2014 UT Schedule Analysis....

[Sin City Vol]

I wish I could feel sorry for you. But, I can't. At least in your darkest hour, you don't have to walk into a room with a framed and signed Tebow jersey.

As a sports memorabilia collector I think that's awesome. Yea he played for Florida but you an autographed jersey in your house of maybe one of the top 10 college football players of all time and a Heisman Trophy winner. I have nothing like that but my stuff is still a work in progress. I honestly could care less who anyone played for as long as they were significant in their sport.

 

[General Jack]

:stop: All that jibber jabber wasn't scientific, it was statistical, dumb-butt

nvm

 

[General Jack]

Not necessarily. I would weigh the home team advantage as approaching 60/40 only when talent is very similar, like the Oklahoma game, or the SCAR game. But those aren't my numbers, so I can only use them in the context that they were used when I read the explanation of the study that provided them. I am still trying to understand how to reconcile that data with my own.

In any other situation, I would pick the team with more talent to win 70%, regardless of venue. Interesting to note that in BCS title games, that number rises to above 90% in favor of the team who recruited better over the preceding 4 years.

I think you can attribute that to a month plus of preparation and laser focus by players and staff. Don't upsets typically happen when the more talented team isn't focused or adequately prepared?

 

[sjt18]

There aren't an infinite number of outcomes, but I understand why you feel that way.

There are really only two outcomes, a win or a loss. There are a large number of variables that can change score differentials, and that can appear to approach infinity. The reason that Vegas doesn't typically allow betting on a simple winner or loser of a game, and instead uses points spreads, is because of this very fact.

Picking a winner or loser is relatively easy stuff and can be done with consistency, picking a score is a bit harder. I have just shown you how to pick the winner or loser at a rate approaching 70%, if not exceeding it.

I admit, there are very advanced models that exceed this and include an ability to beat the spread with consistency. Talent is a huge factor, regardless of how complex the system is.

I understand what you are doing and the advantage of simplification. But it fails to account for match ups. I'm not disagreeing in the broad sense. Much like recruiting rankings though what is "generally" accurate doesn't really translate to particular application.

The games that many here feel are likely losses for UT but that you suggest are 70% wins involve pretty different teams. Mizzou is built on a completely different philosophy than USCe or Ole Miss. If all three of those teams were a bad matchup for UT then the odds go much lower. Good matchup then much higher.

The similarity of style between USCe and Ole Miss gives me some encouragement about the chances of picking up a win in Oxford. IMO however the talent gap between UT and MU is wider and more meaningful.

 

[ntxVOL]

I went back and looked at my numbers again and found a mistake. So, if we consider Ole Miss, USC, and OK 50/50 games based upon recruiting. Then add 20% weighting only to the even or more talented home teams I get a different answer...6 wins.

 

[gnm53108]

You said it.

She's a good woman, and about to be the mother of our first child (due the first week of October, a girl).

This is how my wife told me we were having a child. I fell in love all over again (it doesn't hurt that my wife looks really really good in a bikini).

Haha...awesome.

Congratulations.

 

[daj2576]

I went back and looked at my numbers again and found a mistake. So, if we consider Ole Miss, USC, and OK 50/50 games based upon recruiting. Then add 20% weighting only to the even or more talented home teams I get a different answer...6 wins.

Interesting. I still don't know if I would call any of those a 50/50 game, though. Ole Miss, I would put as a solid 70/30 in our favor, and Oklahoma and SCAR would be 60/40 in their favor, at most. Another complicating factor in these simple evaluations is that they don't account for coach effect or other factors. Some coaches have a significant effect on talent, and a history to suggest that is repeatable. For instance, Petrino and Jones both have a significant positive effect on talent over their history, whereas Dooley, Kiff and 'Champ all have a historical negative effect on talent. As an example, here is Jones' history against talent predictions:



Those are some of the many drawbacks of my "system." When coming up with a "predicted" outcome, it chalks up a win for any game that is 70/30 and a loss for any game that is 30/70. That isn't consistent with the way the probability, randomness or error, will propagate through a system but it gives the viewer an idea of what games, on an individual basis, any team should win or lose.

