1. #1
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    NBA Halftime Lines System... Oppinions and Help MUCH APPRECIATED!

    OK... I don't know how many of you go into the NBA and other betting forums during the course of the year... but if you didn't... I started a betting system last year that I came up with through a lot of research... These were the results pre-all-star game... and I had to stop due to the demands of my Junior year of College to the chagrin of many people who were following me...

    Season (81-61-5
    ) 57%

    3 Star Plays (21-9-1) 70%

    2 Star Plays (36-33-2) 52%

    1 Star Plays (24-19-2) 56%

    +35.8 Units

    Now I can't divulge too much of the information about how I make my picks... but it has to do with the opening lines... the halftime lines... who is home... who is away... etc. etc... I created a spread sheets with stats based upon the 2007 and 2008 NBA seasons to bet on the 2009 NBA season...

    If a play showed 65% or higher in my system I would make it a 3 Star Play... 60-65% would be a 2 Star Play... and 55-60% would be a 1 star play...

    Now my questions for the think tank are first of all based on the results from last year would it be more beneficial to bet purely my strongest plays... or should I continue to bet the smaller plays to try to eliminate losses if I have bad nights with my big plays?

    Second... is 2 years of compiled stats enough to base a system on... or how many years do I have to go back until I truly find something?

    Third... if I start with a $1000 bank roll... how much would u suggest laying on plays? and should I be betting a pure percent of my money on plays... or a set amount?

    Lastly what is the percentage of bets you need to win at normal vig to break even... something like 52.38% I think... verification of that would be great though...

    Any help or advice for the system or answers to any of these questions would be greatly appreciated... thanks so much and I will be on here to discuss more if I do get some answers...

    THANKS!

  2. #2
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Nebody alive in these forums???

  3. #3
    sharpcat
    sharpcat's Avatar Become A Pro!
    Join Date: 12-19-09
    Posts: 4,516

    1) I am not sure how you came up with this system or how you backtested it but if you are confident in your system and you did not data mine your system than last years sampling of results is way too small the poor or good record of each play type could easily be the result of variance.

    2) It depends what kind of stats you are using, If they are team specific than no 200 games is not enough data if they are league specific than yes 3,000 games is likely enough though more is better.

    3) This all comes down to how much of an edge you have I would recommend using Kelly if you can accurately estimate your edge, if not I would just flat bet at 1-2% of your bankroll.

    4) 52.38% is correct.
    Favorites= Line/(100+Line)= 110/(100+110)= 110/210= 52.38%
    Dogs= 100/(100+Line)=100/(100+110)=100/210= 47.62%

  4. #4
    roasthawg
    roasthawg's Avatar Become A Pro!
    Join Date: 11-09-07
    Posts: 2,990

    If you feel a play is +ev then bet it.

  5. #5
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by sharpcat View Post
    1) I am not sure how you came up with this system or how you backtested it but if you are confident in your system and you did not data mine your system than last years sampling of results is way too small the poor or good record of each play type could easily be the result of variance. 2) It depends what kind of stats you are using, If they are team specific than no 200 games is not enough data if they are league specific than yes 3,000 games is likely enough though more is better. 3) This all comes down to how much of an edge you have I would recommend using Kelly if you can accurately estimate your edge, if not I would just flat bet at 1-2% of your bankroll. 4) 52.38% is correct. Favorites= Line/(100+Line)= 110/(100+110)= 110/210= 52.38% Dogs= 100/(100+Line)=100/(100+110)=100/210= 47.62%
    They are not team specific... but I might as well see if I can go back another season and see if it continues to work...

    Would you mind explaining Kelly to me... I'm sure it might be a bit complicated... but... with my system I basically get a pure percentage number... so for a particular game I may get a 63.4% advantage... or might get a 54.8%... if I know the exact percentages can I use the Kelly system to figure out exactly how much of my bankroll to bet? and how much do u suggest I bet on the smallest percentage advantage I will place a play on... which is probably somewhere around 54%... and how much should my max bet be no matter how much the anticipated advantage???

