even if these are just guesstimates. I'll start a new thread with a little maths problem related to Kelly that could be applicable to sports betting. |
|
By:
DFC, do you really think people read your posts ?
|
|
By:
Nice one Ken.
Made me laugh anyway. |
|
By:
Lex, I'll give you a simplified, exaggerated but not altogether outlandish example not from horse racing to demonstrate Kelly.
Let's say that Man City are home to Wolves. You make them a 1/7 shot or in other words give them an 87.5% chance of winning. The odds available are 1/3 (an implied 75% chance). By Kelly, you should bet half your bankroll. As the web tool on Ambion Research explains, 'on 87.5% of similar occasions, you would expect to gain 166.67 in addition to your stake of 500.00 being returned. But on those occasions when you lose, you will lose your stake of 500.00. Your bankroll will grow, on average, by about 4.82% on each bet.' You can feed the values into the Kelly formula fraction of bank to bet = bp-q / b that is, all over b where: b is the net odds received on the wager ("b to 1"); that is, you could win $b (plus the $1 wagered) for a $1 bet p is the probability of winning; q is the probability of losing, which is 1 − p and see the values come out the same. |
|
By:
And City are ahead for all the game and Wolves pop in a 93rd minute goal to steal a draw.
And bingo there goes 1/2 your bankroll, thanks to Kelly. No thank you very much. Imo as there is absolutely no way anybody can accurately forecast the result of any one football game, you have to stake( as others have suggested on here) as if you are at any one time about to go into a worst case extended losing streak. Now this won't maximize your returns ( particularly if analysed and calculated on a hindsight basis), but if you have the ability to predict results at a strike rate marginally better than the odds you get, you will see your bankroll compound up nicely over the years ( I stress over the years). In other words you will sort of become like a Warren Buffet of the sports gambling world. Not too shabby an achievement or goal, providing you really do have the discipline, skill and PATIENCE to do it year in year out. Very few in gambling make big fortunes overnight. And if they do, they probably don't keep them very long. The real pros just grind away with small margins and increasingly bigger stakes over a longish period of time. |
|
By:
FAFH - You know my position on Kelly and short prices. It doesn't work IMO and is way too aggressive a strategy. Small errors in judgement can lead to ridiculous overstakes putting bankrolls in jeopardy.
Kelly is great for 2/1 and above where occasional errors in judgement and bad luck won't hurt a punter with a normally consistent edge. |
|
By:
Many thanks askari1(and tobermory)
 |
|
By:
FATH, Kelly is optimally aggressive in growing yr bankroll. If you are more risk-averse, it makes sense to split your bankroll into something like ten Kelly pots.
I gave an example where (as I supposed) both the odds offered and 'your' estimate of the true probabilities were wrong. The stake was meant to be too aggressive. But if you would get 1/3 about City at home to Terry Connor-era Wolves, would you bet? Personally I don't use Kelly on horses other than in working put entry points into markets for a few reasons. First, price movements are exceptionally informative in the way horse markets develop. E.g. a horse from many small yards that tightens from 14-1 in the morning to 8-1 on bf at off time is probably at least an 8-1 chance. Next, the markets are too illiquid for me to get my stake on at what I would hope is the collapsing top of the price. I'd sooner overstake and try to trade (which almost always means arb) than try to wait for the 'silver bullet' one-off bet. Last, and related to this, the prices on horses are not fine enough (in their increments) to allow the kind of bet size Kelly would dictate. Oh, and of course, my pricing is not consistently good across all events--an effect no doubt owing in some measure to the information asymmetry that characterises horse markets below the very highest level. |
|
By:
Too many ifs for my liking Askari.
I will stick to the simple method of just trying to grow my bank slowly but surely over a long period of time, and the best method to do that is imo to bet a fixed percentage of your bank, irregardless of whether you feel that the odds on offer are particularly attractive ( or good value if you prefer) at any one time. I'm in no hurry to win, in the sense that I don't currently need " gambling" earnings to fund my living/entertainment expenses. Time is my friend not my enemy. I sometimes think that on here the converse seems to be the predominant factor. People want to win immediately and in large amounts. Not really feasible imo. You have to walk before you can run. |
|
By:
Get your edge established first with level stakes and then worry about sophitications like Kelly.
|
|
By:
Kelly is total cobblers as it ignores Probability Theory and you WILL go skint following it.
If you bet 50% winners at some stage Probablity Theory dictates that you WILL hit a losing streak of 10. You will definitely need an additional reserve of 10 bets to cover this. To include a 'comfort zone' you had better have another 10 bets in safe keeping. Therefore 20 to 30 bets to cover the bad times. Therefore your stake should be between 3-5% of your tank. |
|
By:
Under what circumstances CAN Kelly staking be used ?
|
|
By:
Kelly is the best staking method if you are consistent in yr estimate of ye edge. But of course you aren't.
