Deviation Posted February 27, 2009 Share Posted February 27, 2009 I have a greyhound system with an 85.6% SR and average odds of 6.31. How do I calculate the bank size required, given a fixed liability of 1 unit per bet? Thanks. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 27, 2009 Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying required for what ? Quote Link to comment Share on other sites More sharing options...
Deviation Posted February 27, 2009 Author Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying A long term staking plan, with a 1% chance of exhausting the bank in any one year. There are about 1600 bets per month on average. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 27, 2009 Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying 1600 bets per month, sure ? that would be an average of over 50 bets per day. 85 % at average odds 6.31 ..... you sure you don't want to refrase that ? Quote Link to comment Share on other sites More sharing options...
Deviation Posted February 27, 2009 Author Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying I'm laying a dog in many races, so yes on average there are 53 bets per day. 14.4% of the selections actually win their race, which are losses when laying of course. Sorry, I thought lay system strike rates were quoted the other way round. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 27, 2009 Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying So you have a strike rate of winning bets of 85.6 % and the average odds on those winning bets is 1/5.31 = 1.1883 at a strike rate of 85.6% you need minimal odds of 1.1682 to break even so your expected yield would be about 2% If you want to stake at fixed liability then it's quite important to know the average odds of the winning bets, not the average of all bets or losing bets. The winning ones as that is where you need to make the profit and that is where the ups and down are. Quote Link to comment Share on other sites More sharing options...
Deviation Posted February 27, 2009 Author Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying The average betfair odds traded for the winning bets is 6.425, which converts to 1.1843 I think. The yield (profit/number of bets?) is 1.3% after 5% commission. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 27, 2009 Share Posted February 27, 2009 Quote Link to comment Share on other sites More sharing options...
Deviation Posted February 27, 2009 Author Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying First of all the question " how big should the bank be ? " makes no sense. There is no right or wrong answer and there is nothing to calculate. The size of the bank is however much you are willing to commit. However much you can afford' date=' and are willing, to lose.[/quote'] Sorry, I should have worded it better. I meant how many units will I need. My results are based on just over 22 months' worth of data.betfair.com prices, and 35,700 bets. I have live prices captured for 4,250 bets spanning over 4 months. The live prices seem to be in the same ball park as the data.betfair.com prices, but the yield is a little higher :). How confident can I be of the yield at this stage? Is it solely down to how statistically significant the results are based on live prices? Or can I take comfort from the 35,700 bet sample? Thank you very much DP for the time you've spent on this. It is much appreciated. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 27, 2009 Share Posted February 27, 2009 Re: Calculating bank size for fixed liability laying How confident can I be of the yield at this stage? That is indeed the question. Is there a chance that you start out with a positive yield of +2% but then after some time it flips to -3% for a while. The effect would be that you base a stake on the +2% and make great profits for 6 months but then go bancrupt inside a month. Remember once bancrupt there's no way to recover. So to avoid this, as you simply don't know when the bad patch migh occur, you could start with an extremely low risk stake, allowing you to survive the bad patch. But then you won't be making much profit, so will you still be motivated :\ 't is a bitch innit :lol I just don't know how reliable you can call the backtesting at this point. Also i can't estimate how robust the yield is, i.e. will it stay positive. But if you're expecting +1.3% then there's not much room for error is there. You will very likely have periods of negative yields so i'd say caution is in order. Real live testing with real bets and cash would be the next step with something like 0.5% of the bank as fixed liability per bet. So if the bank is 1000 pounds you risk 5 pounds per bet, at odds of over 6.00 a winner will get you less than 1 pound profit. Doesn't sound like very motivating but remember that with the number of bets at over 50 per day you'll see the bank grow very nicely.... ... providing the yield stays positive :lol Quote Link to comment Share on other sites More sharing options...
