1. #1
    PavelSpr
    PavelSpr's Avatar Become A Pro!
    Join Date: 07-19-13
    Posts: 65
    Betpoints: 534

    Money management based on winning percentage

    Signed up here to hopefully catch some fresh and smart opinions, maybe some help as I believe around here are serious bettors with some knowledge, better than mine.

    What do I want to ask for? Let's start from description of what I do. During the last 2 years I managed to achieve over 60% winning percentage in around 600 bets only in soccer matches using asian handicaps lines with average decimal odd minimum 1.85. As I really respect the deep knowledge of what I invest on so I wager basing not only on clear statistics and maths but mostly on real understanding of player's mood after recent results, home-away factors etc. all this stuff which is important for real player and the rest of the team, my dream from childhood was to be a professional soccer player so maybe that's the reason. But my main advantage I think is that I am interested only in semi-professional leagues which are covered maybe only by local TVs but still in some part, not all of them so usually bookmakers are very chaotic setting the fair lines and the average number of oddsmaker's mistakes is higher than in the popular leagues. But all this description is not the point. The point is that I've learned so much in last months that finally I see how many possible additional income I lost because of lower bank and sometimes very risky money management.

    So now I have the question for you guys, I've read many possible money management system but still not decided which is the best one according of what I wrote above. I thought about the simple one which looks like that:
    1-2% of bank - mostly based on feelings because not facts are proved
    3-5% of bank - based on league's knowledge
    6-9% of bank - based on some additional juicy and proved information (like internal team's problems) and knowledge

    and with flexible stake according to current bank's level

    Waiting for your opinions, thanks in advance

  2. #2
    easyliving
    easyliving's Avatar Become A Pro!
    Join Date: 06-25-12
    Posts: 8,876
    Betpoints: 11

    betting 6-9% of your BR is a huge no no in my opinion but if it works for you go with it. In my opinion eliminate the bets based on feelings and flat bet everything else. If your bankroll is large enough betting 3% on each game should return you a nice profit if you are truly hitting at a 60% rate at 1.85.

  3. #3
    HeeeHAWWWW
    HeeeHAWWWW's Avatar Become A Pro!
    Join Date: 06-13-08
    Posts: 5,487
    Betpoints: 578

    Firstly, and you won't want to hear this: 60% means a heavy dose of luck. You won't continue that going forward, not even close. Not trying to be abusive there, that's just simply reality. The problem is that you don't get to know your true stats until you're up into several thousand bets - variance is a bitch.

    The reason I start on that: your estimated edge should determine your stake size. If you over-estimate that edge based on past performance, you'll over stake ......... and over staking is a killer. That's not some glib rule of thumb either - if you bet more than double Kelly, you lose long term even with a positive edge (ie negative expected bankroll growth despite +EV).

  4. #4
    easyliving
    easyliving's Avatar Become A Pro!
    Join Date: 06-25-12
    Posts: 8,876
    Betpoints: 11

    Quote Originally Posted by HeeeHAWWWW View Post
    Firstly, and you won't want to hear this: 60% means a heavy dose of luck. You won't continue that going forward, not even close. Not trying to be abusive there, that's just simply reality. The problem is that you don't get to know your true stats until you're up into several thousand bets - variance is a bitch.

    The reason I start on that: your estimated edge should determine your stake size. If you over-estimate that edge based on past performance, you'll over stake ......... and over staking is a killer. That's not some glib rule of thumb either - if you bet more than double Kelly, you lose long term even with a positive edge (ie negative expected bankroll growth despite +EV).
    how big of a sample size does one need to see variance? what is several thousand? is it 2k? 3k? maybe more?

  5. #5
    HeeeHAWWWW
    HeeeHAWWWW's Avatar Become A Pro!
    Join Date: 06-13-08
    Posts: 5,487
    Betpoints: 578

    Quote Originally Posted by easyliving View Post
    how big of a sample size does one need to see variance? what is several thousand? is it 2k? 3k? maybe more?
    Depends on the RoI. The higher it is, the less a sample size is needed to prove to a particular level of certainty.

