The 615 Index Simulates the Season

[curmudgeon]

Football games aren't played on paper, or on Xbox, but a complex statistical formula can make for interesting conversation two weeks before kickoff.

First, some basics: 3.6 points is given to the home team, based on results of every game involving a BCS team the last three years. This is up slightly from last year, when the index gave a 3.4 point advantage.

Most of our scoring returns, and with QBs and WRs receiving equal credit for passing TDs in my formula, we have a high amount of returning scoring. Defensively, the index just looks at stats and returning starters. Losing two CBs hurts us because the index penalizes us against teams with big passing offenses, as you will read below.

Last year, the index had us at 8-4, with us beating Ole Miss and losing to Arkansas.

First some national predictions.

SEC Champion: Alabama (13-0)
Big 12 Champion: Oklahoma State (10-2)
Big Ten Champion: Ohio State (12-1)
Pac 12 Champion: Stanford (11-1)
ACC Champion: Clemson (11-2)
AAC Champion: Louisville (12-0)

National Championship: Alabama over Louisville

Our season:
Oklahoma State (Neutral Site). Loss 45-24 (0-1)
Alcorn State (FBS, not included in simulation), Win (1-1)
at Auburn. Win 31-21 (2-1, 1-0)
Troy. Win 38-17 (3-1)
LSU. Win 21-17 (4-1, 2-0)
Bowling Green. Win 42-24 (5-1)
Kentucky. Win 34-15 (6-1, 3-0)
at South Carolina. Loss 28-23 (6-2, 3-1)
at Texas A&M. Loss 38-14 (6-3, 3-2)
Alabama. Loss 27-10 (6-4, 3-3)
at Arkansas. Win 32-20 (7-4, 4-3)
Ole Miss. Win 24-17 (8-4, 5-3)

SEC West Standings
Alabama 13-0, 8-0
Texas A&M 11-1, 7-1 (Loss to Alabama)
MSU 8-4, 5-3 (Losses to OSU, USC, Alabama, A&M)
Ole Miss 8-4, 5-3 (Losses to Texas, Alabama, A&M, LSU)
LSU 6-6, 3-5 (wins over Auburn, Ole Miss, Arkansas - loss to TCU in opener)
Arkansas 3-9, 1-7 (win over Auburn - loss to ULL in opener)
Auburn 3-9, 0-8 (loss to Wazzu in opener)

Notes
* A&M was simulated with Manzeil returning. If you take him out, they lose to Ole Miss and LSU but still beat us.

[Goat Holder]

I see it the exact same way as the index.

[Coach34]

interesting- that was exactly as my predictions for us yesterday- but I have us as a toss-up with Ok State

[SignalToNoise]

Fascinating stuff, man. Thanks for sharing.

[curmudgeon]

LSU at 6-6 looked out of whack, but if you look at their schedule, even the eyeball test makes it a possibility

TCU - that's at best a slight lean for LSU if not a complete tossup.
UAB, Kent, Auburn - all pretty definite wins
at Georgia - that's a game where UGA will be favored
at MSU - a winnable game for State
Florida - another one where UF will be favored.

A 3-4/4-3 start isn't out of the question .. then

at Ole Miss - losable for LSU
Furman - definite win
at Alabama - definite loss
Texas A&M - losable for LSU
Arkansas - definite win

A 2-3 finish is very possible. 3-2 at best.

LSU's bowl hopes may ride on winning one of the games vs. Mississippi schools.

[CJDAWG85]

Wouldn't TSUN be 7-5 (4-4) if it has MSU winning the Egg Bowl?

[bulldogsmsu]

You forgot to put MSU as a lost to Ole Miss. Should be 8-5 4-4.


Football games aren't played on paper, or on Xbox, but a complex statistical formula can make for interesting conversation two weeks before kickoff.

First, some basics: 3.6 points is given to the home team, based on results of every game involving a BCS team the last three years. This is up slightly from last year, when the index gave a 3.4 point advantage.

Most of our scoring returns, and with QBs and WRs receiving equal credit for passing TDs in my formula, we have a high amount of returning scoring. Defensively, the index just looks at stats and returning starters. Losing two CBs hurts us because the index penalizes us against teams with big passing offenses, as you will read below.

Last year, the index had us at 8-4, with us beating Ole Miss and losing to Arkansas.

First some national predictions.