    Thanks so much in advance!

  6. #6
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by roasthawg View Post
    If you feel a play is +ev then bet it.
    True True... just got mighty close to that 52.38% number on the plays as the season moved on... so I wondered if continuing those plays was worth it...

  7. #7
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Sharpcat... what does data mining my system mean as well... thanks!!!

  8. #8
    sharpcat
    sharpcat's Avatar Become A Pro!
    Join Date: 12-19-09
    Posts: 4,516

    SBR tools Kelly calculator is the easiest way, simply type in your expected win% and the line you are getting and hit calculate and it will tell you what % of your bankroll to bet to maximize growth. The calculator is set for full kelly which you can change to 1/2 or 1/4 kelly if full kelly is to aggressive for you. Also make sure you readjust your bankroll after each bet because you want to assume that any pending wagers could be losses and therefore should not be consider part of your balance.

    An example of data mining would be if while looking at a set of data you notice a trend and than test that data for that trend. If you notice a trend in a set of data you have to use a completely new set of data to test your hypothesis or else your results will be flawed. After testing your hypothesis against a separate set of data that was used to develop your system you should have a smaller set of data that has not been used yet to back test your system. Data mining can cause a -EV system to appear to be +EV.

  9. #9
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Don't know how to find the SBR tools... how to do that would be great...

    and yea didn't use data mining... that would be absolutely stupid logic... but I guess I could see someone doing that... when I find a trend over the course of a year or 2 I use a past year to see if that information holds up... only logical way to do it obviously... thanks

  10. #10
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    found the kelly calculator... sadly no idea how to use it now... lol... its saying 100% of my bankroll... so obviously not doing it right... I will try to figure it out... or some advice on how to put the info in would be great!

  11. #11
    MarketMaker
    MarketMaker's Avatar Become A Pro!
    Join Date: 07-19-10
    Posts: 44

    Quote Originally Posted by Brp27345 View Post
    found the kelly calculator... sadly no idea how to use it now... lol... its saying 100% of my bankroll... so obviously not doing it right... I will try to figure it out... or some advice on how to put the info in would be great!
    Have to admit this made me laugh.

  12. #12
    Flying Dutchman
    Floggings continue until morale improves
    Flying Dutchman's Avatar Become A Pro!
    Join Date: 05-17-09
    Posts: 2,467
    Betpoints: 759

    Brp, you are not allowing PMs: I also do NBA and will answer privately, but not in an open forum.

  13. #13
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by MarketMaker View Post
    Have to admit this made me laugh.
    yea man... was funny... u got any advice on how to use it?

  14. #14
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by Flying Dutchman View Post
    Brp, you are not allowing PMs: I also do NBA and will answer privately, but not in an open forum.
    didn't know I wasn't... I will try to allow them... and if I can't someone letting meknow how to do that would be appreciated as well!

  15. #15
    Wojo
    Wojo's Avatar Become A Pro!
    Join Date: 03-19-10
    Posts: 1,764
    Betpoints: 9513

    I agree with what was stated earlier. You have too small of data set to make rational decisions from what to bet and for how much.

    52.38% is correct.

    In my mind, you have done data mining when you stated you based your system off of the previous two seasons. However, IMO, basing it on only two seasons is not enough.

    The NBA is one of the best sports to bet, IMO. But, with such vague information, I have no idea on the validity of your system. But, I sincerely wish you the best of luck. (I would worry more about college than betting the NBA at this point of your life, FWIW)

    Wojo

  16. #16
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by Wojo View Post
    I agree with what was stated earlier. You have too small of data set to make rational decisions from what to bet and for how much. 52.38% is correct. In my mind, you have done data mining when you stated you based your system off of the previous two seasons. However, IMO, basing it on only two seasons is not enough. The NBA is one of the best sports to bet, IMO. But, with such vague information, I have no idea on the validity of your system. But, I sincerely wish you the best of luck. (I would worry more about college than betting the NBA at this point of your life, FWIW) Wojo
    Well I mean when u find a system it has to be based off of something... after finding the stats from those 2 years I tested them on a different season... last year... and continued to see positive results as you can see by my record... so I'm pretty sure I'm not data mining...

