Back in the BCS era, "Margin of Victory" was a dirty term. At a certain point, BCS computers were required to strip MOV from their formulas if they wanted to remain in the BCS, even though the formula makers claimed doing so decreased their accuracy. There was a not very well disguised trade-off here: Including MOV in the BCS formula encouraged "running up the score" by coaches, which had the perceived negative side-effects of (a) being unsportsmanlike, and (b) could lead to more injuries, as coaches would keep starters in the game longer so as to run up the score, so it was formally banned from the computers even though according to the computer developers, doing so made those computers less accurate.

Now though, it seems pretty obvious that the CFP committee is using an "eyeball test" to judge teams, and there's no question that a large MOV has better optics than a narrow escape. So, for example, Baylor is likely to suffer because they "only" beat Kansas State by 7 points. That doesn't have very impressive optics. Notre Dame seems to be getting a lot of credit for barely losing to the #1 team on the road. I doubt they would be #5 if Clemson had beaten them 35-0.

Is the Eyeball Test a good thing, or should the CFP downplay MOV in its deliberations for the same reason the BCS did? Or is a heavy dose of MOV inevitable when humans are calling the shots?

I think it's a good point. I do think MOV should be included, but I'd like to see them put their hat on some sort of analytic instead of just the eye test. Be it a MOV heavy computer ranking like Sagarin predictor or ESPN FPI or a "game control" stat, use that and refer to it when explaining the rankings. The "eye test" is too subject to arbitrary analysis. I also think it is too subject to recency bias. Use an analytic for how dominating teams are that takes away some of the subjectivity.

(11-06-2015 07:14 AM)Frog in the Kitchen Sink Wrote:  I think it's a good point. I do think MOV should be included, but I'd like to see them put their hat on some sort of analytic instead of just the eye test. Be it a MOV heavy computer ranking like Sagarin predictor or ESPN FPI or a "game control" stat, use that and refer to it when explaining the rankings. The "eye test" is too subject to arbitrary analysis. I also think it is too subject to recency bias. Use an analytic for how dominating teams are that takes away some of the subjectivity.

I agree that a recency error is much more likely with the Eyeball standard too. A team wins a game impressively NOW and we forget about the close escape they had two months ago. Good point.

Alternatively, one could argue that we should be biased towards recent games, on the basis that it's a good thing if teams get better over time, and that this is relevant information for who makes the playoffs.

When the committee was created, "style points" became a major factor in how they slotted teams. FSU was the poster child for that last season.

not a big fan of MOV. There are better tie-breakers. The bottom line should be did you win or lose.

It should come down to conference championships, wins and losses, and strength of schedule.

One thing I have noticed about CFP committee is they seem to make up their own criteria for ranking each week and often contradict itself and then justify it with the old eyeball test.

But its fine, it will sort itself out in Nov and Dec. The actual rankings right now won't matter much a month from now.

I'd say a pretty good example of the MOV is actually Memphis. Don't think they would be #13 if they had beaten Ole Miss on a last second FG to win by 1 point.

I think the 'Eyeball Test'/MOV is a good thing, because it also shows how dominating a team is or isn't. I can't imagine if a team ranked #1 that keeps wining by 3 pts, is that good of a team, especially considering when they play the conference foes that only have 0 to 4 wins.

There should also be a measure regarding 1st half and 2nd half. Is a team that has a close game against a team with 0 to 4 wins after 3 qtrs, but then blows them out in the 4th as good a team the has the blowout in the 1st half?

I do think there is a point where running up the score has a zero effect in a margin of victory stat, especially is you are still using the starting players. If a team is using the 2nd and 3rd string players but still scoring, then I don't have a problem of that.

Baylor definitely benefitted in the first analysis from eye test. Last year's committee would have put them around 10.

Their human and these teams are closely approximated. Take the Massey composite ratings. Other than Clemson, the computers are all over the place ranking the top 10. And the means (scroll all the way to the right) for teams 2-10 are all clustered between 5 and 10. And the sd are pretty hefty, too. Baylor in particular, with a sd of 11, is all over the place with the MOV heavy rankings being high on them and the SOS rankings being low.

Previous comments in this thread show the all the reason why FBS needs to go to a 8 team playoff. Sports is meant to be settled on the field of battle not in back rooms by humans with all their bias' * teams would do the job 5 P5 spots, 1 G5 and 2 at large.