SEC Champion: Alabama (13-0)
Big 12 Champion: Oklahoma State (10-2)
Big Ten Champion: Ohio State (12-1)
Pac 12 Champion: Stanford (11-1)
ACC Champion: Clemson (11-2)
AAC Champion: Louisville (12-0)

National Championship: Alabama over Louisville

Our season:
Oklahoma State (Neutral Site). Loss 45-24 (0-1)
Alcorn State (FBS, not included in simulation), Win (1-1)
at Auburn. Win 31-21 (2-1, 1-0)
Troy. Win 38-17 (3-1)
LSU. Win 21-17 (4-1, 2-0)
Bowling Green. Win 42-24 (5-1)
Kentucky. Win 34-15 (6-1, 3-0)
at South Carolina. Loss 28-23 (6-2, 3-1)
at Texas A&M. Loss 38-14 (6-3, 3-2)
Alabama. Loss 27-10 (6-4, 3-3)
at Arkansas. Win 32-20 (7-4, 4-3)
Ole Miss. Win 24-17 (8-4, 5-3)

SEC West Standings
Alabama 13-0, 8-0
Texas A&M 11-1, 7-1 (Loss to Alabama)
MSU 8-4, 5-3 (Losses to OSU, USC, Alabama, A&M)
Ole Miss 8-4, 5-3 (Losses to Texas, Alabama, A&M, LSU)
LSU 6-6, 3-5 (wins over Auburn, Ole Miss, Arkansas - loss to TCU in opener)
Arkansas 3-9, 1-7 (win over Auburn - loss to ULL in opener)
Auburn 3-9, 0-8 (loss to Wazzu in opener)

Notes
* A&M was simulated with Manzeil returning. If you take him out, they lose to Ole Miss and LSU but still beat us.

[bulldogsmsu]

Wonder why so many people are so down on LSU. I have seen this a few times.

[Saltydog]

many starters and Miles is a bafoon........That's why..........They still have enough raw talent to overcome the experience, or lack thereof and to cover Les' shenanigans.......I think they win 9 games......

[Coach34]

Wonder why so many people are so down on LSU. I have seen this a few times.


because nobody ever loses that many Jr's at one time to the draft. They had to replace their entire DL
Now, with the Williford injury- they only have 2 starters back on the OL.

So, on both lines- only 2 of 9 starters returning. Thats alot to replace- teams like LSU or Bama are the only schools that could even think about replacing that many starters on their lines and not lose much. And that's not counting losses in other areas.

[maroonmania]

OK, I vote we call off the season and just use the index results. We will be extremely fortunate with this schedule to come out as well in reality. Also, not surprised the index has us losing to OSU but I am a little surprised it has us losing a 21 point beatdown. And believe me, if we lose by 21 to OSU, going 8-4 will be the furtherest thing from my mind.

[curmudgeon]

Coach nailed it.

LSU will struggle, and that gauntlet that they have middle of the season doesn't help.

[Coach34]

OK, I vote we call off the season and just use the index results. We will be extremely fortunate with this schedule to come out as well in reality. Also, not surprised the index has us losing to OSU but I am a little surprised it has us losing a 21 point beatdown.

As I was doing the breakdown for predictions- our schedule could be alot worse. I wouldnt want UPig's, Missouri's, or LSU's for anything

[Technetium]

I'd be really curious to know more about how you calculate this. Is this based on an iterative technique, such as MLEM, or some sort of regression method? Or are you simply assigning point values to various stats and adding/subtracting them for each game? I made a relatively simple regression model last year for creating my own power rankings, but soon gave up when I lacked the time to adequately compile the stats for my model on a weekly basis. Kudos to you for compiling stats.

[PassInterference]

Don't be silly, Tech. Its a complicated formula with a patent pending. State secrets don't get out.

The 615 index is really impressive. Solid results.

[maroonmania]

As I was doing the breakdown for predictions- our schedule could be alot worse. I wouldnt want UPig's, Missouri's, or LSU's for anything

Well, we play FIVE preseason Top 15 teams. Looking at the ones you mentioned Arkansas plays 5, Mizzou plays 4 and LSU plays 4 plus #20 TCU. Only one I probably wouldn't trade for is Arkansas who plays, TA&M/FL/SC/AL in 4 straight weekends which is nasty. I will say though that you stated you didn't think LSU would stay in the Top 15 and I sort of feel the same way about FL. I don't see them being a Top 15 team this year when all is said and done unless their offense has significantly improved.

[biscuit]

To me, that's about right. I hope it works out that way.

