Talent Evaluations - Music City Bowl

#26
#26
Are you saying that you believe the B1G tends to perform significantly above expectations that are closely tied to talent? I would suggest not. And, that's a clear way to evaluate coach effect.

How about this, give me a couple of B1G teams that you believe are evidence of great coaching this season, and I will post their results.

Yeah, I think what this shows is a broad picture of the SEC being more difficult than it has ever been, giving Saban even more credit for what he is doing. Everyone says the SEC is down, and the East is garbage. I would argue it's as tough as its ever been and if we want to win the East consistently we need to be at the top of the recruiting rankings every year. So, are we honestly sure we can beat FL and GA every year in recruiting? I don't think so and I think it's the reason people need to have a realistic view of everything. Does your data go back to the 90's, particularly with that last chart above showing talent averages for all conferences?

Thanks much for your posts! :good!:
 
#27
#27
I don't see any obvious reason to exclude Bama.

But yes, that's what those charts tend to show. I'm not sure what conclusion you are drawing from it, if any.

What I have found is that within a couple of points, the home team should be favored, but not that margins of victory shrink or expand relative to talent differential (at least not in a way that can be sussed out with this evaluation alone).

Vandy is a bad loss, but they also over-performed significantly outside of normal distributions. It'sroughly as bad from a talent stand point as LSU v. Wisconsin, or OSU vs. PSU.

Bama was the only game we didn't compete in. The numbers show there is still a substantial gap That's really the only reason.

My conclusion would be that we played 3 teams with relatively the same numbers as UT. UGA/UF/A&M
We went 2-1 - Acceptable

We dropped 2 games to inferior talent. USC/Vandy. - Wildly Unacceptable (to borrow a word)
 
#28
#28
Bama was the only game we didn't compete in. The numbers show there is still a substantial gap That's really the only reason.

My conclusion would be that we played 3 teams with relatively the same numbers as UT. UGA/UF/A&M
We went 2-1 - Acceptable

We dropped 2 games to inferior talent. USC/Vandy. - Wildly Unacceptable (to borrow a word)

If you can show that the talent gap matters to the likelihood of wins/losses outside of 70/30, I'm all ears. I haven't found that, and I've been doing this for several years with hundreds of thousands of data points.

Interestingly, what I have found is that, generally speaking, the top half talented teams are more stable than the bottom half. Specifically, the top third talented teams show a great deal of stability in these predictions and the bottom third are the least stable.
 
Last edited:
  • Like
Reactions: 2 people
#30
#30
Big 10 was a better conference this year (and perhaps for a few more years) because of coaching. I think everybody would agree that SEC teams still get the best players.

Let me address this in a general way. The narrative is that the B1G is the best conference, ignoring talent, and because of coaching.

That means, explicitly, that you would need to prove that due to coaching, B1G teams would do better against the SEC than itself.

By way of illustration, let me pick a media darling. In this case, let's look at Wisconsin. I believe they have one of the better cases of coaching over-coming talent in the B1G. The question is then by how much, and how would that translate to the SEC? Let's look at Wisconsin and compare them to Vanderbilt. This is an apt comparison because their base level talents are similar, as are their over performances (note, for the average calculations below I threw out the two least talented teams either team played so as not to penalize Wisconsin for playing two teams that talent average can't judge, while Vandy only had one).

wiscy v vanderbilt.jpg

What this tells me is that Vandy actually played a tougher schedule and over-performed at a similar rate to Wisconsin. Vandy had a better average win, and a slightly worse average loss. Not how close Vandy's talent is to Wisconsin's. Note also how if you put Vanderbilt in Wisconsin's schedule, you would see similar predicted, and likely similar actual outcomes based on over-performance. This would be true with Wisconsin in Vanderbilt's schedule as well.

In other words, Vanderbilt and Wisconsin are similarly talented teams, over performing. The difference is that Wisconsin is in a far less talented conference meaning that their talent predicts a base of almost 50% more wins as a starting point. This, to me, is an indictment of the supposed coaching cap, that even if it exists (I would posit that overall the B1G performs closely to talent in the same way as the SEC, with a minority of coaches significantly over or under performing in predictible ways as illustrated by the difference in perception of Wiscy and Vandy), isn't big enough to truly over come the substantial talent gap between the conferences
 
Last edited:
  • Like
Reactions: 2 people
#31
#31
If you can show that the talent gap matters to the likelihood of wins/losses outside of 70/30, I'm all ears. I haven't found that, and I've been doing this for several years with hundreds of thousands of data points.

Interestingly, what I have found is that, generally speaking, the top half talented teams are more stable than the bottom half. Specifically, the top 3 talented teams show a great deal of stability in these predictions and the bottom third are the least stable.

I'm not try to show you anything. Just giving my simple conclusion to the talent #s you provided.

