Comparison of Michigan QBs over 50 years

Comparison of Michigan QBs over 50 years

Submitted by UMProud on November 13th, 2018 at 8:57 AM

Derrick Hutchinson authored a great article contrasting key statistics of Michigan quarterbacks over the last 50 years.  He included a great sortable table of quarterback statistics on a year by year basis.

One key metric, completion percentages, listed Shea Patterson as #1 (YTD obviously).  And #2?  Jim Harbaugh's 1986 performance.

12 game stats - 2016 vs. 2015

12 game stats - 2016 vs. 2015

Submitted by ST3 on November 28th, 2016 at 4:06 PM

I'd like to move on from the Columbus Screwjob, circa 2016, and focus on the improvement in the team from 2015 to 2016. (Yes, this is a coping mechanism, but we all have our own ways of dealing with the pain.) The biggest question going forward is how much the improvement is due to the system being implemented (year 2 of the Harbaughffense, importing Don Brown) and how much is due to the players just getting 1 year older. We did have a lot of seniors on the roster this year. Anyway, links to the stats are here:

2015 season http://www.mgoblue.com/sports/m-footbl/stats/112815aac.html

2016 season http://www.mgoblue.com/sports/m-footbl/stats/112616aac.html

Offense

* The first thing we see is a dramatic increase in points scored (yay.) We went from 367 (30.6 ppg) to 492 (41.0 ppg.) A TD and a FG better per game in 2016.

* First downs increased from 237 to 267. But the mix tilted more heavily to the run in 2016 as the Harbaughffense got closer to full implementation. We went from 85 rushing first downs in '15 to 138 in '16, an increase of 53 first downs.

* Rushing yardage increased from 1832 to 2679 yards and the per carry average increased from a measly 4.1 to a respectable 5.0. I know the criticism will be that we couldn't grind out first downs against Iowa and OSU when we needed them, but only the truly elite teams can impose their will on solid defenses and run the ball effectively when everyone knows that's what they are trying to do.

* Passing yardage decreased somewhat. I would attribute that to us playing well ahead of the opposition for large stretches this year, and the lack of deep balls to Chesson. He and Butt both saw their per game averages go down this season. We will miss Darboh, but there are opportunities for the young WRs to improve upon Chesson's production this year, and a gaggle of more experienced tight ends should mitigate the loss of Butt.

* Regarding QB play, the Total efficiency went from 133.75 to 143.23, a nearly 10 point improvement. Considering we were going with a first time starter this year versus an experienced, 5th year QB last season, that speaks positively to Jim's QB development skills. Can we get another 10 point increase from Speight next season?

Defense

*Oh man, Don Brown. The first thing one checks is sacks. Those increased from 30 to 44. But the major, eye-popping stat is the QBHs. We had 16 in 2015. This year? 52! Was our statistician more generous this year with QBHs? Our opposition had 13 last year to 25 this year, so maybe, but I think Wilton had a tendency to hold onto the ball longer than Rudock did and saw more pressure as a result.

* TFLs are less subjective than QBHs. They went from 82 to 115.

* But all those fancy stats don't mean anything if they don't show up on the scoreboard. Average points allowed dropped from 17.2 to 12.5, a 4.7 point per game reduction. (Hey, the offense showed more improvement than the defense did, at least using raw PPG as a metric.)

There's lot of other interesting stuff in there, so I recommend you take a look. In this time of infinite sadness at the missed opportunities and horrible Ohioness of the reffing, take solace in the fact that even though we only improved the regular season win total by 1 game, the team was dramatically improved over last year's squad in just about every metric. What will year 3 of the Harbaughffense bring? What will year 2 of Don Brown's defense bring? The Knowledge knows, but I don't. I do think the future is bright. If I know one thing about Harbaugh, it's that the result on Saturday is going to put steel in his spine, and he's already a pretty tough S.O.B.

Go Blue.

 

OT: Statistical model predicts with high accuracy the play-calling tendency of NFL teams

OT: Statistical model predicts with high accuracy the play-calling tendency of NFL teams

Submitted by Don on August 12th, 2015 at 11:33 AM

I wonder if college teams are as predictable as the NFL. Seems like the people in Vegas might find this interesting info too.

http://phys.org/news/2015-08-statistical-high-accuracy-play-calling-ten…

 

 

Statistical Review Of Week 1 In The Big Ten - Averages And Extremes

Statistical Review Of Week 1 In The Big Ten - Averages And Extremes

Submitted by LSAClassOf2000 on September 3rd, 2013 at 10:20 AM

I thought that it might be appropriate to give a brief statistical summary of the ups and downs of the Big Ten in its inaugural weekend for 2013, and there were some interesting extremes.  Granted, the averages for 12 games aren’t terribly meaningful, but they are interesting to look at all the same.

THE CONFERENCE ON OFFENSE:

The average performance in the Big Ten on offense was a fairly balanced attack – 214 passing yards and 228 rushing yards over an average of 71 plays. That’s good for 6.23 yards per offensive snap.

The most prolific rushing attack in the first week belonged to Wisconsin, which put 393 yards rushing up against Massachusetts, followed closely by Nebraska’s 375 yards rushing against Wyoming. The more paltry rushing numbers belong to Penn State and Purdue at 57 and 65 yards respectively.  Teams carried it an average of 42 times and did so at a clip of 4.8 YPC as well for about 2 rushing TDs.