I don't want to further complicate it though, as my numbers show that on any given year, about 80% of SEC teams will fall within 2 games of the talent predicted seasonal outcomes. About half of the SEC in any given year will perform EXACTLY as these talent based predictions show. I think that is a reasonable range of understanding.

Here is the 2013 season if you want to see how all SEC teams did in league play versus their talent evaluations.



Notice that the only real statistical anomalies are Florida and Mizzou. Florida and Tennessee have been battling for the biggest under-performers in the SEC approaching the last decade, whereas Mizzou tends to perform about where talent predicts (last year being a glaring exception largely based on Muschamp, injuries to UGA, and a UT team in a new system, low on depth, etc.).

Here is the 2014 season boiled down to predictions:



Oh, as a final note, I wanted to track how well my predictions did (If I Bleed Orange... - my blog) against "the professionals." Here is the talent evaluations predictions versus Athlon magazines 2013 predictions versus actual outcome:

 

[KBVol]

I wish I could feel sorry for you. But, I can't. At least in your darkest hour, you don't have to walk into a room with a framed and signed Tebow jersey.

Touché. Game, set, match to daj.

 

[KBVol]

You said it.

She's a good woman, and about to be the mother of our first child (due the first week of October, a girl).

This is how my wife told me we were having a child. I fell in love all over again (it doesn't hurt that my wife looks really really good in a bikini).

Congrats daj! On the child (and your wife in a bikini&#128521

 

[daj2576]

Congrats daj! On the child (and your wife in a bikini&#128521

Thank you! I never thought I would be so excited to be a father, but I am. I've already picked out little UT cheerleader outfits, and the like, for the baby to wear as an infant. My wife is totally on board with that, by the way, which is a testimony to her grace and love of me as a Vol-grad.

I do have to give my wife a ton of credit for her beach body. She is a personal trainer and a yoga instructor, so that balances out her being a gator (somewhat :birgits_giggle. Plus she can do more pushups, pullups and situps than I can. lol

 

[daj2576]

I think you can attribute that to a month plus of preparation and laser focus by players and staff. Don't upsets typically happen when the more talented team isn't focused or adequately prepared?

Interesting that you ask that.

Check out page 14 of this pdf. Notice that Butch Jones has a winning percentage of about 59% under similar circumstances to the ones you described.

View attachment The-Saturday-Edge-eBook-PDF.pdf

This is actually a really interesting read, all around. There is much I agree with and some things I don't, but it is enlightening, none the less.

 

[ntxVOL]

Interesting. I still don't know if I would call any of those a 50/50 game, though. Ole Miss, I would put as a solid 70/30 in our favor, and Oklahoma and SCAR would be 60/40 in their favor, at most. Another complicating factor in these simple evaluations is that they don't account for coach effect or other factors. Some coaches have a significant effect on talent, and a history to suggest that is repeatable. For instance, Petrino and Jones both have a significant positive effect on talent over their history, whereas Dooley, Kiff and 'Champ all have a historical negative effect on talent. As an example, here is Jones' history against talent predictions:

I did weight those games, although I weighted Ole Miss 60/40 as well. I weighted the Ole Miss game against us because their recruiting is not really that far behind us and their coach also seems to be making the most of the talent they have.

I see what your saying though and find it a hard pill to swallow that we have under-performed so badly the last few years. I agree that keeping it simple is probably the best as we can finagle the numbers in all sorts of ways but it would only muddle them even more.

 

[daj2576]

As a sports memorabilia collector I think that's awesome. Yea he played for Florida but you an autographed jersey in your house of maybe one of the top 10 college football players of all time and a Heisman Trophy winner. I have nothing like that but my stuff is still a work in progress. I honestly could care less who anyone played for as long as they were significant in their sport.

I have to admit it is really cool, even if it makes me sick. It also helped my understanding of the debate over Nike vs. Adidas. I also have a game worn UT jersey in my office. Having held both (but worn neither) it seems the Adidas is actually the superior product. This is most certainly another conversation for another thread.

 

[KBVol]

Thank you! I never thought I would be so excited to be a father, but I am. I've already picked out little UT cheerleader outfits, and the like, for the baby to wear as an infant. My wife is totally on board with that, by the way, which is a testimony to her grace and love of me as a Vol-grad.