UT beat a slightly more talented UGA team

UT beat essentially an evenly talented UF team

UT got hammered by a more talented Bama team.

UT lost essentially to an evenly talented A&M

UT lost to lesser talented USC team

UT lost to a substantially less talented Vanderbilt team.

I take it you disagree me.
 
#32
#32
I'm not try to show you anything. Just giving my simple conclusion to the talent #s you provided.

UT beat a slightly more talented UGA team

UT beat essentially an evenly talented UF team

UT got hammered by a more talented Bama team.

UT lost essentially to an evenly talented A&M

UT lost to lesser talented USC team

UT lost to a substantially less talented Vanderbilt team.

I take it you disagree me.

Again, I'm just not sure what point you are making. I am the one that posted the graphics that show relative talent. If those agree with you, I do. If not, I don't.

Is there some nuance that I am missing, or are you just trying to convince yourself that whatever conclusion you started with is correct?
 
  • Like
Reactions: 2 people
#33
#33
Yeah, I think what this shows is a broad picture of the SEC being more difficult than it has ever been, giving Saban even more credit for what he is doing. Everyone says the SEC is down, and the East is garbage. I would argue it's as tough as its ever been and if we want to win the East consistently we need to be at the top of the recruiting rankings every year. So, are we honestly sure we can beat FL and GA every year in recruiting? I don't think so and I think it's the reason people need to have a realistic view of everything. Does your data go back to the 90's, particularly with that last chart above showing talent averages for all conferences?

Thanks much for your posts! :good!:

The SEC is tougher than it has ever been. As a point of comparison, Vandy and Kentucky fielded about the same level of talent this year as Bama did in 2005.

This data goes back to 2005 (the first year a four year trailing average of Rivals numbers could be created). I have data that goes back to the early 90's using a similar derivative, but I cannot use that as it's owned by my employer (last check it had 250,000 discrete data points).
 
Last edited:
  • Like
Reactions: 2 people
#34
#34
Again, I'm just not sure what point you are making. I am the one that posted the graphics that show relative talent. If those agree with you, I do. If not, I don't.

Is there some nuance that I am missing, or are you just trying to convince yourself that whatever conclusion you started with is correct?

Bama was the only game we didn't compete in. The numbers show there is still a substantial gap That's really the only reason.

My conclusion would be that we played 3 teams with relatively the same numbers as UT. UGA/UF/A&M
We went 2-1 - Acceptable

We dropped 2 games to inferior talent. USC/Vandy. - Wildly Unacceptable (to borrow a word)

If you can show that the talent gap matters to the likelihood of wins/losses outside of 70/30, I'm all ears. I haven't found that, and I've been doing this for several years with hundreds of thousands of data points.

Interestingly, what I have found is that, generally speaking, the top half talented teams are more stable than the bottom half. Specifically, the top third talented teams show a great deal of stability in these predictions and the bottom third are the least stable.

I'm not try to show you anything. Just giving my simple conclusion to the talent #s you provided.

UT beat a slightly more talented UGA team

UT beat essentially an evenly talented UF team

UT got hammered by a more talented Bama team.

UT lost essentially to an evenly talented A&M

UT lost to lesser talented USC team

UT lost to a substantially less talented Vanderbilt team.

I take it you disagree me.

I guess Im not quite sure why you responded with the bold given my posts.

I added some descriptions to my opinion or conclusion but followed your chart as far as where each team stands in talent. Unless I have a typo I missed.
 
#35
#35
I guess Im not quite sure why you responded with the bold given my posts.

I added some descriptions to my opinion or conclusion but followed your chart as far as where each team stands in talent. Unless I have a typo I missed.

You said we played teams with relatively similar talent. That's true, but the more talented team wins with consistency (so the differences, however small, do matter), but in a case of a *very small* margin, picking the home team as the winner stabilizes the results.
 
  • Like
Reactions: 2 people
#36
#36
You said we played teams with relatively similar talent. That's true, but the more talented team wins with consistency (so the differences, however small, do matter), but in a case of a *very small* margin, picking the home team as the winner stabilizes the results.

Thats fine.

I would add SC to the bad Ls column with Vandy. Not to the same degree but bad none the less.
 
#37
#37
Daj bottom line - UT should beat Nebraska
Beecher bottom line - UT will lose if it keeps playing the way it has
 
#38
#38
Daj bottom line - UT should beat Nebraska
Beecher bottom line - UT will lose if it keeps playing the way it has

Generally.

For a recent trend, UT is 3-2 in its last 5 games against teams with less relative talent. According to the chart daj provided Nebraska fits between SC and Vandy being closer to SC.

So UT "should" win the game.
 
#39
#39
After reading all of opening and most answers..... To state, as use to be stated on Laugh In...... I'am So Confused!!!!!!
 