Deviation Posted February 28, 2009 Author Share Posted February 28, 2009 Re: Calculating bank size for fixed liability laying These are the results based on data.betfair.com prices: strike rate: 0.844 avg odds: 6.27 avg win odds: 6.43 profit: -6.2 bets: 1458 yield: -0.004 2007/04 strike rate: 0.853 avg odds: 6.26 avg win odds: 6.36 profit: 17.1 bets: 1646 yield: 0.010 2007/05 strike rate: 0.858 avg odds: 6.22 avg win odds: 6.33 profit: 23.4 bets: 1565 yield: 0.015 2007/06 strike rate: 0.870 avg odds: 6.21 avg win odds: 6.33 profit: 45.3 bets: 1611 yield: 0.028 2007/07 strike rate: 0.853 avg odds: 6.18 avg win odds: 6.30 profit: 23.8 bets: 1583 yield: 0.015 2007/08 strike rate: 0.851 avg odds: 6.33 avg win odds: 6.49 profit: 3.7 bets: 1609 yield: 0.002 2007/09 strike rate: 0.848 avg odds: 6.42 avg win odds: 6.51 profit: 3.1 bets: 1779 yield: 0.002 2007/10 strike rate: 0.855 avg odds: 6.36 avg win odds: 6.49 profit: 11.2 bets: 1619 yield: 0.007 2007/11 strike rate: 0.865 avg odds: 6.10 avg win odds: 6.21 profit: 45.4 bets: 1403 yield: 0.032 2007/12 strike rate: 0.866 avg odds: 6.50 avg win odds: 6.63 profit: 27.5 bets: 1759 yield: 0.016 2008/01 strike rate: 0.841 avg odds: 6.41 avg win odds: 6.56 profit: -18.4 bets: 1625 yield: -0.011 2008/02 strike rate: 0.863 avg odds: 6.34 avg win odds: 6.44 profit: 27.9 bets: 1620 yield: 0.017 2008/03 strike rate: 0.864 avg odds: 6.39 avg win odds: 6.50 profit: 27.2 bets: 1561 yield: 0.017 2008/04 strike rate: 0.837 avg odds: 6.41 avg win odds: 6.48 profit: -21.1 bets: 1528 yield: -0.014 2008/05 strike rate: 0.864 avg odds: 6.25 avg win odds: 6.35 profit: 32.5 bets: 1471 yield: 0.022 2008/06 strike rate: 0.868 avg odds: 6.43 avg win odds: 6.52 profit: 34.2 bets: 1582 yield: 0.022 2008/07 strike rate: 0.857 avg odds: 6.38 avg win odds: 6.49 profit: 17.9 bets: 1605 yield: 0.011 2008/08 strike rate: 0.862 avg odds: 6.45 avg win odds: 6.54 profit: 70.0 bets: 1697 yield: 0.041 2008/09 strike rate: 0.857 avg odds: 6.41 avg win odds: 6.51 profit: 15.6 bets: 1754 yield: 0.009 2008/10 strike rate: 0.851 avg odds: 6.21 avg win odds: 6.30 profit: 14.0 bets: 1619 yield: 0.009 2008/11 strike rate: 0.845 avg odds: 6.24 avg win odds: 6.42 profit: -3.8 bets: 1523 yield: -0.002 2008/12 strike rate: 0.857 avg odds: 6.07 avg win odds: 6.17 profit: 65.9 bets: 1718 yield: 0.038 2009/01 strike rate: 0.852 avg odds: 6.23 avg win odds: 6.35 profit: 3.8 bets: 372 yield: 0.010 2009/02 ---------------------------------------------------------------------------------------------------------------------------------- strike rate: 0.856 avg odds: 6.31 avg win odds: 6.43 profit: 460.1 bets: 35707 yield: 0.013 Total - The profit is in units after 5% Betfair commission, and the stake is 1 unit liability. - To my untrained eye this looks quite stable. Where is the risk? Is it that the yield will drop for a few consecutive months and wipe out the bank? Is that what a Monte Carlo simulation will show? - Should I tune the system to increase/stabilise the yield? - Why aren't the drops in yield simply the effects of deviation? - This is my first real system (maybe). - I would like to get a monthly income out of it rather than build up the bank, for the time being. - A subject for another thread, but would I be better off with level stakes as I'm a beginner? My estimate is the bank would have to be 200 units, but the profit is a lot more volatile and was negative for 2 months at one point. Sorry for all the questions. I have put a lot of time in (rightly so), but of course it takes months to learn what you can in a hour on here. Thank you so much for your time, once again. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted February 28, 2009 Share Posted February 28, 2009 Re: Calculating bank size for fixed liability laying the stake is 1 unit liability Just checking we're talking about the same thing here: " 1 unit liability " at decimal odds of 6.00 means: if the selection loses, the bet wins and you win: 0.20 if the selection wins, the bet loses and you lose: 1.00 so you would enter 0.20 in the input field at odds of 6.00 to have a liability of 1.00 Where is the risk? Is it that the yield will drop for a few consecutive months and wipe out the bank? yes Is that what a Monte Carlo simulation will show? The Monte Carlo simulation shows 2 things: One is that even with a stake of only a few % of the bank, the risk of bancrupty is real and cannot be dismissed as trivial. The second thing is that a positive or negative yield has an exponential effect on the progress of a bank.The more negative the yield becomes the faster it heads to bancruptcy. So with a stake that's too high a bad patch, even thou it may be temporary, can see you go bancrupt in a short period of time. Should I tune the system to increase/stabilise the yield? Well yes, as with any betting system. But it's exactly this which is so notoriously difficult to do. Your next step is simple: get a series of bets in at minimal stake so you have a real time record. Back testing is fine, recording live prices even better but there is no substitute for real live action data. The reason for this is all about timing. You simply don't know you can get the same prices as you got in backtesting, or even live recording, until you actually put bets on. Quote Link to comment Share on other sites More sharing options...
