Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

I like the computer models. I think they are useful. Same for yours, Medina.

Your version looks like it may struggle with the strenght of schedule: Oregon played a weaker schedule than Ohio State, but it still won all of its games. They also don't do a good job of capturing the teams that are better or worse in close games. Sometimes it's just luck, but when you get a consistent set of results, it starts to suggest that some teams handle the stress of close games better. That is a behavior that can be learned/coached, so when you have a team that consistently wins or consistently loses those close games, I think it's fair to take it out of the luck category.

I think the goal of such a model is to TRY to correct for strength of schedule issues. 

However I think the blind spot might be that it incorporates a certain level of MOV advantage that can benefit schools who run up the score vs schools who are pulling their starters in the 3rd quarter and coasting. 

I.e. I was finding it interesting that Indiana was still as high as they are... But I think Indiana knows that when it comes to the committee, and when your schedule is Charmin soft, that "style points" mean something. If you're playing to try to prove something because you know you're not ever getting the benefit of the doubt, you run up the score. I mean, their starting QB was in the game against Purdue until they were up 59-0. I know it's a rivalry game so you also want to stick the knife in deep, but Indiana knew they couldn't take their foot off the gas in any game. 

So a team like OSU that had a tougher schedule than Oregon may get a benefit of being the best against tough opponents because they played more tough opponents. Whereas someone like Oregon knew that in certain games they already were beating the weak opponents badly, and decided to just pound the rock in the running game to chew clock because they KNEW the other team wasn't going to come back. 

So a model that tries to correct for strength of schedule might actually exacerbate some strength of schedule issues...

Anyway @medinabuckeye1 -- I'm just spitballing here, not criticizing your model. Just interested in it for academic purposes. 

If I've offended, PM me your address and I'll send you an early Christmas gift  

I think the goal of such a model is to TRY to correct for strength of schedule issues.

However I think the blind spot might be that it incorporates a certain level of MOV advantage that can benefit schools who run up the score vs schools who are pulling their starters in the 3rd quarter and coasting.

I.e. I was finding it interesting that Indiana was still as high as they are... But I think Indiana knows that when it comes to the committee, and when your schedule is Charmin soft, that "style points" mean something. If you're playing to try to prove something because you know you're not ever getting the benefit of the doubt, you run up the score. I mean, their starting QB was in the game against Purdue until they were up 59-0. I know it's a rivalry game so you also want to stick the knife in deep, but Indiana knew they couldn't take their foot off the gas in any game.

So a team like OSU that had a tougher schedule than Oregon may get a benefit of being the best against tough opponents because they played more tough opponents. Whereas someone like Oregon knew that in certain games they already were beating the weak opponents badly, and decided to just pound the rock in the running game to chew clock because they KNEW the other team wasn't going to come back.

So a model that tries to correct for strength of schedule might actually exacerbate some strength of schedule issues...

Anyway @medinabuckeye1 -- I'm just spitballing here, not criticizing your model. Just interested in it for academic purposes.

If I've offended, PM me your address and I'll send you an early Christmas gift

I hadn't really looked that closely but now that you mention it, IU was #1 in the league against:

• 0-9 Purdue
• 3-6 UCLA
• 3-6 MSU
• 3-6 UNL

It probably isn't coincidental that none of those teams are very good.  It would make sense that Oregon/Ohio State/Penn State wouldn't "need" to score 66 on Purdue but Indiana might.  

LoL @ your gift, funny!

I hadn't really looked that closely but now that you mention it, IU was #1 in the league against:

• 0-9 Purdue
• 3-6 UCLA
• 3-6 MSU
• 3-6 UNL

It probably isn't coincidental that none of those teams are very good.  It would make sense that Oregon/Ohio State/Penn State wouldn't "need" to score 66 on Purdue but Indiana might. 

LoL @ your gift, funny!