    How many seasons do you think I need to go through until the info is legit? I mean 2 seasons is about 2,500 games... and then u add on last year I am up near 4,000... but I think I might go back to '07 for some more data testing if I have the time...

    And I know worrying about college is more important... get a 3.8 G.P.A. going into my Senior year though... so not struggling there... don't worry... gambling is not taking over my life... just putting as much effort as I can into it on my summer off... and looking for some help on collecting as much outside info/adive as I can...

    Thanks for your input!

  17. #17
    Wojo
    Wojo's Avatar Become A Pro!
    Join Date: 03-19-10
    Posts: 1,764
    Betpoints: 9513

    Brp,
    I was considering that you only have around 150 plays that fell within your system and just 30 that were top-rated. That is too small of sample, IMO, to make any calculated decisions on. Also, thanks for realizing I wasn't trying to bash you about my college pick. I wish I would have had a 3.8 going into my senior year. Good job!

    GL,
    Wojo

  18. #18
    brxbmbers42
    brxbmbers42's Avatar Become A Pro!
    Join Date: 07-26-10
    Posts: 4,312

    Quote Originally Posted by sharpcat View Post
    SBR tools Kelly calculator is the easiest way, simply type in your expected win% and the line you are getting and hit calculate and it will tell you what % of your bankroll to bet to maximize growth. The calculator is set for full kelly which you can change to 1/2 or 1/4 kelly if full kelly is to aggressive for you. Also make sure you readjust your bankroll after each bet because you want to assume that any pending wagers could be losses and therefore should not be consider part of your balance.

    An example of data mining would be if while looking at a set of data you notice a trend and than test that data for that trend. If you notice a trend in a set of data you have to use a completely new set of data to test your hypothesis or else your results will be flawed. After testing your hypothesis against a separate set of data that was used to develop your system you should have a smaller set of data that has not been used yet to back test your system. Data mining can cause a -EV system to appear to be +EV.

  19. #19
    brxbmbers42
    brxbmbers42's Avatar Become A Pro!
    Join Date: 07-26-10
    Posts: 4,312

    What is a +ev bet?

  20. #20
    wrongturn
    Update your status
    wrongturn's Avatar Become A Pro!
    Join Date: 06-06-06
    Posts: 2,228
    Betpoints: 3726

    Brp, to make it simple, you can bet 3%,2%,1% of bankroll on your 3*,2*,1* plays, and recalculate daily. I think using the optimal kelly is not important at beginning, the important thing for you is to verify that your system is not a short term fluke.

  21. #21
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by brxbmbers42 View Post
    What is a +ev bet?
    + Estimated Value I believe... It means that a given bet should have an estimated positive return value over a given period of time...

  22. #22
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by wrongturn View Post
    Brp, to make it simple, you can bet 3%,2%,1% of bankroll on your 3*,2*,1* plays, and recalculate daily. I think using the optimal kelly is not important at beginning, the important thing for you is to verify that your system is not a short term fluke.
    Yea I think that might not be a bad idea to start with.. thanks for the advice... do u think that taking all the info and going game by game in the 2007 season would be the best way to attempt to verify the results before the season starts?

  23. #23
    sharpcat
    sharpcat's Avatar Become A Pro!
    Join Date: 12-19-09
    Posts: 4,516

    Quote Originally Posted by wrongturn View Post
    Brp, to make it simple, you can bet 3%,2%,1% of bankroll on your 3*,2*,1* plays, and recalculate daily. I think using the optimal kelly is not important at beginning, the important thing for you is to verify that your system is not a short term fluke.
    ^^This or flat betting at 2% until you have enough wagers to truly know if you really have an edge or not.