Deviation Posted March 1, 2009 Author Share Posted March 1, 2009 Re: Calculating bank size for fixed liability laying Just checking we're talking about the same thing here: " 1 unit liability " at decimal odds of 6.00 means: if the selection loses, the bet wins and you win: 0.20 if the selection wins, the bet loses and you lose: 1.00 so you would enter 0.20 in the input field at odds of 6.00 to have a liability of 1.00 Yes, that's it. There are small rounding errors, of course. So with a stake that's too high a bad patch, even thou it may be temporary, can see you go bancrupt in a short period of time. Why am I not better off with a 60 unit bank, which may make for example 95 - 160 units in 6 months, judging on past results (I know, I know!). Then it goes horribly wrong and I lose all my original 60 units. Isn't that more profitable than splitting the bank into 200 units and (hopefully) almost never going bankrupt? I need to maximise an income from this as soon as possible. I have spent an awful long time trying to make an income from home, and am climbing the walls!! I don't mind some risk as long as it's calculated. Fixed stakes give an average yield of 4.5%. Is this likely to be a better path for me? Or shall I put it in another thread? Your next step is simple: get a series of bets in at minimal stake so you have a real time record. Yes, already been doing this, thanks. Quote Link to comment Share on other sites More sharing options...
Datapunter Posted March 1, 2009 Share Posted March 1, 2009 Re: Calculating bank size for fixed liability laying Why am I not better off with a 60 unit bank, which may make for example 95 - 160 units in 6 months, judging on past results (I know, I know!). Then it goes horribly wrong and I lose all my original 60 units. Isn't that more profitable than splitting the bank into 200 units and (hopefully) almost never going bankrupt? You wouldn't make the same profit would you. The more risk the more profit, in the end a matter of choice. Only you can determine what level of risk, size of bank, stake per bet, you find acceptable. No-one can do that for you. I need to maximise an income from this as soon as possible. I have spent an awful long time trying to make an income from home, and am climbing the walls!! I don't mind some risk as long as it's calculated. Some would say this is not something to base an income on. But i'm not here to preach. I'll leave that to someone else. Quote Link to comment Share on other sites More sharing options...
Deviation Posted March 3, 2009 Author Share Posted March 3, 2009 Re: Calculating bank size for fixed liability laying You wouldn't make the same profit would you. I'm not quite sure what you mean by this. :unsure I feel like I should split my funds equally into 4 bankrolls as I'm a beginner, and use the first quarter on this system. I think this is the prudent thing to do. - So given that, wouldn't I be better off if on average I gain (95+60)/2=127 units and lose a 60 unit bank every 6 months, rather than make 127 units and retain a 200 unit bank? - Or should I use all the funds as one bankroll and play it very safe? - Or is there no right and wrong way, just what's best for me and my situation? Some would say this is not something to base an income on. Do you mean it won't provide a reliable and/or a sufficient income? Thanks once again, DP, I'm so grateful for your help. :notworthy I'll lay off the questions now and start a new thread to give you a rest. :) Quote Link to comment Share on other sites More sharing options...
Datapunter Posted March 3, 2009 Share Posted March 3, 2009 Re: Calculating bank size for fixed liability laying You wouldn't make the same profit would you. Stake/liability (as in risk to lose) 1 unit of 60 is more than 1 unit of 200 , ( if you split the same money bank into either 60 or 200 units ) that 1 unit gets you more profit per bet. This whole thing is about balancing risk with potential profit. Or is there no right and wrong way, just what's best for me and my situation? Once all aspects have been examined all that remain is a personal choice, the rest becomes history. Do you mean it won't provide a reliable and/or a sufficient income? Both, by the very nature of the business any profit will never ever be reliable, there will always be a large degree of faith/karma/luck. But then again some will argue that having a steady job is just as unreliable. After all, who predicted the credit crunch and how many people have lost their "steady" job because of it. If you are dependend to pay the rent from this then it is an absolute requirement to reduce the risk as much as possible. The only thing you have actual control over is the amount you stake, or the amount of the total bank. If you have a huge bank then you can reduce the risk to close to 0 but then again if you have a huge bank there's loads of other different ways to get an income much less risky. So if your bank is limited and you still want a very low risk, you need to stake little. But that only gets you small profits. I'm not going to bore you with a lot of been there done that but my current situation is that i have a part-time job, 20 hour a week, as a postman. Gets me out of the house, some fresh air and excercise and i get paid as well. The salary from that covers the rent and utilities. Anything else comes from various betting related activities. Works for me. Good luck, now go talk to some other people :lol Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.