#40
#40
As a data scientist, this makes me tingle in places we don't talk about at parties. I love it.

One big question for me: did you remove any players that have gone down with season-ending injuries and recompute?

Hurd, Tuttle, McKinsie, JRM, etc are on the roster... but they shouldn't be used when calculating the one-game prediction. Our talent level that we'll put on the field will be different than the roster talent level.

Suggestion for version 2 of this model: apply a multiplier for each year of experience in the program. A true freshman would receive a 0.75X multiplier, while a senior might get a 1.5X multiplier. If you stacked the team FULL of 4* and 5* players, but they were all true freshmen, that wouldn't be a very good team. A bunch of talented seniors is a different ballgame.
 
  • Like
Reactions: 2 people
#42
#42
Are you saying that you believe the B1G tends to perform significantly above expectations that are closely tied to talent? I would suggest not. And, that's a clear way to evaluate coach effect.

How about this, give me a couple of B1G teams that you believe are evidence of great coaching this season, and I will post their results.

The best example I could give of a coach that gets the most out of his players would be Dantonio, but they were awful this year and have lost everyone off of the really good 2013-15 teams.

You're a data guy so maybe the numbers don't bear this out, but I just think that the top 4 or 5 Big 10 coaches are better collectively than the top 4 or 5 SEC coaches. In the SEC there is a huge drop off after Saban, IMO. I don't even think it is clear who the second best coach in the SEC is after him. Is it McElwain? Sumlin? Malzahn? There is just so much mediocrity, in large part because Saban has gotten good coaches fired or they have left schools.

In the Big 10 you've got Urban, Harbaugh, and Franklin and overachievers like Dantonio, Ferentz, even Pat Fitzgerald. Paul Chryst isn't a bad coach and has gotten a lot out of somewhat limited talent at Wisconsin in his 2 years.
 
#43
#43
As a data scientist, this makes me tingle in places we don't talk about at parties. I love it.

One big question for me: did you remove any players that have gone down with season-ending injuries and recompute?

Hurd, Tuttle, McKinsie, JRM, etc are on the roster... but they shouldn't be used when calculating the one-game prediction. Our talent level that we'll put on the field will be different than the roster talent level.

Suggestion for version 2 of this model: apply a multiplier for each year of experience in the program. A true freshman would receive a 0.75X multiplier, while a senior might get a 1.5X multiplier. If you stacked the team FULL of 4* and 5* players, but they were all true freshmen, that wouldn't be a very good team. A bunch of talented seniors is a different ballgame.

You are correct, there are ways to increase the predictability of this model. But, at some point the time invested goes near vertical with very little return. As a free analysis, posted on a message board, used to illustrate likely outcomes and seasonal expectations 70/30 is about as good as you're going to get. Understand that there are more advanced analysis that I simply cannot, nor will I share, as I view those as being the property of my employer. I don't want to ever entangle my work with my employer's.

I have been doing this since approximately 2011, and I have tested almost every variable that one could imagine. In fact, I spent the better part of a year inputting and analyzing all sorts of variables and data for my employer. In years past, I have adjusted teams for attrition. I have also tried to weight teams based on experience. Both are popular ''what if'' scenarios. Neither really tweaked the outcome in a way that was worth the effort. In fact, of the samples that I tested, while the evaluation of a team might have adjusted in relation to itself, the relative evaluation didn't change much at all. In the rare event that there was a change, one way or another, it was rather insignificant. Are there cases where attrition and experience make huge differences? I am sure there are. They are the anomaly as far as I can tell.

My hypothesis is that talent is effected more by experience when the relative talent is low. For example, Pinkel at Missouri had an interesting trend. He recruited consistently in the mid thirties (quite mediocre). His coach effect trend line was 3 years or so of performing to talent expectations, then a two year bump of exceeding expectations. The irony is that he entered the SEC right as that explosion occurred, right on time. His recruiting was consistent, only the performance in relation to that relative talent changed. I believe that is a verifiable case of experience modifying talent. I don't tend to see that much with teams that recruit in the top 3rd. That isn't to say those situations don't exist, only that the change in attrition and experience are within ranges across the majority of teams to not greatly effect their relative position to each other.

I hope that answers your questions.

EDIT: As an aside, here is a graphic I used to gauge the coach effect of a few "hot name" coaches back to the 2014 season. Note Pinkel's performance in relation to talent to help illustrate the point I was making. If you were to add last year to it, the performance returned to expectations.

coaching trends.jpg
 
Last edited:
#44
#44
The best example I could give of a coach that gets the most out of his players would be Dantonio, but they were awful this year and have lost everyone off of the really good 2013-15 teams.

