The basic explanation is here:
After every tenth of the NBA regular season (126 games) I am looking back to see what the results are saying. I have done it twice so far, right on schedule. On Dec 21 we hit the 3/10ths point of the season. Because of the Xmas schedule break, I waited through the 12.23.06 games and am counting all those games in this review.
The headlines:
1. The 5%-OFF method is calling for plays on a lot of games – about three out of every four games played so far (297 out of 401).
2. The Edgar Line numbers have been consistently good and consistently profitable. The fluctuations and cold streaks have been minor. Just blindly playing every game where the Pinny ML closes 5% or more away from the Edgar Line has produced a steady profit.
Make a unit $100. All calculations below assume that a play-on team gets both a $100 ATS bet (risking $105 to make $ 100,) and a ML bet (if on the dog risking $100; if on the favorite to make $100.) .
All games:
Moneyline plays
297 plays
112 wins, 155 losses
$3714
($12.51 per play)
ATS
297 plays
160-135-2
$1825
($6.14 per play)
Blindly dropping $100 ML and $100 ATS on every team where the Pinny ML reaches the strike price would mean a profit through December 23 of $5539 (plus another $205 on Xmas Day.)
There has been little by way of a losing skid to test anyone's nerve. The ML plays have shown a profit in 7 or the 8 weeks; the ATS plays in 6 of the 8 weeks
3. Obviously, discovering which subgroups of plays did best in the past says nothing certain about the future. But it’s still worth looking backward. I’ve now sliced these 297 past plays every way I can, to try to identify for future use what the characteristics of a prime play might be. In the posts below I’ll just set out some of the results. But the main conclusion is:
a. So far this year, at a certain point when Edgar Line is too far off from the market, the market tends to be somewhat more right.
The Edgar Line is still much more right when it is at least 5% but less than 12% off from the closing Pinny ML.
But the market has been more right overall when the Edgar Line is 12% or more (actually, 17% or more) off the closing Pinny ML, especially when the Edgar Line is pointing to the home team.
b. Overall, focusing just on when the Edgar Line is between 5% and 11% off from the Pinny closing line, taking the “play-on” team for 1 unit on the ML has been worth $29.37 per play or a total of $4582 over 156 bets (69 wins, 87 losses).
Betting them one $100 unit ATS has been worth $18.53 per bet or a total of $2890 over 156 bets (94 wins, 62 losses).
Overall, when the Edgar Line is 12% or more off from the Pinny closing line, taking the “play-on” team for 1 unit ($100) on the ML has been worth -$6.16 per bet or a total loss of -$868 over 141 bets. Betting one unit ATS has been worth -$7.55 per bet or a total loss of -$1065 over 141 bets (66-73-2).
But the performance of the “12%+” teams also has a wide home/road differential. That is, they are worse at home.
82 plays on a “12%+” home team:
ML -$676 (-$8.24 per play).
ATS -$1435 (-$17.50 per play)
Better on the road:
59 plays on a "12%+" road team:
ML: -$192 (-$3.25 per play)
ATS: $370 (+6.27 per play)
A couple of other subgroups have been losers so far. They are:
Plays against the best western teams (PHX, DAL, SAN) Yes there was the +815 moneyline hit on CHA against SAN. But otherwise, nothing but misery. Those three teams just take care of business far too well. ML plays against the big three are down -$1373 over 64 plays (-$21.45 per play) ( 14-50). ATS plays are down -$1490 (-$23.28 per play) over the same 64 plays (25-38-1).
Playing true favorites has also been a small negative. Of course the system doesn’t really call for many of these plays – so far just 12 plays on teams giving 3.5 points or more. They are 6-6 on the ML (for -$393) and 5-7 ATS (for -$235).
