Penn St Recap: Pitchfork not included

Submitted by The Mathlete on

We are getting worse and that nice September we had isn’t as nice as it used to look.

We’ll ease in to the pain and start with the offense.

The Bad

image

After three of the first four FBS games produced stratospheric results, real Big Ten play, even after being opponent adjusted as it is here, the offensive performance has come back to earth. This is not the same as being bad, this is still a very good offense. In fact, the +5 average over the last three games would still be good enough to rank in the top 25, it’s just not good enough to bail our defense out like it did early.

After three great games through the air in his first four, Denard has regressed as a passer. His last three games have been slightly above +0 after averaging nearly +10 against UConn and Indiana.

Michigan is still ranked #2 nationally in rushing and #3 overall, in Points Above Normal but the game scores are coming down.

The Bad and The Ugly

image

This chart is not upside, it just shows no need to display positive numbers. Last year’s defense was bad, this year’s is terrible. Last year’s defense didn’t make it down to –2 on the season, this year is worse than –7. We are 113th after adjusting for opponents and the only BCS conference school worse is Washington St.

The “good” news is that next week shouldn’t be too bad. The defense has alternated between worst game of the season and slightly below average for every week this year.

If the season were to end today, Michigan’s gap between offensive success and defensive failure would be the largest I have recorded in 8 full seasons on file.

Projections

Not nearly as fun as they used to be.

Illinois: 34% chance of victory

@Purdue: 65% chance of victory

Wisconsin: 38% chance of victory

@Ohio St: 12% chance of victory

Home field is factored in. The season win projection is now at 6.5.

image

Ballot

Still trying to turn my system, which is more like a power ranking into something more similar but definitely not identical to a traditional poll setup. For now I have settled on ranking teams first by number of losses and secondarily by power rating and then making one off adjustments to account for teams that are rated lower but have beaten a team with a identical losses.

Rank Team PAN
1 Auburn 22.8
2 Boise St 18.6
3 TCU 17.0
4 Oregon 14.4
5 Utah 10.9
6 Nebraska 23.0
7 Alabama 15.9
8 Arkansas 19.2
9 Oklahoma St 18.4
10 Stanford 18.1
11 Missouri 17.7
12 Oklahoma 17.8
13 Wisconsin 11.6
14 Ohio St 14.3
15 Michigan St 14.1
16 LSU 13.4
17 Arizona 10.5
18 Nevada 5.0
19 Iowa 18.7
20 S Carolina 16.9
21 Virginia Tech 16.3
22 Hawaii 14.4
23 Mississippi St 12.8
24 NC St 9.2
25 Baylor 7.2

Teams of note:

  • Oregon: my numbers still aren’t sold on their defense and strength of schedule.
  • Nebraska: might have the most dangerous offense/defense combination in football
  • Alabama: the numbers don’t see last year and think there are a lot of better 1 loss teams than the Tide
  • Wisconsin: hurt by three lackluster non-conference wins against underachieving opponents, propped up by their win against OSU

Comments

joeyb

November 1st, 2010 at 4:16 PM ^

I love that your numbers always reflect my mood and feeling about the team. Illinois and Wisconsin look winnable if we play a perfect game and only because we are at home. Purdue is horrific. If we don't beat them, I think he can kiss any support he has left goodbye. OSU just isn't a winnable game this year. It never has been and even if we win the next 3, that game still won't look winnable.

If we do end up beating Illinois, it would be nice to see the prediction jump to 7.8 or something like that because Wisconsin starts to look like a better matchup too.

joeyb

November 1st, 2010 at 4:44 PM ^

Well, they both lost to MSU and Illinois had to go to OSU while Wisconsin got to play them at home. You also consider that they didn't get blown out by Missouri and the numbers are probably going to lean in favor of Illinois.

UMaD

November 1st, 2010 at 5:20 PM ^

We are bad at D.  We were always bad at D...Unless you mean Mike Martin getting hurt, which sucks, but you have to expect injuries.  I think other teams might be improving their offenses as the season progresses, while ours (due to injury and position changes) is still trying to get off the starting block.

The early season offense seemed too good to be true, and it was.

To your system's credit, the early season numbers, which looked ridiculous (80% chance of beating Wisc??), are now coming around and looking far more reasonable.  6.5 was my preseason expectation.  This seems, intuitvely, spot on.

Interesting that your system likes our chances against Wisc at least as much as Illinois.

Yinka Double Dare

November 1st, 2010 at 5:39 PM ^

If Nebraska's receivers hadn't had a bizarre case of the drops against Texas,  I think they'd be in the top 2 right now in the polls.  It wasn't as bizarre as their insane game against Iowa State last year, but it wasn't that far off.

UMinTroyOh

November 1st, 2010 at 8:28 PM ^

It would be interesting to see a stat breakdown of the offensive performance after a punt vs after a kick-off return. There is the yardage issue which seems immense as the KO returns are inept  at best (at least fewer fumbles). Does the offense operate looser after a returened punt? The pressure on the offense seems immense as they need to score every time they touch the ball.

jackw8542

November 1st, 2010 at 11:19 PM ^

It is interesting that the only game of the last three where the O scores high is the one game played primarily by Tate Forcier.  I would still rather have a passing QB who can run than a running QB who can pass, and for the last 3 games, Denard has looked like a running QB who can sort of pass, particularly in the first half.

jackw8542

November 1st, 2010 at 11:19 PM ^

It is interesting that the only game of the last three where the O scores high is the one game played primarily by Tate Forcier.  I would still rather have a passing QB who can run than a running QB who can pass, and for the last 3 games, Denard has looked like a running QB who can sort of pass, particularly in the first half.

jackw8542

November 1st, 2010 at 11:20 PM ^

It is interesting that the only game of the last three where the offense showed a strong score is the one game played primarily by Tate Forcier.  I would still rather have a passing QB who can run than a running QB who can pass, and for the last 3 games, Denard has looked like a running QB who can sort of pass, particularly in the first half.

nazooq

November 2nd, 2010 at 4:59 PM ^

Mathlete: can you construct a PAN ranking just based on conference games?  I'd like to see how Michigan compares to other Big Ten teams when non-conference opponents are filtered out.