(11-06-2015 07:20 AM)quo vadis Wrote:  

(11-06-2015 07:14 AM)Frog in the Kitchen Sink Wrote:  I think it's a good point. I do think MOV should be included, but I'd like to see them put their hat on some sort of analytic instead of just the eye test. Be it a MOV heavy computer ranking like Sagarin predictor or ESPN FPI or a "game control" stat, use that and refer to it when explaining the rankings. The "eye test" is too subject to arbitrary analysis. I also think it is too subject to recency bias. Use an analytic for how dominating teams are that takes away some of the subjectivity.

I agree that a recency error is much more likely with the Eyeball standard too. A team wins a game impressively NOW and we forget about the close escape they had two months ago. Good point.

Alternatively, one could argue that we should be biased towards recent games, on the basis that it's a good thing if teams get better over time, and that this is relevant information for who makes the playoffs.

There is a whopping difference between EOV and MOV and it is EOV that should be considered not MOV. EOV (Ease of Victory) is a distinction that anyone who actually watches a game can determine. EOV is the level of control that one team exhibits over another during the competition. With EOV you aren't counting garbage points, or docking a really strong squad whose coach actually plays his subs in the second half of a game to both protect his starters and to earn his bench some game experience. Frequently during sub play a large margin amassed in the first half of a game actually gets whittled down to a much smaller but still distinctive win total. Great schools and great coaches shouldn't be penalized by these kinds of moves.

The committee should be required to watch the complete games of the top 12 schools on their ranking. Then without just looking at the point spread they would know if Team A controlled the game the first two quarters and gave their bench some meaningful time, or did they keep their first string in to the last play and continue to throw deep with 2:00 minutes left of the clock going for that 70th point.

To me Ease of Victory is much more revealing of the overall strength of a team than Margin of Victory. But you will never know that if you don't watch the whole game. So far this year Clemson is the school with the greatest Ease of Victory margin. They have really distanced themselves in the majority of their games without continuing to pour it on a hapless victim. L.S.U. is building in that category while Alabama is playing down in some cases and dominating in others. Both Notre Dame and Alabama have had tighter matches than first glance would indicate.

It's difficult to say about Baylor. Until last night they had really played nobody. And then last night in a 7 point win over Kansas State they were playing a sub at QB, albeit one that looked as seasoned as most folks #1. These last few weeks will tell the tale on Baylor and T.C.U..

(11-06-2015 09:11 AM)JRsec Wrote:  

(11-06-2015 07:20 AM)quo vadis Wrote:  

(11-06-2015 07:14 AM)Frog in the Kitchen Sink Wrote:  I think it's a good point. I do think MOV should be included, but I'd like to see them put their hat on some sort of analytic instead of just the eye test. Be it a MOV heavy computer ranking like Sagarin predictor or ESPN FPI or a "game control" stat, use that and refer to it when explaining the rankings. The "eye test" is too subject to arbitrary analysis. I also think it is too subject to recency bias. Use an analytic for how dominating teams are that takes away some of the subjectivity.

I agree that a recency error is much more likely with the Eyeball standard too. A team wins a game impressively NOW and we forget about the close escape they had two months ago. Good point.

Alternatively, one could argue that we should be biased towards recent games, on the basis that it's a good thing if teams get better over time, and that this is relevant information for who makes the playoffs.

There is a whopping difference between EOV and MOV and it is EOV that should be considered not MOV. EOV (Ease of Victory) is a distinction that anyone who actually watches a game can determine. EOV is the level of control that one team exhibits over another during the competition. With EOV you aren't counting garbage points, or docking a really strong squad whose coach actually plays his subs in the second half of a game to both protect his starters and to earn his bench some game experience. Frequently during sub play a large margin amassed in the first half of a game actually gets whittled down to a much smaller but still distinctive win total. Great schools and great coaches shouldn't be penalized by these kinds of moves.

The committee should be required to watch the complete games of the top 12 schools on their ranking. Then without just looking at the point spread they would know if Team A controlled the game the first two quarters and gave their bench some meaningful time, or did they keep their first string in to the last play and continue to throw deep with 2:00 minutes left of the clock going for that 70th point.

To me Ease of Victory is much more revealing of the overall strength of a team than Margin of Victory. But you will never know that if you don't watch the whole game. So far this year Clemson is the school with the greatest Ease of Victory margin. They have really distanced themselves in the majority of their games without continuing to pour it on a hapless victim. L.S.U. is building in that category while Alabama is playing down in some cases and dominating in others. Both Notre Dame and Alabama have had tighter matches than first glance would indicate.