Saturday night has been the worst night of the week by far. Don’t know why. The 63 Saturday night plays so far are down -$2026 (-$32.16 per play) on the ML and down -$465 (-$7.38 per play) ATS. My first thought was that it might have something to do with Saturday night games having the largest share of unrested teams. While I do try to account for no-rest in the Edgar Line, maybe that factor makes things too volatile? But that doesn’t really check out. The no-rest factor overall is close to break even. (All unrested play-on teams are up $1.20 per play on the ML and down -$6.20 per play ATS over 82 plays no matter the day of the week. All plays against unrested teams are up $4.57 per play on the ML and down -$4.03 per play on the ML.)
But, for whatever reason, the “12%+” teams in particular have been beyond brutal on Saturday nights. Over 24 games they are 3-21 on the ML for -$1718 units or an amazing minus $71.58 per $100 play. ATS they are 10-14 for -$470. That group of 24 ugly plays includes Memphis six times and the Knicks four times; it includes 19 plays on unrested teams, but 16 plays against unrested opponents. Somehow Saturday night has been home to the worst of the worst. Even the teams in that Edgar Line sweet spot, at 5-11% off the closing Pinny ML, are also worse on Saturday night than in the rest of the week. Saturdays they lose $8 per play on the ML and are only break even ATS. For now, I can only think that the Saturday thing is probably just an aberration.
Of course, just for fun, I have to run this completely bogus, backfitted analysis:
Playing teams 5%-11% off only, never fading DAL-PHX-SAN, and avoiding favorites of 3.5 or more or any action on Saturday nights:
ML plays
45-51
+$3442
(+$35.85 per play)
ATS plays
62-34
(.646)
+2630
(+$27.40 per play)
I think it says something very good that this "dream set" of games does not actually dramatically outperform the entire set of 5% OFF plays. The dream set is up $6072 overall, while the full set is up $5539.
And the 5-11% subset of all games, with nothing else taken out, is up $7472 -- even better than the cherry picked plays. In other words, cherry-picking among the 5-11% (or 5-17%, which is essentially the same net) is probably not helpful or necessary. The main thing is recognizing that if the market is way way off from the Edgar Line, the market may "know" and be accounting for something that the Edgar Line isn't.
I'll make a few posts below dumping some of the data, for anyone who is curious.
After every tenth of the NBA regular season (126 games) I am looking back to see what the results are saying. I have done it twice so far, right on schedule. On Dec 21 we hit the 3/10ths point of the season. Because of the Xmas schedule break, I waited through the 12.23.06 games and am counting all those games in this review.
The headlines:
1. The 5%-OFF method is calling for plays on a lot of games – about three out of every four games played so far (297 out of 401).
2. The Edgar Line numbers have been consistently good and consistently profitable. The fluctuations and cold streaks have been minor. Just blindly playing every game where the Pinny ML closes 5% or more away from the Edgar Line has produced a steady profit.
Make a unit $100. All calculations below assume that a play-on team gets both a $100 ATS bet (risking $105 to make $ 100,) and a ML bet (if on the dog risking $100; if on the favorite to make $100.) .
All games:
Moneyline plays
297 plays
112 wins, 155 losses
$3714
($12.51 per play)
ATS
297 plays
160-135-2
$1825
($6.14 per play)
Blindly dropping $100 ML and $100 ATS on every team where the Pinny ML reaches the strike price would mean a profit through December 23 of $5539 (plus another $205 on Xmas Day.)
There has been little by way of a losing skid to test anyone's nerve. The ML plays have shown a profit in 7 or the 8 weeks; the ATS plays in 6 of the 8 weeks
3. Obviously, discovering which subgroups of plays did best in the past says nothing certain about the future. But it’s still worth looking backward. I’ve now sliced these 297 past plays every way I can, to try to identify for future use what the characteristics of a prime play might be. In the posts below I’ll just set out some of the results. But the main conclusion is:
a. So far this year, at a certain point when Edgar Line is too far off from the market, the market tends to be somewhat more right.
The Edgar Line is still much more right when it is at least 5% but less than 12% off from the closing Pinny ML.