[Ronny]

..getting prison raped 45-24 by OSU yet still go on to have an 8 wins season is one of the most ridiculous things I've every read on a messageboard.

I'm sure various M.I.T. level calculations went into formulating your "index," but sometimes common sense is all it takes to see the big picture.

not this...



http://images.wikia.com/breakingbad/images/0/04/LabNotes4.jpg

[curmudgeon]

0 credit for a returning pass defense that wasn't great last year to begin with has us not faring well against teams with strong passing offenses.

Oklahoma State was 6th in the nation last year at 331/gm
Texas A&M was 11th in the nation at 316/gm

Remember how the A&M game went last year? And we still won 8 games. Can't really see how that is ridiculous.

[PassInterference]

How about giving up, what, almost 800 yards to aTm but still winning 8?

[Coach34]

Well, we play FIVE preseason Top 15 teams. Looking at the ones you mentioned Arkansas plays 5, Mizzou plays 4 and LSU plays 4 plus #20 TCU. Only one I probably wouldn't trade for is Arkansas who plays, TA&M/FL/SC/AL in 4 straight weekends which is nasty. I will say though that you stated you didn't think LSU would stay in the Top 15 and I sort of feel the same way about FL. I don't see them being a Top 15 team this year when all is said and done unless their offense has significantly improved.

I dont give a shit about current rankings

LSU plays one of the favorites to win the Big 12 at a neutral site
LSU plays Georgia and Fla from the East
LSU plays at Bammer, Georgia- the two SEC favorites...while also playing at both Mississippi schools- who will be solid teams in 2013.
LSU plays Georgia, State, Florida, and Ole Miss in 4 straight weeks

DO NOT WANT

UPig?

UPig's East games are Florida and SC

DO NOT WANT

Missouri???

They get Texas A&M and then travels to OM from the West- in addition to their East schedule
They play swing games Kentucky and Vandy on the road

Their schedule is comparable I guess

[Political Hack]

Hack index says UGA wins the SECCG and plays Ohio State in the NCG.

LSU will also be better than A&M. Met will struggle at times under his 4th OC in 4 years, but he will be good enough to not lose many games for them. Athletes win games.

A&M will also struggle with a new OC, new WRs, and loss of some absolute studs in the trenches.

[curmudgeon]

At least one thing always happens to throw off common sense.

In 2010, it was Cam Newton at Auburn. Turned a 7-5 team into an undefeated nation champion.
In 2011, it was the craziness that led to the LSU/Alabama national championship game.
In 2012, it was the emergence of a freshman QB at Texas A&M and a motorcycle ride gone awry.

This season, there are many candidates:

The rape allegations at Vandy will destroy a promising season coming off of 9 wins, which could lead to Tennessee making their way back to the top of the SEC East.

The rumors around Ole Miss could come to fruition and throw off a promising season for the Rebels.

If Manzeil doesn't play at A&M, that throws the entire SEC West into a frenzy below Alabama.

[archdog]

Hell Manziel is still running on us, the game was just called for pity, but Eulles was playing in the wrong spot so....

[curmudgeon]

Some other national surprises from the Index.

ACC
The tightest indexed game in 2013 is the first week. Georgia has a 0.003 advantage over Clemson and is predicted to win 28-27 in Death Valley.
Duke will be 7-0 heading into an October 26 game at Virginia Tech and is predicted to finish 9-3, and lose a tiebreaker to VT to play Clemson for ACC Championship.

AAC
Louisville is projected to win every game they play by double digits. A 10 point win at South Florida is their closest game.

Big 12
Best title race goes to the Big 12. Oklahoma State, Oklahoma, TCU, Texas, West Virginia and Kansas State will be within one game of the title. Index has:
OSU losing to TCU and Texas
OU losing to OSU and WVU
TCU losing to OU and K-State and WVU
Texas losing to OU, TCU and OSU
WVU losing to OSU, K-State and Texas
K-State losing to Texas, OSU, and OU

Baylor, Iowa State, Kansas and Texas Tech have a combined 0 wins over the top 6.

Big Ten
Michigan/Ohio State championship game the week after Ohio State beats Michigan in a battle of 11-0 teams.

Pac 12
Stanford is undefeated most of the season, but slips up on road at USC and against Notre Dame to end up at 10-2.

[Political Hack]

Hell Manziel is still running on us, the game was just called for pity, but Eulles was playing in the wrong spot so....

I laughed.