I do have to give my wife a ton of credit for her beach body. She is a personal trainer and a yoga instructor, so that balances out her being a gator (somewhat :birgits_giggle. Plus she can do more pushups, pullups and situps than I can. lol

With all due respect..... personal trainer and yoga instructor balances/counteracts her gatordom by more than a substantial amount. Well done sir.

 

[daj2576]

I did weight those games, although I weighted Ole Miss 60/40 as well. I weighted the Ole Miss game against us because their recruiting is not really that far behind us and their coach also seems to be making the most of the talent they have.

I see what your saying though and find it a hard pill to swallow that we have under-performed so badly the last few years. I agree that keeping it simple is probably the best as we can finagle the numbers in all sorts of ways but it would only muddle them even more.

I'm actually excited that you said this! This is where perception doesn't mesh with reality, and also the reason that I think my evaluation is valuable. Freeze is increasing talent at Ole Miss, but I don't find that he is actually exceeding any talent based results. This sort of thing tends to be magnified when one coach comes in and performs with competence, when previous coaches at the same school couldn't do that. That is why Malzahn looks so great. He didn't make an Auburn team super human, he just showed that Chizik could make a super human team bad.

I admit that it is terribly easy to fall prey to hype when the media and other sports fans are telling you otherwise. Maybe even your own eyes are pleading with you to conclude differently. I call this the Saban conundrum. Saban is REALLY good at winning games but in relation to talent he actually loses more games than he should. He has a long history of that. His history shows that he doesn't tend to do much better than the classes of talent he brings in, but he does tend to exponentially increase talent. The end result is success, but not for the reasons most believe. See also Fulmer pre 2002ish.

 

[daj2576]

With all due respect..... personal trainer and yoga instructor balances/counteracts her gatordom by more than a substantial amount. Well done sir.

Lol. Thank you.

Let me offer this tid-bit to counteract your perception. It is more difficult than you would believe to watch her team beat mine with consistency, and then she smashes me in the weight room too.

She asks for no quarter, and gives none. ha.

 

[99gator]

I wish I could feel sorry for you. But, I can't. At least in your darkest hour, you don't have to walk into a room with a framed and signed Tebow jersey.

you are blessed.

 

[daj2576]

you are blessed.

I do have to kneel when I enter that room, if that's what you mean. lol

 

[Catbone]

Resisting.... to make.... perverted joke... aghhh

 

[NotoriousB-I-G]

I said that UF is the most probable upset, not that it was a sure thing, definite, or even likely. And yes, being at home is an advantage, especially against under-achieving coaches.

Also, I can find no consistent time during modern history where UT actually had a four year trailing recruiting average that was better than UF's. That in itself can explain why the series is so lopsided.

You would probably be interested in reading the book "Mathletics" or "Scorecasting", especially the discussion about how bad people are at understanding runs in a series. Many people believe that the preceding event effects the next. I don't, I think they are independent events. Studies have been done that tend to side with me, but I wouldn't say those studies are definitive.

Even in a coin toss, where both outcomes are as likely as the other, there is a much higher chance of a long run of similar outcomes than most people would ever believe. As I mentioned above, I cannot find a time in this series where the outcome should be a coin toss. It is typically 70/30 in favor of UF.

I suggest looking up the so-called "Wald Wolfowitz Runs Test" which illustrates this very thing. Put in a baseball perspective, many people believe a team gets hot during a season. But, math shows that in a 162 game season, that there is a "20% chance that a team with a .60 probability of winning each game will have a winning steak longer than ten games during the season." (see: Mathletics, pg. 91).

That's a helluva lot of homework and mental gymnastics to understand 40 years of butt-whoopings. Do I have to enroll at Princeton to understand 9 straight to those ********?

Even the agreed upon underachieving Muschamp beat the crap out of us two years ago.

By the way, "underachieve" is a perfectly accpetable word, no need to hyphenate it. You can find that word and other accpetable words here: (see: Underachieve | Define Underachieve at Dictionary.com) page 148.)

eace2:

 

[daj2576]

That's a helluva lot of homework and mental gynastics to understand 40 years of butt-whoopings. Do I have to enroll at Princeton to understand 9 straight to those ********?

Even the agreed upon underachieving Muschamp beat the crap out of us two years ago.

