[curious]

NCAAB Free money plays are 58-23-2 since Jan 19th

[Gundog]

Curious
Can you explain your method? Or is there a site i can go to and do some reading on your method. You are pretty darn good.
Thanks

[curious]

Curious
Can you explain your method? Or is there a site i can go to and do some reading on your method. You are pretty darn good.
Thanks

It is pretty simple. I take the pythagorean win % for each team. I then multiply the pythagorean win % by a factor to determine what the spread should be. I then compare this "should be" spread to the actual spread. I subtract the "should be" spread from the actual spread to determine the edge. If the edge is above 10 then I have found a play.

The pythagorean win% is determined by using a pythageorean equation using the offensive efficiency and the defensive efficiency of each team. http://kenpom.com/stats.php has done this for us already.

I found the factor to use by taking every line for every game since the 3rd week of the season and dividing it by the pythagorean win % for that team on that day. The factors were all over the place so I had to use the mean.

I have found that spreads over -15 cannot be trusted so I throw those out. I have also found that "should be" spreads over about -22 are a contraindicator so I throw those out also.

I have also found that if a team should have been a fav or a small dog and are instead a big dog that is an automatic win. So far those are hitting 100%.

A couple of other things I have found which are very interesting. The formula is much better at predicting weekday games than Saturday games. Also, it is better at predicting early games than late night games. The western conference is absolutely impossible to predict. Just blindly bet the dogs in that conference. Also I found that days with fewer games are better than days with lots of games. I can't imagine why that would be true.

I can improve the effectiveness of this formula by creating factors for spreads within ranges, I have found that different types of spreads behave differently. By type I mean how far away from 0. I can also improve the formula by modifying it for different "levels" of teams. I have found that really bad teams are very unpredictable.

Also, I am not taking into consideration injuries, emotion, regional rivalry, days of rest etc. If I add those things the formula will probably improve.

However, I am reluctant to mess with the formula because it is winning at a 70% clip and it is simple. I think over ANALyzing is a very bad thing to do.

[stingyrivers]

curious,

I would like to apply the math myself to each slate of games....

1st) when you say take the win% for each team and multiply by the factor, do you mean individually for each teams win % or combine their win % and multiply that by the factor?

2nd) what was the actual number you came up with for the factor?

thanks...

[curious]

curious,

I would like to apply the math myself to each slate of games....

1st) when you say take the win% for each team and multiply by the factor, do you mean individually for each teams win % or combine their win % and multiply that by the factor?

2nd) what was the actual number you came up with for the factor?

thanks...

Here I am talking about adjusted pythagorean win %.

I subtract the lower pythagorean win % from the larger pythagorean win % and multiply the result by -50 and assign the result to the team with the larger pythagorean win % and then I assign the absolute value of the result to the team with the lower pythagorean win %. Here is an example where the team with the larger pythagorean win % has the edge:

College of Charleston
1 1/2 actual line
0.446818 pythagorean win %
-7.7127 (.446818 - .292563) * -50 ("shoud be" line)
9.21275 edge (actual line - "should be" line)

wofford
-1 1/2 actual line
0.292563 pythagorean win %
7.71275 abs((.446818 - .292563) * -50) ("should be" line)
the edge here is negative so I ignore it. College of Charleston has the edge in this game.

I have found when the team with the lower pythagorean win % has the edge (and the edge is greater than 8 or so) it is almost always a winner

words of advice:
1. Huge favs are not predictable ATS. I don't play favs where the spread is larger than -15 except in special situations like Gonzaga vs Portland.
2. When the line is - and the edge is huge, say over 20 you cannot trust this game by the formula alone you need to handicap it to see if you still like the play because the formula doe not work very well for these sutuations For example, say Team A is -8 and the "should be" line is -38, making the edge 30, Team A will probably lose.
3. West coast teams cannot be predicted when playing each other. Don't ask me why. If you want to bet on teams from the west coast conference bet 2 units on the dog spread and 1 unit on the dog moneyline. LT has some stats about this.
4. Early games are more predictable than night games.
5. Saturday is not a good day for the formula, except the Sat AM games. DO NOT bet too many games on Saturday. I would say keep it to 5 or 6 on Saturday. I have not had a winning Saturday since I started this.
6. Week day games are very predictable (Mon-Thurs).
7. Days with a small number of games seem to be more predictable.
8. I don't account for emotion, regional rivalry, injuries, home court, or other factors in the formula. I have had good success picking a team that has a small edge when I know that the team is a fierce competitor via regional rivals or at home. Cincinnati is a good example of a team like this. If they have any edge at all and are playing at home or against a regional rival I will take them.

[stingyrivers]

curious, thanks for the reply....

ok so to see if i followed you correctly, i applied the formula to the ohio st at penn st game tonight, are my results correct with:

actual line 7 1/2

"should be spread" = 11 edge is 3.5???


OHio St
-7 1/2 actual line
.947 PW %
-11 (.947 - .727) * -50
edge = 3 1/2

P ST
7 1/2 actual line
.727 PW %
11 (.727 - .947) * -50

... i think i went wrong somewhere, let me know

[MoneySportsGuy]

Ok so can somebody explain this more or can you PM me curious I am not understand all of this.

[WestsidePete]

curious what are your plays today based on this system??

[Irish Lumberjack]

If i'm getting this right you take team which are favored if the edge is over 10 and dogs if the edge is 8 or so, so then would the two plays for tonight be:
Western Mich - 3.5 & Missouri St +1

??????