How do you make an NFL line?

[QuadzillaXXX]

I asked this in main street but thelink is right and this is probably the place to ask.

How do you set your own NFL line to compare it to what the spread really is?

And if possible, will someone use an actual example to show maybe past opponents etc so that I can see what process you use instead of just listing the criteria?

Also, is it the same for the NFL and NCAAF?

thanks.

[inSite]

sorry I missed this I already posted in the Main Street

[inSite]

there tends to be less exaggeration in the NFL. even if you see a blowout one week lines only go so high..

[QuadzillaXXX]

i saw it and thank you.....just wasn't sure exactly where the best place to post was

[uva3021]

the best way is typically to find various statistics and run regressions onto the line, with an adequate sample size

i like to use the average line and find the difference between the two teams

[TheLink]

Originally Posted by uva3021

the best way is typically to find various statistics and run regressions onto the line, with an adequate sample size

i like to use the average line and find the difference between the two teams

I would think with football would be harder to get adequate sample size on certain teams due to the turnover at the coaching position and limited number of games.

would be interesting to run regressions on the head coaches/coordinators and see if you can gain any information/edge that way.

[uva3021]

well running regressions with fixed stats, those that determine victories regardless of coach

maybe teams with a high rushing YPA differential tend to win more games, or punt/kick return ability and net special teams, etc...

[QuadzillaXXX]

so you mean take the current line and alter it based on whichever stats you use to cap games?

[uva3021]

use whichever variables more closely resemble the line

take maybe the last ten years, pick some statistics and regress onto the line and find a reasonable combination of correlation and margin of error that best fits the optimal standards set forth by a regression process

[inSite]

use an inaccurate line, for example Sagarin, other computer systems, or lines that were influenced heavily by public money, then go 3 points in the opposite direction of the opening line. now you have a line that is statistically better than all of the above

[uva3021]

^^

elegant, overly simple, yet with immense utility and explanatory power

[QuadzillaXXX]

Originally Posted by inSite

use an inaccurate line, for example Sagarin, other computer systems, or lines that were influenced heavily by public money, then go 3 points in the opposite direction of the opening line. now you have a line that is statistically better than all of the above

you guys rule, learning to actually gamble on football is pretty sweet lol

okay so to make sure i understand.....sagarin ratings pgh@tb
PGH: 26.58
TB + Home field: 21.1+3.16

so it comes to pgh -2.32

so what do i do with that......the "right" line should then be PGH +1?

[inSite]

Originally Posted by QuadzillaXXX

you guys rule, learning to actually gamble on football is pretty sweet lol

okay so to make sure i understand.....sagarin ratings pgh@tb
PGH: 26.58
TB + Home field: 21.1+3.16

so it comes to pgh -2.32

so what do i do with that......the "right" line should then be PGH +1?

no.. the first part is right but the key is the opening line which was Pitt -3 so the -2 is not a noticeable difference.

my point is the opening line is very accurate especially after several weeks in it beats other lines on average including the closing line. so if you notice Pitt opens at -3 and Sagarin and Statfox both say it should be a pk then Pitt could be a consideration for a play.

i couldnt tell you how to put it into practice since its just an average. i dont have the bankroll to make betting worth the time spent researching right now so i just play parlays based on opinion

[inSite]

I have also come up with some good alternate lines before by using a formula similar to the one i posted in the other thread but using rushing off vs. rushing def and passing off vs. passing defense then assigning a point value per unit of yards gained

[Champ10]

Originally Posted by inSite

I have also come up with some good alternate lines before by using a formula similar to the one i posted in the other thread but using rushing off vs. rushing def and passing off vs. passing defense then assigning a point value per unit of yards gained

Let's see it!

What are your guys' thoughts on Sagarin's predicted point-spreads and how to use them?

Week 3 NFL, his plays were:

Atlanta (Sagarin line +1, Real line +4)= Winner Outright
Seattle (Sagarin line +2, Real line +4.5)= Winner Outright
Buffalo (Sagarin line +11, Real line +14)= Winner

Jax (Sagarin line -1, Real line +3) = Loser

There were a few others week 3, but those were the main ones.

Looking ahead to NFL week 4 - There are not many underdog plays like the previous week. Sagarin is suggesting (by some pretty wide margins) that it could be a favorite week. Thoughts?