    Z-score is good for this.

  6. #6
    PavelSpr
    PavelSpr's Avatar Become A Pro!
    Join Date: 07-19-13
    Posts: 65
    Betpoints: 534

    Quote Originally Posted by easyliving View Post
    betting 6-9% of your BR is a huge no no in my opinion but if it works for you go with it. In my opinion eliminate the bets based on feelings and flat bet everything else. If your bankroll is large enough betting 3% on each game should return you a nice profit if you are truly hitting at a 60% rate at 1.85.
    6-9% appear very rarely, only in case of really some gifts in odds, so far my average stake per 1 bet is between 3-4% as most of the bets are from the range of 3-5% (bets from 1-2% are for Under/Over and friendly matches). So I think that it's almost the same if I would use flat stakes but I compared these two and flat staking gave less risk but also less profit with similar ROI like it was in the 1st one

    Quote Originally Posted by HeeeHAWWWW View Post
    Firstly, and you won't want to hear this: 60% means a heavy dose of luck. You won't continue that going forward, not even close. Not trying to be abusive there, that's just simply reality. The problem is that you don't get to know your true stats until you're up into several thousand bets - variance is a bitch.

    The reason I start on that: your estimated edge should determine your stake size. If you over-estimate that edge based on past performance, you'll over stake ......... and over staking is a killer. That's not some glib rule of thumb either - if you bet more than double Kelly, you lose long term even with a positive edge (ie negative expected bankroll growth despite +EV).
    So seems that I need at least few more years to discover this variance as per month I try to keep the number of bets to 30-35 but average it's 25-30. And you r writing about over-estimating - I think my staking system won't let me to do this because it's simple - if a team A is weakened and I know that replacers will shit in pants in that game and probably a Team B will eat them without problem is it still would be over-estimating?

  7. #7
    HeeeHAWWWW
    HeeeHAWWWW's Avatar Become A Pro!
    Join Date: 06-13-08
    Posts: 5,487
    Betpoints: 578

    Quote Originally Posted by PavelSpr View Post
    So seems that I need at least few more years to discover this variance as per month I try to keep the number of bets to 30-35 but average it's 25-30.
    Sort of: google z-score. What number of bets counts as "sufficient sample size" isn't arbitrary, it depends on your results, and the distribution of stake sizes. Once you have a z-score of 2+, you have a decent reason to believe it's not just luck, although it still can be. Once you have 3+ you can be fairly sure (although still possible to be luck).

    There's a separate element that's less easy to judge, either via statistics or subjectively: year to year changes in the sport. A method/model might do well one season, then not the next, because some underlying element in the sport has changed (or "evolved" might be a better word). My understanding of this is that the structure of a model needs to be fluid enough to adapt, but that's complex beyond belief.



    And you r writing about over-estimating - I think my staking system won't let me to do this because it's simple - if a team A is weakened and I know that replacers will shit in pants in that game and probably a Team B will eat them without problem is it still would be over-estimating?

    OK - here's a different way of looking at stake sizing: at some point you will have a nightmare bad streak. The laws of probability make this inevitable, regardless of how good your picks are. You might lose ten in a row, or 17 of 20. I've even had worse.

    Let's take the latter. Can you cope with a few days where you go a net negative 14? At 5% stake, that's going to wipe out 70% of your roll.

    At bare minimum, half your current stake sizes, and preferably set to 1/3. They are far, far, far, far too large and when the badluck-hammer hits you, it's going to kill your bankroll.

