ChiPOTLE 2016

[CHONGS]

End of the Regular Season Update

Top 25


Alabama is the clear number 1.  They have the second best efficiency stats and the best accomplishments.  2-4 are all very close and probably interchangeable.  This year was a real missed opportunity for Michigan.  That overtime loss was devastating. Interesting to see that Stanford at 14 still ended up as the best team we faced this year.

Big 12

Our best win of the season (over #56 TCU) is rewarded by getting KSU into the top 50 for this first time this year and rated 4th in the conference.  KSU ends up with the #3 (#49 nationally) defense and the #4 (#36 nationally) offense. 

Here are teams comparable to KSU:


Conferences

Here is how the P5 conferences ended up.  Each bar is 5 wide, so the Big Ten has 2 teams in the top 5, and 1 team ranked between 6 and 10.

Judging the strength of a conference can be tricky, there are many different philosophies.  I tend to think one should judge a conference by the top half.  In that case we have:

where the second column is the average ranking of the top half of the conference.  Here the weakness of the Big 12 is clear, we have no elite teams. It might surprise some to see the ACC ranked the highest.

If you want the whole conference:

While the Big Ten has 5 teams in the top 25 (tied with the ACC for the most of any conference), team like Purdue (119),  Illinois (124), and Rutgers (127) drag them waaay down.

KSU Weekly

OS and DS where places where KSU improved the most.   It was nice to see us drop the terrible DS of 2.80 down to something more reasonable and lower than our DE.  Over the past 6 weeks the offense has been pretty steady, but the defense has improved.

Texas Bowl Matchup

KSU with a 57% chance to win! 

The rating difference between us and A&M is solely due to their quality of wins this year.  By the "numbers" we have a slightly more efficient team.

[CHONGS]

Also out of the 60 P5 + Notre Dame teams, KSU is #31.  Smack dab in the middle.  Almost the very definition of an average Power 5 team.

[MakeItRain]

Also out of the 60 P5 + Notre Dame teams, KSU is #31.  Smack dab in the middle.  Almost the very definition of an average Power 5 team.

Speaking of Notre Dame, how are they that high with their four amazing wins of Nevada, Syracuse, Army, and Miami? That's a murders row of wins.

App State ran the gauntlet too, lots of quality wins. I have no idea how they went through the three week stretch of Idaho, Georgia Southern, and Texas State unscathed.

[CHONGS]

Also out of the 60 P5 + Notre Dame teams, KSU is #31.  Smack dab in the middle.  Almost the very definition of an average Power 5 team.

Speaking of Notre Dame, how are they that high with their four amazing wins of Nevada, Syracuse, Army, and Miami? That's a murders row of wins.

App State ran the gauntlet too, lots of quality wins. I have no idea how they went through the three week stretch of Idaho, Georgia Southern, and Texas State unscathed.

Kansas State wins:
56 TCU
60 Texas
62 Baylor
77 Texas Tech
101 Iowa State
115 Florida Atlantic
125 Kansas

Notre Dame wins:
27 Miami
67 Army
90 Syracuse
100 Nevada

App State wins:
52 Old Dominion
84 Idaho
87 Louisiana Lafayette
91 Georgia Southern
94 Akron
108 Georgia State
112 New Mexico State
116 Louisiana Monroe
128 Texas State

For Notre Dame it came down to Miami, which is a lot better than any win we have.  They also have a bit higher WP.   App state also has a slightly better win (against #52 Old Dominion), but also notice that App State actually has a smaller ACC score than KSU. They are, however, 0.14 higher in WP.  That is why they are higher.

[Rage Against the McKee]

Does your system throw out all results against FCS competition?

[CHONGS]

Does your system throw out all results against FCS competition?

Yes

[mocat]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

[Rage Against the McKee]

Throwing out FCS results probably inflates WSUs ranking a little bit, right?

[CHONGS]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

Why?

[CHONGS]

Throwing out FCS results probably inflates WSUs ranking a little bit, right?

Maybe, but losses arent explicitly punished.

[mocat]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

Why?

i think those wins are dragging them down. better just to lose to them (like they did to -these are snowbrag rankings- #74 texas, #99 michigan state, #97 duke, and #55 nc state) to keep that average win strength sky high

[mocat]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

Why?

or, more succinctly:

losses arent explicitly punished.

[CHONGS]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

Why?

i think those wins are dragging them down. better just to lose to them (like they did to -these are snowbrag rankings- #74 texas, #99 michigan state, #97 duke, and #55 nc state) to keep that average win strength sky high

I don't take an average.  Its cumulative. You get rewarded for wins.

[CHONGS]

sounds like in the chipotle system, notre dame could have an even better accomplishments had they managed to lose to Army, Syracuse, and Nevada

Why?

or, more succinctly:

losses arent explicitly punished.

Losses aren't explicilty punished, but they hurt you nonetheless because they cost you the opportunity for getting points for wins.

My rules of thumb:  a win will never cost you raw ACC points AND a loss will never get you raw ACC points. 