    Kelly is suggested assuming that you are accurately predicting your edge if you believe that your edge is 7% when you really have no edge at all Kelly can be suicide.

  24. #24
    wrongturn
    Update your status
    wrongturn's Avatar Become A Pro!
    Join Date: 06-06-06
    Posts: 2,228
    Betpoints: 3726

    Quote Originally Posted by Brp27345 View Post
    Yea I think that might not be a bad idea to start with.. thanks for the advice... do u think that taking all the info and going game by game in the 2007 season would be the best way to attempt to verify the results before the season starts?
    2 seasons is definitely too short. Like sharpcat said, it is best to be categorized as trend, which can come and go. It helps to verify the system works if you can go through all available seasons, but I don't know if anybody has found a working system yet in that way.

  25. #25
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by wrongturn View Post
    2 seasons is definitely too short. Like sharpcat said, it is best to be categorized as trend, which can come and go. It helps to verify the system works if you can go through all available seasons, but I don't know if anybody has found a working system yet in that way.
    Yea I mean all seasons would take forever too!!! I'm sure I'm doing it incorrectly... but going through game by game the way I am right now.. I wouldn't be over shooting it if I were to say it probably takes between 30 and 40 hours of work to finish a season...

    Kind of upsetting that 2 seasons is only to be considered a trend... but I guess I will just keep on putting in the man hours and see if what I am seeing continues on...

  26. #26
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Another random question btw as well... These aren't the exact examples... but I am trying to figure out if this has any truth behind it...

    Let's say the Magic are 20-10 at home for a 66.67% winning percentage... and they are 15 and 10 on B2B nights or something at 60%...

    Would it be a fair assumption to assume that at home on a B2B night their winning percentage should be above the 66.67%... since they have another trend above 60%...

    Does that logic make any sense at all??? and if the situations are somewhat correlated does it make a difference???

    Thanks!

  27. #27
    Dark Horse
    Deus Ex Machina
    Dark Horse's Avatar Become A Pro!
    Join Date: 12-14-05
    Posts: 13,764

    Quote Originally Posted by Brp27345 View Post
    Now my questions for the think tank are first of all based on the results from last year would it be more beneficial to bet purely my strongest plays... or should I continue to bet the smaller plays to try to eliminate losses if I have bad nights with my big plays?
    Sports can fluctuate from one season to the next. Just because something worked last season, that doesn't mean it will work again in the next season. One of the most costly mistakes bettors can make is to keep pounding systems that were once good. So the first thing you would want to do in the new season is to find out if the light is still green.

  28. #28
    sharpcat
    sharpcat's Avatar Become A Pro!
    Join Date: 12-19-09
    Posts: 4,516

    Quote Originally Posted by Brp27345 View Post
    Another random question btw as well... These aren't the exact examples... but I am trying to figure out if this has any truth behind it...

    Let's say the Magic are 20-10 at home for a 66.67% winning percentage... and they are 15 and 10 on B2B nights or something at 60%...

    Would it be a fair assumption to assume that at home on a B2B night their winning percentage should be above the 66.67%... since they have another trend above 60%...

    Does that logic make any sense at all??? and if the situations are somewhat correlated does it make a difference???

    Thanks!
    This is the problem with systems it is very important to note that if the system does not take into consideration the opponent the team is facing or the price of the wager it is IMO almost guaranteed to fail long term.

  29. #29
    SolidDala
    SolidDala's Avatar Become A Pro!
    Join Date: 12-14-09
    Posts: 1,696
    Betpoints: 48

    Retrosheets have HUGE data and a seasons data in 1 zip-file and even 10 seasons. This can be opened in excel, the problem I have with this is how to use it with VBA, can one do a search function to find out what one is looking for. No one can go through say 5000 games by hand or even 1000, well you could but the amount of time doing it is just to long, no?