By the way, "underachieve" is a perfectly accpetable word, no need to hyphenate it. You can find that word and other accpetable words here: (see: Underachieve | Define Underachieve at Dictionary.com) page 148.)

eace2:

Thank you for the English lesson. I prefer to hyphenate-that-word, though, as I use both "under" and "over" as modifiers. Hyphenating urges the reader to pause and emphasize the difference instead of breezing over the word and perhaps missing that very important distinction.

Mental gymnastics sometimes are necessary to test if our perception of the world around us is legitimate, or just a mis-under-stand-ing (did I do that right?).

I doubt you need a degree-from Princeton-to participate in that sort of exercise. Of course, most people aren't interested in discovering that their perception might be wrong. "Facts be damned, I will feel how I want to feel" seems to be the form of data analysis most participate in. As Dave Bartoo suggests on his website (cfbmatrix.com) when he resurrects Neo and Morpheus from the Matrix... you can take the red pill or the blue pill, it is your choice.

What is the point of going back to an arbitrary point 40 years ago, by the way? Between 1974 and 1990, UT and UF only played 4 times. Yes, UF won all of those. It seems to me that if you are going to split the record, you should do it prior to 1990, and after 1990 (the point where the game was played every year). Prior to 1990, the UT v. UF series was played 19 times over 73 years. UT held a 13-6 advantage. Since 1990, the game has been played 24 times with UF holding an 18-6 lead. The historical record only swung in UF's favor over the past decade. Of note, most of that same past decade has had both UT trending well below expectations, and UF with a consistent and significant edge in recruiting.

It isn't like there is some magic to this understanding. The team who recruits better wins significantly more often than not. UF has done that. My point is to try to communicate to you, and others, that the run of nine wins is easily explainable while giving you sources on how to test this hypothesis on your own.

This is neither gymnastics, nor voodoo, nor Ivy League academics.

 

[KBVol]

you are blessed.

That's funny.

 

[OrangeIsLife]

If anybody thinks we're going 4-8, YOU BETTER CALL SOMEBODY!

 

[IluvdoubleD's]

Mods please file this away in archives we will win atleast 7 games or i will allow BTO to shave my head bald and stick a Vanderbilt star on it

oh yeah got 2 words for the teams we play this year. Suck It.

get some!

 

[NotoriousB-I-G]

Thank you for the English lesson. I prefer to hyphenate-that-word, though, as I use both "under" and "over" as modifiers. Hyphenating urges the reader to pause and emphasize the difference instead of breezing over the word and perhaps missing that very important distinction.

Mental gymnastics sometimes are necessary to test if our perception of the world around us is legitimate, or just a mis-under-stand-ing (did I do that right?).

I doubt you need a degree-from Princeton-to participate in that sort of exercise. Of course, most people aren't interested in discovering that their perception might be wrong. "Facts be damned, I will feel how I want to feel" seems to be the form of data analysis most participate in. As Dave Bartoo suggests on his website (cfbmatrix.com) when he resurrects Neo and Morpheus from the Matrix... you can take the red pill or the blue pill, it is your choice.

What is the point of going back to an arbitrary point 40 years ago, by the way? Between 1974 and 1990, UT and UF only played 4 times. Yes, UF won all of those. It seems to me that if you are going to split the record, you should do it prior to 1990, and after 1990 (the point where the game was played every year). Prior to 1990, the UT v. UF series was played 19 times over 73 years. UT held a 13-6 advantage. Since 1990, the game has been played 24 times with UF holding an 18-6 lead. The historical record only swung in UF's favor over the past decade. Of note, most of that same past decade has had both UT trending well below expectations, and UF with a consistent and significant edge in recruiting.

It isn't like there is some magic to this understanding. The team who recruits better wins significantly more often than not. UF has done that. My point is to try to communicate to you, and others, that the run of nine wins is easily explainable while giving you sources on how to test this hypothesis on your own.

This is neither gymnastics, nor voodoo, nor Ivy League academics.

Shorten' it up and I'll bother to read. This is a sports forum, not a seminar where some dude looks down his nose at the audience and lectures about how smart he is.

Check this out, you can benefit:

The Art of Writing for Brevity | The Daily Anchor