I think the 615 index is going to be pretty close with exception of the few issues I think we'll see. I've expected a dip from LSU the last few seasons but they have unreal athletes.

And of course I thought OM was going to be terrible last season and they were good enough to celebrate a 6 win season.

[PMDawg]

You forgot to put MSU as a lost to Ole Miss. Should be 8-5 4-4.

smh buddy, smh. haha

[PMDawg]

Football games aren't played on paper, or on Xbox, but a complex statistical formula can make for interesting conversation two weeks before kickoff.

First, some basics: 3.6 points is given to the home team, based on results of every game involving a BCS team the last three years. This is up slightly from last year, when the index gave a 3.4 point advantage.

Most of our scoring returns, and with QBs and WRs receiving equal credit for passing TDs in my formula, we have a high amount of returning scoring. Defensively, the index just looks at stats and returning starters. Losing two CBs hurts us because the index penalizes us against teams with big passing offenses, as you will read below.

Last year, the index had us at 8-4, with us beating Ole Miss and losing to Arkansas.

First some national predictions.

SEC Champion: Alabama (13-0)
Big 12 Champion: Oklahoma State (10-2)
Big Ten Champion: Ohio State (12-1)
Pac 12 Champion: Stanford (11-1)
ACC Champion: Clemson (11-2)
AAC Champion: Louisville (12-0)

National Championship: Alabama over Louisville

Our season:
Oklahoma State (Neutral Site). Loss 45-24 (0-1)
Alcorn State (FBS, not included in simulation), Win (1-1)
at Auburn. Win 31-21 (2-1, 1-0)
Troy. Win 38-17 (3-1)
LSU. Win 21-17 (4-1, 2-0)
Bowling Green. Win 42-24 (5-1)
Kentucky. Win 34-15 (6-1, 3-0)
at South Carolina. Loss 28-23 (6-2, 3-1)
at Texas A&M. Loss 38-14 (6-3, 3-2)
Alabama. Loss 27-10 (6-4, 3-3)
at Arkansas. Win 32-20 (7-4, 4-3)
Ole Miss. Win 24-17 (8-4, 5-3)

SEC West Standings
Alabama 13-0, 8-0
Texas A&M 11-1, 7-1 (Loss to Alabama)
MSU 8-4, 5-3 (Losses to OSU, USC, Alabama, A&M)
Ole Miss 8-4, 5-3 (Losses to Texas, Alabama, A&M, LSU)
LSU 6-6, 3-5 (wins over Auburn, Ole Miss, Arkansas - loss to TCU in opener)
Arkansas 3-9, 1-7 (win over Auburn - loss to ULL in opener)
Auburn 3-9, 0-8 (loss to Wazzu in opener)

Notes
* A&M was simulated with Manzeil returning. If you take him out, they lose to Ole Miss and LSU but still beat us.

how did your national predictions fare last year (conference champs, NC, etc.)?

[jumbo]

I would be very very happy if that's how the season played out

[curmudgeon]

Had Alabama beating Oregon last year.

3/6 in conference champions.

SEC: Alabama
Big 12: West Virginia
Pac 12: Oregon
ACC: Virginia Tech
Big 10: Nebraska
Big East: Louisville

WVU kind of bombed. Everyone else was in the hunt.

[Technetium]

Don't be silly, Tech. Its a complicated formula with a patent pending. State secrets don't get out.

The 615 index is really impressive. Solid results.

Not asking for specific values/weighting factors/etc. More interested in the algorithm used so I know how to interpret the results.

[Political Hack]

Not asking for specific values/weighting factors/etc. More interested in the algorithm used so I know how to interpret the results.

it's a COTS combination of a phonetic based and Slavic spelling algorithm... of wait, wrong subject.

[smootness]

Why are we not included as one of Ole Miss' losses?

[curmudgeon]

I may write a detailed description of how each specific game shapes up, but here's a little look inside. And admittedly, when looking at last year's numbers, its harder to be exact, but as the season progresses, it becomes more accurate.

Oklahoma State Passing vs. Mississippi State Passing Defense

OSU attempts 38.07 passes/game. They gain 8.71 yards/attempt.
The Index expects them to improve to 9.14 yards/attempt based on returning personnel.

MSU gives up 31.8 passes/game, and 6.96 yards/attempt.
Based on losing two starting cornerbacks, the Index expects them to give up 7.65 yards/attempt.