[mocat]

#27 nd
#56 ksu
#60 ksu
#62 ksu
#67 nd
#77 ksu
#90 nd
#100 nd
#101 ksu
#115 ksu
#125 ksu


essentially what chipotle boils down to is that

1 win vs #27 is more valuable than 3 wins vs #56, 60, and 62, because everything below that sort of washes out

[MakeItRain]

The problem with not not giving credit for quality losses is that a 9-3 sun belt school with two FCS wins ends up with a higher rating than a power 5 school with five losses. The Sun Belt and C-USA have far too much weight because they didn't beat anyone at all in the non con they the good teams in these conferences all beat each other giving the appearance of a good win.

What is this supposed to measure? Best team? Is the 20th team expected to be significantly better than the 63rd?

If Iowa State played ODU on a neutral field ISU would by all accounts be a favorite, you have ODU 50 spots better. You would have ISU as ODUs 7th best win, that's prosperous.

[Rage Against the McKee]

It's not a ranking of who is the best. It's a ranking of who has accomplished the most. I think it works for what it is intended to do. It's really the sort of system that should be used to set the playoff field.

[CHONGS]

#27 nd
#56 ksu
#60 ksu
#62 ksu
#67 nd
#77 ksu
#90 nd
#100 nd
#101 ksu
#115 ksu
#125 ksu


essentially what chipotle boils down to is that

1 win vs #27 is more valuable than 3 wins vs #56, 60, and 62, because everything below that sort of washes out

Close. Here is the value of each win:

[CHONGS]

The problem with not not giving credit for quality losses is that a 9-3 sun belt school with two FCS wins ends up with a higher rating than a power 5 school with five losses. The Sun Belt and C-USA have far too much weight because they didn't beat anyone at all in the non con they the good teams in these conferences all beat each other giving the appearance of a good win.

What is this supposed to measure? Best team? Is the 20th team expected to be significantly better than the 63rd?

If Iowa State played ODU on a neutral field ISU would by all accounts be a favorite, you have ODU 50 spots better. You would have ISU as ODUs 7th best win, that's prosperous.

ChiPOTLE is not really a predictor, it is a listing sorted by accomplishments.  Teams listed to the top "deserve" to be there based on their statistical numbers and the teams they have beaten.

Of course we will use OE and DE (points per drive on offense and defense), but boy oh boy are there are a couple of major changes!  There are now two more metrics used to calculate the expected Win % (WP):

OS: Offensive Stamina --- number of points scored on offense per drive when staring from deep within your own territory
DS: Defensive Stability --- number of points given up by the defense per drive when the opponent is starting from deep withing their own territory

And

Acc: Accomplishments --- on a scale between 0 and 1, how impressive are your wins? There is no credit for simply playing good teams, you must beat the team to get points here.  The team with 1 for that week has the best accomplishments.

ChiPOTLE is calculated by taking the WP (based on OE, DE, OS, and DS) and combining that with Acc.

[MakeItRain]

So winning a home game against Toledo is a greater accomplishment than winning a road game against Nebraska?

[CHONGS]

So winning a home game against Toledo is a greater accomplishment than winning a road game against Nebraska?

In this objective system they would actually be worth the same (0.25 raw points).   The only inputs are OE, DE, OS, DS, and who you have beaten.  I don't give credit for being a "name" team or how big your stadium is.

[MakeItRain]

So winning a home game against Toledo is a greater accomplishment than winning a road game against Nebraska?

In this objective system they would actually be worth the same (0.25 raw points).   The only inputs are OE, DE, OS, DS, and who you have beaten.  I don't give credit for being a "name" team or how big your stadium is.

The circumstances in which a game is being played is somewhat important, especially if you're using it as a measure of playoff worthiness.

[CHONGS]

Perhaps, but I haven't figured out an objective way to quantify these circumstances.  What should be used?  Stadium size? Ticket sales?  Opponent's home win percentage  over the past three years?  All possibly right answers, but in the end it needs to be objectively quantified.

[MakeItRain]

Perhaps, but I haven't figured out an objective way to quantify these circumstances.  What should be used?  Stadium size? Ticket sales?  Opponent's home win percentage  over the past three years?  All possibly right answers, but in the end it needs to be objectively quantified.

I'm feeling a little facetiousness here. But to answer the question that I don't think you are looking for is that I'm not sure any of these individual metrics will give you that answer. I think ultimately the flaw with ChiPOTLE is that it is too stripped down. We are not watching robots here. You and I and anyone with a tiny bit of objectivity know a 21-17 win for Mississippi State over a 4-2 Auburn in the plains is a greater accomplishment than Toledo beating a 7-0 Kent State 35-27 in the Glass Bowl. It appears that ChiPOTLE views these two events more closely than it should. I don't know what the specific metric is, you're much smarter than me when it comes to that, but it seems the more data points you have the more likely these things would be fleshed out.

[CHONGS]

Post Bowl Season Update

Top 25


OU finally gets into the top 10.

Big 12

I suspect getting jumped by Baylor is going to not make sense to a lot of people.  They did have two really good wins though (vs OSU and BSU).

Going in to next year we will have a pretty efficient offense (ranked #3 in the conference and #34 overall) and defense (#3 in Big 12, #51 overall)

2016 Recap


39 Baylor --> 0.25
42 Texas A&M --> 0.25
61 TCU --> 0.2
62 Texas --> 0.2
76 Texas Tech --> 0.15
101 Iowa State --> 0.1
114 Florida Atlantic --> 0
126 Kansas --> 0

2017 Preview