Posted:
#776
Hey guys,
Those lucky to get the spreadsheet from USASurfer, can someone email it to me at Thanks! Or if you want PM me first then we'll do it through covers.
Recon
0
New YPPT Method
Posted:
#778
Quote Originally Posted by roughshod31:
lou4062
I use the Total Plays & Yards Per Play as a means on double checking the accuracy of YPPT calculation. I will show you using Texas @ Oklahoma St. as an example.
Offense Plays (Differential) Offense YPP (Differential)
UT 78.6 ( + 6.7 ) 5.6 ( + 0.5 )
OK St. 68.0 ( - 3.5 ) 6.1 (+ 0.1 )
Defense Plays (Differential) Defense YPP (Differential)
UT 65.9 ( - 5.6 ) 3.6 ( - 1.7 )
OK St. 71.7 ( + 0.3 ) 4.9 ( - 0.4 )
Texas SOS (Mult.) 1.030
OK St. SOS(Mult.) 0.970
Texas 78.6 + 0.3 = 78.9
71.7 + 6.7 = 78.4 (78.9 + 78.4 = 157.3/2 = 78.7)
78.7 x 1.030 = 81.1 Expected Plays
Texas 5.6 + -0.4 = 5.2
4.9 + +0.5 = 5.4 (5.2 + 5.4 = 10.6/2 = 5.3)
5.3 x 1.030 = 5.5 Expected YPP
Texas 81.1 Plays x 5.5 YPP = 446.1 Total YDS.
430.1 Total Yds. YPPT Calc.
Do the same for Oklahoma St. and you have a good way to double check your YPPT calculation.
Hope this helps
GL
(
0
Posted:
#779
Quote Originally Posted by lou4062:
Thanks roughshod for sending me a thorough example for this weeks game. That is a nice way to double check everything just to be sure. Checking back to your earlier posts, I can see what your doing but I can't try it because the games are gone. I wish that you could go back to the match ups of prior games but you can not. The games disapear pretty quickly, once they are played.
Here is what I got for the tex at ok.st
yds yppt pts hfa scor
tex 430.1 12.6 34.1 -1.25 32.8 (34)
ok 240.7 15.9 15.1 +1.25 16.4 (17)
I rounded up for tex so I rounded up for ok also.
I only gave ok +2.5 because of what you said about the 7 point difference in power ratings.
tex = 86.49
ok st = 79.26
0
Posted:
#790
Quote Originally Posted by powerade:
:(
now that you say that and I think about it....
I'm 
at myself
Don't be
I don't want to take anything away from the work these guys are doing I was just amazed by the amount of people doing the email thing, and then right at the end there was a captain...
I was just takin cheap shots
0
Posted:
#793
Some people have been asking what is the record on these plays since Roughshod31 created the thread in the NFL forum.
These are the record for the plays that differed by +/- 7 points from Vegas closing line (scoresandodds.com) (Roughshod31 himself set a +/-4 point difference in the beginning. I changed it to a full TD to get better results). Also, when I started to post predictions, I counted those games, not Roughshod31's predictions, since he kept adjusting the system, adding HCA and other factors. (No disrespect to the him)
NFL:
18 wins - 6 losses - 1 PUSH (75%)
NCAAF:
24 wins - 15 losses - 1 PUSH (61.5%)
TOTAL:
42 wins - 21 losses - 2 PUSH (66.7%)
These are the record for the plays that differed by +/- 7 points from Vegas closing line (scoresandodds.com) (Roughshod31 himself set a +/-4 point difference in the beginning. I changed it to a full TD to get better results). Also, when I started to post predictions, I counted those games, not Roughshod31's predictions, since he kept adjusting the system, adding HCA and other factors. (No disrespect to the him)
NFL:
18 wins - 6 losses - 1 PUSH (75%)
NCAAF:
24 wins - 15 losses - 1 PUSH (61.5%)
TOTAL:
42 wins - 21 losses - 2 PUSH (66.7%)
0
uld someone PM me with the spreadsheet. Thanks in advance
Posted:
#799
[
NFL:
18 wins - 6 losses - 1 PUSH (75%)
NCAAF:
24 wins - 15 losses - 1 PUSH (61.5%)
TOTAL:
42 wins - 21 losses - 2 PUSH (66.7%) [/Quote]
There are much better ways to test the accuracy of your model than just the ATS record of when the games meet your criteria.
Action points is one of them. Actions points is how the team did in relation to the spread. So if the thread is +7 and the team looses by 3 your are -4 action points. If they win by 3 you are +10 action points. Your goal is to be as positive as possible
Ther other is how often does your model say to bet an opening line but pass on a closing line. This is showing you if your model is in agreement with the market. This is the strongest way to predict the sccuracy of your model. If your model isconstantly predicting the correct movement of the market it IS A WINNER.
Also keep in mind that almost no models do well in al situations. Dr. Bob boasts the most famous and accurate math model on the planet (that we know of atleast). And even he admitted that his model seriously struggles on large road favorites. This comes from some serious testing and work and if you guys are willing to do that than give up on the whole"math to give you and edge" mindet, because you wont get one.
I understand you were just answering their question. this is for everyone else 
0
Hot Forum Topics