You're a data guy so maybe the numbers don't bear this out, but I just think that the top 4 or 5 Big 10 coaches are better collectively than the top 4 or 5 SEC coaches. In the SEC there is a huge drop off after Saban, IMO. I don't even think it is clear who the second best coach in the SEC is after him. Is it McElwain? Sumlin? Malzahn? There is just so much mediocrity, in large part because Saban has gotten good coaches fired or they have left schools.

In the Big 10 you've got Urban, Harbaugh, and Franklin and overachievers like Dantonio, Ferentz, even Pat Fitzgerald. Paul Chryst isn't a bad coach and has gotten a lot out of somewhat limited talent at Wisconsin in his 2 years.

It all depends on which statement you want to defend. If it is that the B1G has better coaches than the SEC, I would let you argue that all day because I view it as inconsequential to predictive analytics. If the statement is that the B1G is the better conference than the SEC despite a severe lack of talent, I would tell you the numbers don't bear that out.

Most coaches will simply perform tightly leashed to their talent. Even those that significantly over-perform given multiple seasons (Petrino at Arkansas / Spurrier at SCAR), don't typically recruit well enough to do much more than play spoiler to another team's season a couple of times a year (the reason for that might be that those coaches tend to run unique systems and look for unique players. These systems can shock a team not accustomed to it, but can and usually are defeated with significantly superior talent). The problem seems to be that as you win yourself into a position to face more talented teams, the numbers actually become more predictive and stable meaning that the lessor talented team has a less chance of over-performance (see the reliability of talent evals in national championship games to 2005).
 
  • Like
Reactions: 1 person
#45
#45
Hey Daj...just want to say I always enjoy your posts and I have been following this work since you started posting.

I know you have discussed attrition before, but I was wondering if the numbers across three classes amounts to 25 players, and then with the injuries we have sustained this season which seem abnormally high, especially at DT and LB where you have guys that are highly ranked but miss half the season (O'Brien, Tuttle, McKenzie, Sapp, JRM, etc...)

Is there a necessary adjustment or is it relatively even across all programs?
 
#46
#46
So what we are really saying is we went 8-4 not because Butch can't coach, but cause he can't recruit. :)

Or put another way, he under achieved by 1 game whereas everyone thinks he under achieved by 2 games. Or did he over achieve in 1 game and under achieve in 2?
 
  • Like
Reactions: 1 person
#47
#47
It all depends on which statement you want to defend. If it is that the B1G has better coaches than the SEC, I would let you argue that all day because I view it as inconsequential to predictive analytics. If the statement is that the B1G is the better conference than the SEC despite a severe lack of talent, I would tell you the numbers don't bear that out.

Most coaches will simply perform tightly leashed to their talent. Even those that significantly over-perform given multiple seasons (Petrino at Arkansas / Spurrier at SCAR), don't typically recruit well enough to do much more than play spoiler to another team's season a couple of times a year (the reason for that might be that those coaches tend to run unique systems and look for unique players. These systems can shock a team not accustomed to it, but can and usually are defeated with significantly superior talent). The problem seems to be that as you win yourself into a position to face more talented teams, the numbers actually become more predictive and stable meaning that the lessor talented team has a less chance of over-performance (see the reliability of talent evals in national championship games to 2005).

I think that bolded part might have occurred this season, but not over any other significant time period. I think that the Big 10 clearly had a better year as a conference despite being less talented than the SEC. Maybe it persists beyond this year - I think the Big 10 has clearly upgraded their coaching in recent years at two major schools in their conference (Michigan and Penn State). Those schools should recruit better and play better in coming seasons after several years of mediocrity.

Totally agree that coaches for the most part are tethered to where they recruit. There are a lot of people who like to knock star ratings and recruiting rankings but it is no coincidence that the teams that consistently recruit well according to the recruiting services are also the ones consistently playing for championships.
 
#48
#48
I believe we've had this conversation before, even if we hadn't, I will gladly look at your research that supports your conclusions.

What conclusions? I'm using your data.

Are you not stating that the higher rated team wins 70% of the time?

I simply made the observation that UT, aTm, UF, & UGA were all within the toss-up range of 3.50 points. UT played all 3 and went 2-1.

So UT starts the season 5-2 with 1 win it shouldn't have (UGA) and 2 losses it should have (aTm & Bama).

I see what you're saying by UT should have gone 9-3 based on the recruiting rankings but that doesn't excuse or dismiss losing to USC after a bye week or Vandy. At some point coaching comes into play.
 
  • Like
Reactions: 1 person
#49
#49
When teams recruit bigger, faster, and stronger athletes, they have a better chance of winning? :search:
 
#50
#50
I guess, daj,

Let me ask this another way. When you see UT lose in OT to aTm and you see UT beat App St. in OT, do you simply say, "Meh, it was predictable."

I don't see an issue with UT and aTm going to OT but App St. taking UT to OT had to be a head scratcher based on talent.
 
Advertisement



Back
Top