NBA OU Systems

I finally realized developing a sucessful NBA system is not going to happen in a few weeks or even months.  As I said about my NFL system I worked for years developing it, so I am going to try and be patient and do the same for the NBA.  I thought I had something that would work, but so far I am not even getting 50% overall.  I think my problem is I am not patient ... if it doesn't work I tweak it and tweak it and tweak more.  That's good and bad.  So now I have a multitude of systems and variations ... and backtesting shows some sucess, but I keep playing different ones and getting 50%.

 

Soooooo .. I'll share some of them and if someone is patient, and sticks with one for a while they may show a profit, where I couldn't (stick with it that is.)  So what I'll do is track them and post the results and stop making "predictions" so to speak, but let you make the choice from various systems.  And if you have one I can understand I'll track it too and include it.  I'm looking for concensus in systems and see how that works.

 

This may take a while to get all together, but as you guys may know, I'm pretty dedicated to make this work.  So let me know what ideas you may have and any suggestions.

 

First off I want to post like I did in Cappers thread the backtesting results for my latest system.  Formatting this for this forum absoluty sucks, but I will do my best.

 

Games    Win    Lose   Push   Win %
    25         15      10        0       60%
    50         27      23        0       54%
    75         42      33        0       56%
   100        56      44        0       56%
   125        69      56        0       55.2%
   150        84      66        0       56%
   175        97      78        0       55.4%
   200       113     86        1       56.5%
   250       144     105      1       57.6%
   300       176     121      3       58.7%
   318       188     126      4       59.1%

 

As you can see the backtesting looks promising.  Now I just need to either keep it up, figure out what I am doing wrong theory wise, or find the mistakes I may have made.  It just looks to good to be true, so it probaly is.  This system is for all games ... I have some filters I have added to make the results look a bit better.  This filter cuts out about 60% of the total games played.

 

Games    Win    Lose   Push   Win %
    25         6        4        0       60%
    50         9        9        0       50%
    75         16      10       0       61.5%
   100        19      13       0       59.4%
   125        27      16       0       62.8%
   150        36      20       0       64.3%
   175        40      24       0       62.5%
   200       48       28       0       63.2%
   250       67       34       0       66.3%
   300       88       39       1       68.8%
   318       92       41       1       68.7%

 

Believe it or not this system does not rely on the line!!!  I'm trying to think of a way to include the line, but that will take some work.  I basically let the books figure out the day to day adjustments to the line .. ie injuries and such.  I'm looking more at the tendencies of the two teams past, if that makes sense.

 

I'll post the picks for these two systems in a few minutes.

Opps, turns out I do need the line for the filter aspect of my system, but here is the all games part...

 

Memphis/Toronto  Over
Atlanta/Orlando  Over
Charlotte/Philadelphia  Over
Boston/Cleveland  Over
Houston/Oklahoma City  Under
LA Clippers/New Orleans  Over
New Jersey/Milwaukee  Under
Washington/Chicago  Over
Detroit/Denver  Over
Miami/Sacramento  Under
Indiana/LA Lakers  Under
Dallas/Phoenix  Over

 

I'm tired and need to get to bed .. I'll post the other part tomorrow.

...are you going to add a ...less than 6 pt ....6-10py ....+10 filter? Reverse of the +10 angle?.....etc....trying to find the "best" bet out of all them...most consistent win rate..

Ken Writes....

Gas ... I may end up with "ranges" or something similiar, but I kinda doubt it.  More likely to end up with just percentages ... its a totally different system.  I just compared my picks to Caps sys and we agree on 7 and disagree on 5, so let me figure out my "top plays" and make a post later.

 

Yes, I agree on finding the most "consistant bet" ... that is why I started this thread to compare systems and see if when they agree they win more???

Memphis/Toronto  Over  Top Play
Atlanta/Orlando  Over
Charlotte/Philadelphia  Over
Boston/Cleveland  Over  **CTP**
Houston/Oklahoma City  Under
LA Clippers/New Orleans  Over  **CTP**
New Jersey/Milwaukee  Under  Top Play  **CTP**
Washington/Chicago  Over  Top Play
Detroit/Denver  Over  Top Play
Miami/Sacramento  Under
Indiana/LA Lakers  Under
Dallas/Phoenix  Over  Top Play

 

So the consensus play is NJ/Mil Under, both my systems agree which makes it a top play, plus Capper has it as a top play.  It's going to be interesting tho ... my system disagrees with Cappers other two top plays and his system disagrees with two of my top plays ... his system does agree with two of my top plays (Mem/Tor Over and Det/Den Over)

 

Games we agree on are underlined, Cappers top plays have **CTP** following.

....so how can the "away/home offense" numbers be more that the vegas line and an UNDER be the call....

 

...I don't get it...

 

I think it should be....