It's difficult to say about Baylor. Until last night they had really played nobody. And then last night in a 7 point win over Kansas State they were playing a sub at QB, albeit one that looked as seasoned as most folks #1. These last few weeks will tell the tale on Baylor and T.C.U..

That is a better illustration and explanation and thought of my 1st half vs 2nd half comment. Good post and agree with your ease of victory insight.

I think ease of victory is akin to ESPN's "game control" metric. I like that too, as scores can be deceiving both big and small.

http://espn.go.com/college-football/playoffPicture

I'd like to see the committee hang their hats on that sort of thing more. Changing the subject slightly, but speaking of hanging their hat, I was annoyed with the "wins over .500+ teams" arguments Long was making in defense of their rankings. First off, not all .500+ teams are created equal, and second off, it seemed like he applied that defense arbitrarily. That is not a good metric to use, IMO. Too flawed.

Not a fan of "the eyeball test". There have been times throughout history where teams that "pass the eyeball test" are not that very good. Conversely there have been teams who win "ugly" who are very good. To illustrate I will use the example of one of the Chip Kelly Oregon teams. They spanked everyone all year long and set all kids of offensive records. They "passed the eyeball test". One year they went to the Rose Bowl and got waxed by a Jim Tressel era Ohio State team that did not play "sexy" football. In the current set up that Oregon team would have easily been a top 4, but in reality they were probably not one of the 4 best teams in football that year.

(11-06-2015 09:39 AM)CliftonAve Wrote:  Not a fan of "the eyeball test". There have been times throughout history where teams that "pass the eyeball test" are not that very good. Conversely there have been teams who win "ugly" who are very good. To illustrate I will use the example of one of the Chip Kelly Oregon teams. They spanked everyone all year long and set all kids of offensive records. They "passed the eyeball test". One year they went to the Rose Bowl and got waxed by a Jim Tressel era Ohio State team that did not play "sexy" football. In the current set up that Oregon team would have easily been a top 4, but in reality they were probably not one of the 4 best teams in football that year.

The year that you are talking about is 2009..... Oregon didn't win as big as you say...
Purdue 2
Utah 7
UCLA 14
Arizona 3 in 2 OT
Oregon St 4

Ohio St was actually more impressive. The 7 Big Ten games they won- they average won by 18.4 PPG.

(11-06-2015 09:55 AM)stever20 Wrote:  

(11-06-2015 09:39 AM)CliftonAve Wrote:  Not a fan of "the eyeball test". There have been times throughout history where teams that "pass the eyeball test" are not that very good. Conversely there have been teams who win "ugly" who are very good. To illustrate I will use the example of one of the Chip Kelly Oregon teams. They spanked everyone all year long and set all kids of offensive records. They "passed the eyeball test". One year they went to the Rose Bowl and got waxed by a Jim Tressel era Ohio State team that did not play "sexy" football. In the current set up that Oregon team would have easily been a top 4, but in reality they were probably not one of the 4 best teams in football that year.

The year that you are talking about is 2009..... Oregon didn't win as big as you say...
Purdue 2
Utah 7
UCLA 14
Arizona 3 in 2 OT
Oregon St 4

Ohio St was actually more impressive. The 7 Big Ten games they won- they average won by 18.4 PPG.

I am a little off on the margin of victory, but if you put yourself back into that time period you will recall everyone thought Oregon was sexy as hell and Ohio State was the beneficiary of playing in a weak arse conference that played boring football.

FWIW, I am not trying to get Ohio State props here. There are a number of other programs that have fallen into these categories over the years.

dumb thread....the committee has said over and over that they are using a wide variety of factors...

They're not just using the "eye test"....people who keep guessing at what they're trying to base their rankings on don't get it

all of this analysis and splitting hairs is pointless at this point....I think the rankings are very fluid, and what happened one week may not carry over to the next week.

For example, I don't think Oklahoma State needs a lot of teams in front of them to lose to jump up in the rankings....beat TCU, and I bet they are in the top 6, if not the top 4

(11-06-2015 09:58 AM)CliftonAve Wrote:  

(11-06-2015 09:55 AM)stever20 Wrote:  

(11-06-2015 09:39 AM)CliftonAve Wrote:  Not a fan of "the eyeball test". There have been times throughout history where teams that "pass the eyeball test" are not that very good. Conversely there have been teams who win "ugly" who are very good. To illustrate I will use the example of one of the Chip Kelly Oregon teams. They spanked everyone all year long and set all kids of offensive records. They "passed the eyeball test". One year they went to the Rose Bowl and got waxed by a Jim Tressel era Ohio State team that did not play "sexy" football. In the current set up that Oregon team would have easily been a top 4, but in reality they were probably not one of the 4 best teams in football that year.