  30. #30
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by Dark Horse View Post
    Sports can fluctuate from one season to the next. Just because something worked last season, that doesn't mean it will work again in the next season. One of the most costly mistakes bettors can make is to keep pounding systems that were once good. So the first thing you would want to do in the new season is to find out if the light is still green.
    How would I go about finding out if the light is still green??? and so wat ur saying basically is that no matter how much back tracking I do that there is no way at all to attempt to solidify the validity of a system...

    I appreciate all of the advice and opinions... but you guys are very depressing when it comes to someone trying to prove a theory... lol

  31. #31
    Dark Horse
    Deus Ex Machina
    Dark Horse's Avatar Become A Pro!
    Join Date: 12-14-05
    Posts: 13,764

    Quote Originally Posted by Brp27345 View Post
    How would I go about finding out if the light is still green??? and so wat ur saying basically is that no matter how much back tracking I do that there is no way at all to attempt to solidify the validity of a system...

    I appreciate all of the advice and opinions... but you guys are very depressing when it comes to someone trying to prove a theory... lol
    You test it out in the new season. Like testing the waters with a toe, before diving in again.

    I can't begin to tell you how many profitable systems I have tossed out. They simply were no longer relevant. Things change. It's not depressing. It teaches you to stay awake. And when you do find something that is good for all seasons, who could put a price on that? It's worth the quest.

  32. #32
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by sharpcat View Post
    This is the problem with systems it is very important to note that if the system does not take into consideration the opponent the team is facing or the price of the wager it is IMO almost guaranteed to fail long term.
    No the system does not take into consideration who is playing... or who that team is playing... it does take into consideration the price of the wager though... by price of the wager I am guessing you mean either one... the line of the game at the half... or two the vig on the half time line...

    As to the line at the half of the game in relation to the opening line that has A LOT to do with the formula that I created... and therefore the line of the game is of upmost importance when deciding what games fit into my system...

    And if you mean the vig on the games then that has little to know effect since I am betting with the spread in all situations... and the vig on halftime lines never goes above -115 and I can often get it at -105... so it evens itself out... and even if it doesn't has little to no effect on the outcome since I am taking into consideration -110 and that is what it averages out to minus or plus 1 or 2 points at the absolute most over the course of a year...

    Do you think not taking team matchups into consideration will completely foil the system alone... because if so I would be willing to go back and look at team matchups... the only issue with that is through 2 seasons teams don't come up with the same stat numbers as the overall NBA totals do... I thought these team numbers would not be nearly as accurate so I was avoiding them...

    Thanks for the continued advice in advance!

  33. #33
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Quote Originally Posted by Dark Horse View Post
    You test it out in the new season. Like testing the waters with a toe, before diving in again. I can't begin to tell you how many profitable systems I have tossed out. They simply were no longer relevant. Things change. It's not depressing. It teaches you to stay awake. And when you do find something that is good for all seasons, who could put a price on that? It's worth the quest.
    Yea exactly... I mean a guy offered me $5,000 last year for it... but I had absolutely no way of figuring out if it was a scam or not so I backed away...

    Just wish I knew more about how to back test and use excel and stuff like that... I'm using excel but not too its fullest extent I'm sure... and the amount of time it takes me to back-test is painstaking...

    I would ask for help but to be honest I don't understand most of the termanology that you guys all use... and I don't think that I would be able to figure out how to do it more effectively anyways...

  34. #34
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    By the way continued chat and thoughts can only help... so let's keep this talk going... thanks so much again for all of your advice and help!!!

  35. #35
    Brp27345
    Most likely a winner...
    Brp27345's Avatar Become A Pro!
    Join Date: 01-01-10
    Posts: 3,692
    Betpoints: 177

    Another random question... does changing ur betting amounts based on ur account value into percentages make sense???

    Like If I go from 1,000 to 1,100 raising my bets... and going from 1,000 to 900 lowering my bets... is this really the preferd way of betting?

    If so I will go along with it... just wondering if it is absolutely the best way...

12 Last
Top