I'd much rather Kelly-stake in very mature horse markets near the off than in the mornings, but the price I'd be taking will on average be worse and I'd have missed out on the value of market fluctuations/volatility to the trader. As for probability theory, it enshrines Kelly as the optimal way to bet. The chance of ten true 1/7 losing in a row is 1.164 X 10 to the minus 10, and your bankroll is lost before ten lose anyway. |
|
By:
I'd have missed out on the value of market fluctuations/volatility to the trader.
But you could use Kelly to manage your position throughout these fluctuations as well and get the best of both worlds? |
|
By:
Jabmast, I try to, or something similar, but my estimate of the true odds changes throughout the day according to price movements.
Let's take a young, talented sprint handicapper from the Willie Muir yard, now four, that missed the first part of its 3yo season. I'd expect that horse to be backed if genuinely fit and 'off'. If the price is worse three mins. to race time, I would want to have backed, reversed the bet and then have the choice of coming back in for a back. I'll try to be guided by Kelly, but my 'priors' are updating too quickly to stake optimally according to any system. |
|
By:
Ok. As I started this mess, perhaps we can run an experiment, where I'll post some selections and we can compare the success of level stakes (AT 5 PERCENT) against Kelly.
Askari1...could you be kind enough to calculate what stake Kelly would suggest to place??? I'm afraid I'm still confident in calculating it accurately. Two starting banks, each of 1000 points called 1) LEVEL and 2) KELLY. So to start with LEVEL=1000, KELLY =1000. BET 1 = Doncastor v Birmingham OVER 2,5 goals, taking the current available odds of 2.08. That's a level stakes bet of 50 points. Askari1, how much would the Kelly stake be?? I don't know what info you need ?? Odds are 2.08, but I estimate them to be 1.90 (that's a bit of a guess really though). I'd feel more confident calculating it using my average profit (edge)to date which is 16 percent. So what is Kelly stake? BET 2 = Murray to beat Nadal @ 2.28 LEVEL STAKE = 50 again. But what is Kelly Stake ?? Does it cause a problem if Bets 1 and 2 are running at the same time?? I'd estimate(guess) Murray's true odds to be 2.1. Average profit remains 16 percent. |
|
By:
Sorry, that should have read...I'm still NOT confident in calculating Kelly.
|
|
By:
Go to the page at Ambion Research, h-bomb, feed in the actual odds offered and your estimate of your chance of winnings (they make it even easier than asking for 'true odds') and you will see.
Split your bank into e.g. eight mini-rolls before you start, then follow optimal staking. That wd be my advice. |
|
By:
For the Doncaster-Birmingham game, you bet 8.77% of your total capital.
For the Murray-Nadal game, you bet 6.7% of your total capital. So you're betting more than 5% but not by much so far. Kelly is very aggressive and usually too aggressive. |
|
By:
For the Doncaster-Birmingham game, you bet 8.77% of your total capital.
For the Murray-Nadal game, you bet 6.7% of your total capital. So you're betting more than 5% but not by much so far. Kelly is very aggressive and usually too aggressive. |
|
By:
Ok, thanks for that askari1.
I'm going to chuck in a 3rd bank as well for high risk taking. This will use Kelly where I'm using my edge as calculated from previous bets ( that's 36%!! for football overs/unders bets, and 16% for all other sports bets) |
|
By:
Bet 1 won @ 2.08. Bet 2 was a non-runner.
I'll have to do all the calculations later. Tips for today: Bet 3: Wolves v Bolton. Over 2.5 goals @ 1.78 My rating: 1.66 Bet 4: Portsmouth v Burnley: Goals under 2,5 @ 1.95. My rating: 1.75. Bet 5: Wigan to beat Stoke @ 2,5. My rating 2.20. |
|
By:
Prices wrong way round. Wolves v Bolton. Over 2.5 goals @ 1.95. Portsmouth v Burnley: Goals under 2,5 @ 1.78.
|
|
By:
Bet 1: Winner @2.08
Level stake bet =50, Return (5%comm) = 101.3 Kelly stake = 8.77% = 87.70, Return = 139.98 SuperKelly stake (36% EDGE) = 38.37% = 383.7, Return = 776.65 BANKS AFTER BET 1 LEVEL(5%) = 1051.30 KELLY = 1052.28 SUPERKELLY = 1392.95 |
|
By:
BET 2: Void
Bet 3: Winner @ 1.95 Level stake = 50, Return = 95.12 Kelly stake = 18.39% = 193.48, Return = 368.09 SuperKelly stake = 34.32% = 478, Return = 909.39 BANKS AFTER BET 3 LEVEL(5%) = 1096.42 KELLY = 1226.89 SUPERKELLY = 1824.34 |