In a neutral site game, there is no advantage given, so the Index simply splits the difference. It expects OSU to gain 8.4 yards/attempt on us. (For comparison case, Texas A&M gained 8.5/attempt on us.)

In this game, OSU is expected to throw at us 39.97 times. Gaining 8.4 yards/attempt, that has them gaining 336 yards passing on us. Last year, they averaged 331.69 yards/game.

In gaining 331.69 yards per game, they scored 45.67 points. When you throw in expected rushing totals, it evens out to 43.62 expected points vs. MSU, that rounds up to 45 since 44 is not a likely football number.

[FISHDAWG]

that's some analytic stuff right there and reason for concern but I just don't think there is any way they score over 35-38 points

[LiterallyPolice]

I may write a detailed description of how each specific game shapes up, but here's a little look inside. And admittedly, when looking at last year's numbers, its harder to be exact, but as the season progresses, it becomes more accurate.

Oklahoma State Passing vs. Mississippi State Passing Defense

OSU attempts 38.07 passes/game. They gain 8.71 yards/attempt.
The Index expects them to improve to 9.14 yards/attempt based on returning personnel.

MSU gives up 31.8 passes/game, and 6.96 yards/attempt.
Based on losing two starting cornerbacks, the Index expects them to give up 7.65 yards/attempt.

In a neutral site game, there is no advantage given, so the Index simply splits the difference. It expects OSU to gain 8.4 yards/attempt on us. (For comparison case, Texas A&M gained 8.5/attempt on us.)

In this game, OSU is expected to throw at us 39.97 times. Gaining 8.4 yards/attempt, that has them gaining 336 yards passing on us. Last year, they averaged 331.69 yards/game.

In gaining 331.69 yards per game, they scored 45.67 points. When you throw in expected rushing totals, it evens out to 43.62 expected points vs. MSU, that rounds up to 45 since 44 is not a likely football number.

Is there a last-year-strength-of-schedule component? In other words, we gave up 6.96 yrds / pass attempt last year... but that may have been against above-average passing teams. So against an average team, we may only give up close to 6 yds / attempt (For example).

Does that make sense?

[curmudgeon]

There is. But its based on stats instead of names.

For example, Ole Miss has a more effective pass offense than Alabama. Therefore, our performance against Ole Miss is more relevant to our performance against Oklahoma State than Alabama because Bama runs it down your throat and has an NFL defense.

[curmudgeon]

that's some analytic stuff right there and reason for concern but I just don't think there is any way they score over 35-38 points

And that's fine. I'm certainly not saying its exact on the scores, but it is interesting to look at. A 45-24 prediction is still decently accurate if the game ends up 38-31. It underestimates our offense and our defense by one score.

[FISHDAWG]

we need you as the boards official recruiting dept head

[curmudgeon]

we need you as the boards official recruiting dept head


Negative. I hate three things.

Prosperity Gospel, Pretentiousness and College Football Recruiting

Joel Osteen in Oxford, Mississippi on the first Wednesday in February is what hell looks like to me.

[PMDawg]

I may write a detailed description of how each specific game shapes up, but here's a little look inside. And admittedly, when looking at last year's numbers, its harder to be exact, but as the season progresses, it becomes more accurate.

Oklahoma State Passing vs. Mississippi State Passing Defense

OSU attempts 38.07 passes/game. They gain 8.71 yards/attempt.
The Index expects them to improve to 9.14 yards/attempt based on returning personnel.

MSU gives up 31.8 passes/game, and 6.96 yards/attempt.
Based on losing two starting cornerbacks, the Index expects them to give up 7.65 yards/attempt.

In a neutral site game, there is no advantage given, so the Index simply splits the difference. It expects OSU to gain 8.4 yards/attempt on us. (For comparison case, Texas A&M gained 8.5/attempt on us.)

In this game, OSU is expected to throw at us 39.97 times. Gaining 8.4 yards/attempt, that has them gaining 336 yards passing on us. Last year, they averaged 331.69 yards/game.

In gaining 331.69 yards per game, they scored 45.67 points. When you throw in expected rushing totals, it evens out to 43.62 expected points vs. MSU, that rounds up to 45 since 44 is not a likely football number.

This is awesome, but I do see a potential problem. Auburn this year will look nothing like auburn last year. How do you account for something like that? Arkansas I understand is a different beast. They should suck out loud this year so using last years data seems acceptable. Auburn is an odd situation.

[curmudgeon]

You can't. It's not perfect, but after three games, it starts to look more like current teams rather than last year.