But the market has been more right overall when the Edgar Line is 12% or more (actually, 17% or more) off the closing Pinny ML, especially when the Edgar Line is pointing to the home team.
b. Overall, focusing just on when the Edgar Line is between 5% and 11% off from the Pinny closing line, taking the “play-on” team for 1 unit on the ML has been worth $29.37 per play or a total of $4582 over 156 bets (69 wins, 87 losses).
Betting them one $100 unit ATS has been worth $18.53 per bet or a total of $2890 over 156 bets (94 wins, 62 losses).
Overall, when the Edgar Line is 12% or more off from the Pinny closing line, taking the “play-on” team for 1 unit ($100) on the ML has been worth -$6.16 per bet or a total loss of -$868 over 141 bets. Betting one unit ATS has been worth -$7.55 per bet or a total loss of -$1065 over 141 bets (66-73-2).
But the performance of the “12%+” teams also has a wide home/road differential. That is, they are worse at home.
82 plays on a “12%+” home team:
ML -$676 (-$8.24 per play).
ATS -$1435 (-$17.50 per play)
Better on the road:
59 plays on a "12%+" road team:
ML: -$192 (-$3.25 per play)
ATS: $370 (+6.27 per play)
A couple of other subgroups have been losers so far. They are:
Plays against the best western teams (PHX, DAL, SAN) Yes there was the +815 moneyline hit on CHA against SAN. But otherwise, nothing but misery. Those three teams just take care of business far too well. ML plays against the big three are down -$1373 over 64 plays (-$21.45 per play) ( 14-50). ATS plays are down -$1490 (-$23.28 per play) over the same 64 plays (25-38-1).
Playing true favorites has also been a small negative. Of course the system doesn’t really call for many of these plays – so far just 12 plays on teams giving 3.5 points or more. They are 6-6 on the ML (for -$393) and 5-7 ATS (for -$235).
Saturday night has been the worst night of the week by far. Don’t know why. The 63 Saturday night plays so far are down -$2026 (-$32.16 per play) on the ML and down -$465 (-$7.38 per play) ATS. My first thought was that it might have something to do with Saturday night games having the largest share of unrested teams. While I do try to account for no-rest in the Edgar Line, maybe that factor makes things too volatile? But that doesn’t really check out. The no-rest factor overall is close to break even. (All unrested play-on teams are up $1.20 per play on the ML and down -$6.20 per play ATS over 82 plays no matter the day of the week. All plays against unrested teams are up $4.57 per play on the ML and down -$4.03 per play on the ML.)
But, for whatever reason, the “12%+” teams in particular have been beyond brutal on Saturday nights. Over 24 games they are 3-21 on the ML for -$1718 units or an amazing minus $71.58 per $100 play. ATS they are 10-14 for -$470. That group of 24 ugly plays includes Memphis six times and the Knicks four times; it includes 19 plays on unrested teams, but 16 plays against unrested opponents. Somehow Saturday night has been home to the worst of the worst. Even the teams in that Edgar Line sweet spot, at 5-11% off the closing Pinny ML, are also worse on Saturday night than in the rest of the week. Saturdays they lose $8 per play on the ML and are only break even ATS. For now, I can only think that the Saturday thing is probably just an aberration.
Of course, just for fun, I have to run this completely bogus, backfitted analysis:
Playing teams 5%-11% off only, never fading DAL-PHX-SAN, and avoiding favorites of 3.5 or more or any action on Saturday nights:
ML plays
45-51
+$3442
(+$35.85 per play)
ATS plays
62-34
(.646)
+2630
(+$27.40 per play)
I think it says something very good that this "dream set" of games does not actually dramatically outperform the entire set of 5% OFF plays. The dream set is up $6072 overall, while the full set is up $5539.
And the 5-11% subset of all games, with nothing else taken out, is up $7472 -- even better than the cherry picked plays. In other words, cherry-picking among the 5-11% (or 5-17%, which is essentially the same net) is probably not helpful or necessary. The main thing is recognizing that if the market is way way off from the Edgar Line, the market may "know" and be accounting for something that the Edgar Line isn't.
I'll make a few posts below dumping some of the data, for anyone who is curious.