 

Thanks Josh GL to you and your "clients"

 

Gas, When I tested Cappers system I declared it was at 47.1%, not good, but you should never look at a system below 50% ... ie if it is BAD, that is good, just reverse the picks, or some call it fading.  So the system at 47.1% is really 52.9% ... get it?  If you look back at Cappers picks he was reversing most all plays where the difference was > 7, (somtimes 8, and I think he posted 9)   All I did is say reverse them all, even if less then 7 (or 8, or 9).  Reverse them even if the off road/home #'s are 40 points over/under the Vegas line.  Now that trend is changing, but overall it says reverse them all.  Reversing them when in the 7-10 range is the best play (considering # of games played as well) from that system ... 57.9%.  Reversing in the >16 (same as 16-50) is even better, but fewer games.  I agree it doesn't make sense, but hey that's the NBA and betting OU's. 

It was a pretty crappy night ... Our "concensus pick" was over a few minutes after the opening tap.  I just have to be patient I keep telling myself.  LOL 

 

Almost 3-1 Gas ...

 

Here's my picks for all games ..

 

Charlotte/Washington  Over
Milwaukee/Minnesota  Under
New York/Houston  Under
Oklahoma City/Chicago  Under
Detroit/Utah  Over
Golden State/Portland  Under

 

If I don't get to involved in football I'll post the top picks.

...ken....now I am really confused...

 

.....your pics are exactly as calculated instead of "reversed" like you say expertcapper plays them....

 

...interesting that all of the plays today are in the 6-10 calculated range....calculated the "gasman" regular way was 2-0 yesterday

 

 

...do we play them as you have listed or the "reverse "way???

 

...what am I saying...

 

OkC/Chi, Det/Uta, Mil/Min....7-10 range

 

....sorry

Gas,  You are totally confused.  The pics I listed above have nothing whatsoever to do with Cappers system.  It's a totally different system as I said a few posts ago.

 

It is intersting tho, the picks are exactly opposite of Cappers system.  His system has none in the 7-10 range, but two >10 which he may list as top plays.

 

I'll list my top plays as

Gas to avoid confusion I may have used different lines when I say Capper's system has no plays in the 7-10 range ... you'll just have to wait to see what he posts.

Glad you are back and hope you are feeling better Capper ...

Thanks Capper ... yea I read a little on your 1st half stuff, keep it up.  I am behind on working on the new system stuff, but I did manage to make a list of the records.  My systems have done absolutely terrible and it looks like it is time to fade them.  I hope everyone can understand the tables ... it is the record for the last 25, 50, 75 etc games.  The win % is the % if played normal and the fade % is if played opposite.  # is the number of games that are played with the three systems I am tracking, and the ones to the right and the last line being with filters.

 

What it shows, for example, is that for the last 50 games Capper's 7-10 system is 2-10, with 24% of the games being played.  So playing the opposite is 10-2 or 83.3%.  Playing the opposite means NOT reversing them.  I have no idea why these trends are reversing or more important when, or if, to switch the plays, but I can track them and post to the above link so we can all see what they are doing.

 

I am going to reverse the picks for my system and post in a bit.

I found errors in the past records for Caps system, plus it will be strickly following the computer picks and not meant to reflect Cappers record.  I usually use the line when I bet as the official record that is recorded and that causes differences as well.

 

Ok, here are my picks for tonight, fading my system ...

 

Milwaukee/Washington Over 192 
Toronto/Boston Under 186.5 
Oklahoma City/New Jersey Over 202.5 
New York/New Orleans Over 203.5  ***
Portland/Chicago Under 193.5  ***
Indiana/Utah Over 220.5  ***

 

Interestingly enough it matches the Cappers system picks exactly ... I don't think it matches cappers actual picks tho. 

 

It's not fade the system, but fade whatever I pick ... LOL  Had I not faded I woulda been 5-1 (2-1)  Not a very good job of being patient.  No more fading for a while.

 

Charlotte/Detroit Over 176 
Cleveland/Memphis Under
Miami/Minnesota Under 198.5 
LA Lakers/Houston Under  ***
Atlanta/Phoenix Over 212  ***
Dallas/Denver Over 207 
Orlando/Sacramento Under 201.5  *** 

 

Even tho I don't use the lines I will include them, so make it

Cle/Men under 185

LAL/Hou under 195

 

I have the three *** games but I also have two games that show up weaker ...

Cha/Det and Mia/Min

 

Gas, the decision to fade was a one time experiment gone wrong ... these picks are straight from the revised system as is.

 

Capper,  So these picks look opposite of what I have been usually picking???

 

I'll explain a little about my new system.  What I do is compare up to 37 different aspects of the game from the stats I generate.  The stats are like what I use in the NFL ... either positive or negative #'s.  I then find a value that gives me the largest spread in past over and under values for a given parameter.  Here's an example, if the home off > 0 and the road off - home def > -4.5 then 63.4% of the time the game is an over.  Do that for 36 more similiar comparisons and sum up the percentages.  Positive means over and neg means under.

 

So that system gives me the numbers I had listed under "New System" in the previous link, it has done ok, I have just not been patient.  So last nite I revised it ... I cut back the 37 tests to about 15 and I only go back 199 games to Dec 16.  The overall backtest is at 57.6%.    Here is the graph of the backtest results.  So tonight is the first nite of this new  variation.