The year that you are talking about is 2009..... Oregon didn't win as big as you say...
Purdue 2
Utah 7
UCLA 14
Arizona 3 in 2 OT
Oregon St 4

Ohio St was actually more impressive. The 7 Big Ten games they won- they average won by 18.4 PPG.

I am a little off on the margin of victory, but if you put yourself back into that time period you will recall everyone thought Oregon was sexy as hell and Ohio State was the beneficiary of playing in a weak arse conference that played boring football.

FWIW, I am not trying to get Ohio State props here. There are a number of other programs that have fallen into these categories over the years.

actually that was one of Ohio St's more sexy teams quite frankly. And also one of Oregon's more gritty teams(this was Kelly's 1st year- had the Boise St debacle- but they overcame that). In future years- Oregon got sexier and sexier- but in 2009- not so much.

(11-06-2015 10:00 AM)EvilVodka Wrote:  dumb thread....the committee has said over and over that they are using a wide variety of factors...

They're not just using the "eye test"....people who keep guessing at what they're trying to base their rankings on don't get it

all of this analysis and splitting hairs is pointless at this point....I think the rankings are very fluid, and what happened one week may not carry over to the next week.

For example, I don't think Oklahoma State needs a lot of teams in front of them to lose to jump up in the rankings....beat TCU, and I bet they are in the top 6, if not the top 4

I don't think so. I don't think the committee has much respect at all for the big 12. So wins in the conference don't mean anywhere near as much as if it was in a different conference. I think part of the conference's problem frankly is Big 12 defenses are absolute garbage.

(11-06-2015 10:09 AM)stever20 Wrote:  

(11-06-2015 10:00 AM)EvilVodka Wrote:  dumb thread....the committee has said over and over that they are using a wide variety of factors...

They're not just using the "eye test"....people who keep guessing at what they're trying to base their rankings on don't get it

all of this analysis and splitting hairs is pointless at this point....I think the rankings are very fluid, and what happened one week may not carry over to the next week.

For example, I don't think Oklahoma State needs a lot of teams in front of them to lose to jump up in the rankings....beat TCU, and I bet they are in the top 6, if not the top 4

I don't think so. I don't think the committee has much respect at all for the big 12.

well ya, because they haven't played anyone....they all play each other though...whoever comes through unscathed will be high in the rankings.

I do think the committee will penalize the Big XII for not having a championship game though. I think their champ will have to be undefeated to get in. An 11-1 TCU is going to get passed by a 12-1 conference championship winner every time

(11-06-2015 09:11 AM)JRsec Wrote:  

(11-06-2015 07:20 AM)quo vadis Wrote:  

(11-06-2015 07:14 AM)Frog in the Kitchen Sink Wrote:  I think it's a good point. I do think MOV should be included, but I'd like to see them put their hat on some sort of analytic instead of just the eye test. Be it a MOV heavy computer ranking like Sagarin predictor or ESPN FPI or a "game control" stat, use that and refer to it when explaining the rankings. The "eye test" is too subject to arbitrary analysis. I also think it is too subject to recency bias. Use an analytic for how dominating teams are that takes away some of the subjectivity.

I agree that a recency error is much more likely with the Eyeball standard too. A team wins a game impressively NOW and we forget about the close escape they had two months ago. Good point.

Alternatively, one could argue that we should be biased towards recent games, on the basis that it's a good thing if teams get better over time, and that this is relevant information for who makes the playoffs.

There is a whopping difference between EOV and MOV and it is EOV that should be considered not MOV. EOV (Ease of Victory) is a distinction that anyone who actually watches a game can determine. EOV is the level of control that one team exhibits over another during the competition. With EOV you aren't counting garbage points, or docking a really strong squad whose coach actually plays his subs in the second half of a game to both protect his starters and to earn his bench some game experience.

I would disagree about the ease of determining EOV. Baylor last night is a good example. On one hand, they led the whole game, and had a huge lead, 31-10, early in the 4th quarter. On the other, they almost let that lead slip away. Overall, I say it was a close game. That's what the scoreboard says.

All 60 minutes in a game count.