  8. #8
    PavelSpr
    PavelSpr's Avatar Become A Pro!
    Join Date: 07-19-13
    Posts: 65
    Betpoints: 534

    HeeeHAWWWW, thanks for all and seems that as you wrote, better would be half my stakes to only 3-level system, it's more than fair in my case

    I've never heard about z-score. Looked about it in the network but didn't manage to use it in a proper way, perhaps need a time to understand this tool

  9. #9
    bihon
    bihon's Avatar Become A Pro!
    Join Date: 11-03-09
    Posts: 731

    Quote Originally Posted by HeeeHAWWWW View Post
    Firstly, and you won't want to hear this: 60% means a heavy dose of luck. You won't continue that going forward, not even close. Not trying to be abusive there, that's just simply reality.
    60% on 1.85 odds means 5-6% edge, which is big but not really dramatically lucky.

    Being not so huge fan of Kelly, my main concern for determining stake is solely acceptable risk of ruin. The calc could be found around.

  10. #10
    HeeeHAWWWW
    HeeeHAWWWW's Avatar Become A Pro!
    Join Date: 06-13-08
    Posts: 5,487
    Betpoints: 578

    Quote Originally Posted by bihon View Post
    60% on 1.85 odds means 5-6% edge, which is big but not really dramatically lucky.
    Yeah, sure, was assuming on 50/50 stuff. Win ratio on anything else is meaningless because it's totally dependent what odds you're betting.

  11. #11
    Kaplan
    Kaplan's Avatar Become A Pro!
    Join Date: 01-15-11
    Posts: 165
    Betpoints: 865

    Quote Originally Posted by HeeeHAWWWW View Post
    Sort of: google z-score. What number of bets counts as "sufficient sample size" isn't arbitrary, it depends on your results, and the distribution of stake sizes. Once you have a z-score of 2+, you have a decent reason to believe it's not just luck, although it still can be. Once you have 3+ you can be fairly sure (although still possible to be luck).
    Can You further elaborate, for the sake of my simple mind< on this comment?

  12. #12
    Sawyer
    Sawyer's Avatar Become A Pro!
    Join Date: 06-01-09
    Posts: 7,592
    Betpoints: 6650

    Betting more then %5 of your bankroll is suicide. My advice for recreational players, never bet more then %5 of your bankroll.

    As you get more professional and your bank grows, I suggest to be more conservative. First make it %4, then %3, then %2 and finally %1 when you're a fukkin real pro.

    You can"t bet %5 of your bank when your bankroll is 50k.

  13. #13
    statnerds
    Put me in coach
    statnerds's Avatar Become A Pro!
    Join Date: 09-23-09
    Posts: 4,047
    Betpoints: 103

    Ditto on just about everything H Hawwww wrote.

    another important factor i would need to consider is when you are placing these wagers. are you betting openers, mid market or closers?

    big difference where luck is concerned when you are leading the Market or following it.

  14. #14
    Kaplan
    Kaplan's Avatar Become A Pro!
    Join Date: 01-15-11
    Posts: 165
    Betpoints: 865

    Is a z score of 3 (+) meaningful on a small sample though? I am working with a sample of only 400 games. Does a z score of 3 or more still indicate strength?

  15. #15
    chunk
    chunk's Avatar Become A Pro!
    Join Date: 02-08-11
    Posts: 805
    Betpoints: 19150

    Quote Originally Posted by statnerds View Post
    Ditto on just about everything H Hawwww wrote.

    another important factor i would need to consider is when you are placing these wagers. are you betting openers, mid market or closers?

    big difference where luck is concerned when you are leading the Market or following it.
    Seriously? There are wagers that you want to hit openers and there are others that you don't. You should know this....you've been around long enough.

  16. #16
    HUY
    HUY's Avatar Become A Pro!
    Join Date: 04-29-09
    Posts: 253
    Betpoints: 3257

    Quote Originally Posted by Kaplan View Post
    Is a z score of 3 (+) meaningful on a small sample though? I am working with a sample of only 400 games. Does a z score of 3 or more still indicate strength?
    A z-score takes the sample size into account.

Top