Yeah, I wonder if you could pick a number, 21, 24, 28, whatever, and rather than rank by MOV once you get over that number, you would rank any win based on how much time was left in the game when you surpassed that margin.  So if Oregon went up 28-0 on Purdue in the 2nd quarter, and then pulled starters and played conservatively in the 2nd half, played backups, and only won 42-7, that would still be better than Indiana not taking a 28-0 until into the third quarter, but keeping their foot on the gas and winning 56-0

Yeah, I wonder if you could pick a number, 21, 24, 28, whatever, and rather than rank by MOV once you get over that number, you would rank any win based on how much time was left in the game when you surpassed that margin.  So if Oregon went up 28-0 on Purdue in the 2nd quarter, and then pulled starters and played conservatively in the 2nd half, played backups, and only won 42-7, that would still be better than Indiana not taking a 28-0 until into the third quarter, but keeping their foot on the gas and winning 56-0

I agree with this in theory but in practice that is a LOT of work.  

I've mentioned here before that I evaluate blowouts based on time rather than margin, almost exactly as you laid out.  

My go-to example is two tOSU games that I attended:
1995 Ohio State vs Iowa:
The Buckeyes built a 56-0 lead in the first half then rested their starters.  Eddie George only had 110 yards and 4 TD's because he didn't play after the break.  In the second half Iowa outscored Ohio State 35-0 leading to a final MoV of "only" 21 at 56-35.  

2006 Ohio State vs Penn State:
The Buckeyes and Nittany Lions played a fairly evenly matched defensive slugfest in which the difference in the game was that Ohio State's Goal Line units bested Penn State's Goal Line units.  Deep in the fourth quarter the Buckeyes had an eight point lead at 14-6 based on two TD's for tOSU and two FG's for PSU.  Eight points, however, is only one score so the game was still 'in play' and PSU was driving with just over 2:00 to go in the game when PSU QB Anthony Morelli tossed a pass that tOSU DB Malcolm Jenkins caught and returned for a pick-6 to effectively end the game at 21-6 with just over two minutes to go.  Then, on the very next drive, with just over one minute to go, PSU tossed yet another pick-6 leading to a final MoV of 22 points at 28-6.  

The problem with a lot of objective measures such as mine above is that it would treat tOSU's 22 point win over PSU in 2006 as "better" than tOSU's 21 point win over Iowa in 1995 because 22>21.  However, any of us can look at those two games and realize that the 2006 tOSU/PSU game was MUCH more evenly matched than the 1995 tOSU/Iowa game. 

• OREGON
• PENN STATE
• INDIANA
• MINNESOTA
• IOWA
• ILLINOIS
• MICHIGAN 😈
• OHIO STATE😈
• USC
• WASHINGTON
• WISCONSIN
• UCLA
• RUTGERS
• MICHIGAN STATE
• NEBRASKA
• NORTHWESTERN
• MARYLAND
• PURDUE

very BOLD my friend

Massey Composite Computer Rankings - 88 rankings (last week in parentheses)

• OREGON (2)
• Texas (3)
• Notre Dame (4)
• OHIO STATE (1)
• PENN STATE (5)
• Georgia (6)
• SMU (9)
• INDIANA (7)
• Alabama (10)
• Tennessee (12)
• Ole Miss (11)
• South Carolina (13)
• Miami (8)
• Iowa State (15)
• BYU (17)
• Boise State (16)
• Arizona State (22)
• Louisville (19)
• Clemson (14)
• Colorado (23)
• LSU (24)
• Texas A&M (18)
• Missouri (25)
• Kansas State (20)
• IOWA (-)

• 26. Illinois (27)
• 27. Michigan (33)
• 34. Minnesota (39)
• 35. USC (30)
• 46. Nebraska (47)
• 47. Washington (45)
• 52. Rutgers (61)
• 54. UCLA (65)
• 57. Wisconsin (54)
• 76. Michigan State (70)
• 80. Maryland (79)
• 89. Northwestern (89)
• 120. Purdue (117)