Through the air, the 228 yards came about as the result of a typical performance being 17-28 with 13.4 yards per completion and 8.1 yards per attempt being the conference average. Further, the Big Ten averaged 2 interceptions and 2 passing TDs in its games. The average completion percentage of a Big Ten team this past weekend was 61.43%. Illinois gained the most yards in the air – 415 on a conference-best 29 completions (37 attempts). The least successful passing attack belonged to Minnesota – 10-23 for 99 yards.

Big Ten teams scored an average of 39.5 points, including a typical performance of 2 FGs and 5 TDs. Indiana scored the most points far and away with 73 of them on Thursday, and then at the other end of the spectrum, there was Purdue and their 7 against Cincinnati.

There was a wide variation when it came to the number of first downs earned as well. Against the conference average of 21, Indiana managed 29 first downs against Indiana State, and at the other end, Purdue managed 12 first downs in their game. On third down, the average performance was 7-15 – Michigan converted the most with 10, but Wisconsin had the highest percentage at 72.7% (8-11).

THE CONFERENCE ON DEFENSE:

On the other side of the ball, the Big Ten gave up an average of 357 yards, broken down into about 240 yards passing and 117 yards rushing in a typical game. They defended an average of 73 plays, yielding 4.8 yards per play.

Although this could be blamed on a host of things, Michigan State yielded only 11 yards rushing to Western Michigan on 27 attempts, or an average of 0.4 YPC, far and away the best performance statistically over the weekend. Purdue owns the worst performance in rush defense, yielding 221 yards to Cincinnati across 47 attempts, or 4.7 YPC. It is important to note, however, that UNLV was getting nearly 6 yards per carry against Minnesota. The conference average was 3.4 YPC on defense in the first week.

Wisconsin owns the best overall performance in pass defense this weekend, giving up only 112 yards in the air, followed by Michigan at 144 yards. By a considerable margin, Northwestern’s  455 yards given up in the air to Cal rates as the worst performance on pass defense, although Purdue, Illinois and Nebraska were giving up longer passes on average.

The conference allowed an average of 18 first downs to teams, with the most being the 35 that Nebraska handed to Wyoming, and the fewest being  11, or the number allowed by Penn State against Syracuse. Third downs were difficult to convert if you played the Big Ten in general – the average success rate was 5-16, but Northwestern allowed Cal to convert 10 times, and Nebraska allowed Wyoming to convert just once. That being said, Wyoming didn’t need to get to third down very much.

THE CONFERENCE WHEN KICKING / PUNTING:

Big Ten teams average 4 punts for 177 yards over the weekend, or a shade over 40 yards per punt. Michigan State did an awful lot of punting – 11 of them for 423 yards, whereas Michigan and Wisconsin only have one each. Interestingly, Michigan State turned it around to have one of the best punt return performances, returning 5 of them for 52 yards, but Indiana returned 4 punts for 90 yards.

When it comes to kickoff coverage, Indiana and Michigan got quite a bit of practice in, being in this situation 12 and 10 times respectively. Indiana managed 777 yards of it for an average of 64.8 yards per kick with 6 touchbacks. That’s against the typical performance of 8 for 479 yards and 3 touchbacks. As for kick returns, Minnesota and Illinois both took one to the house and racked up 148 and 135 yards of return yardage respectively. Despite scoring the most points, Indiana owns Week 1’s least impressive performance on returns – 3 for 31 yards.

CONCLUSION:

As the programmers at Infocom would say, "The point of the game is to discover the point of the game."

OBLIGATORY:

 

APR And Big Ten Football: A High-Level Summary

APR And Big Ten Football: A High-Level Summary

Submitted by LSAClassOf2000 on June 18th, 2013 at 1:50 PM

APR AND BIG TEN FOOTBALL: A HIGH LEVEL SURVEY    

SOME NOTES FIRST…

A few threads lately have touched on the subject of the Academic Progress Rate (APR) and where Michigan has been at in the recent past, but I thought it would be even more interesting to take a look at the entire Big Ten over the last several years. As it might garner the most interest, I chose to compare football programs.

I will first say that there was an interesting quandary that presented itself in collecting this data. It is simple enough to look up the rolling averages for the past eight years, but the reports published by the NCAA only had the individual team APR for four years prior, so I had to recreate the formula for finding the individual APRs using the rolling averages and I went back as far as 2004-05. I double checked my results and they seem reasonable.

ACADEMIC PROGRESS RATE:

Here is what we’re measuring when we talk about the Academic Progress Rate of a team.

For a given team, each student receiving aid will receive one point for retention (staying in school) and one point for remaining eligible to play. So, for a football team in Division I that is fielding the full complement of 85 scholarship athletes, there are a possible 170 points.  If you have in a given year, for example, four players who drop out and are ineligible (subtract 8 points), and two players who remain but are merely ineligible (subtract 2 points), you would have (160 / 170)*1000 or an APR of 941.

It is also important to note that, when we enter the new championship structure, teams must earn a 930 four-year minimum average or a 940 for last two seasons to be eligible to participate in the championships. In 2015-16, it will simply be the 930 rolling average as the benchmark for participation. So, if you look at this from the perspective of how many “points” do you lose to get to 930, 93% of 170 is 158.1, so say, 159 or 11 points.

Of course, it is a metric, and the manner in which teams keep players in school and eligible can always be debated. It isn’t perfect by any means, but it is an interesting measure as it stands.

ROLLING AVERAGES:

First, here is the table with the rolling averages (thumbnail is due to size of original):

 photo APRRollingAvg_zpsb122ffea.png

There’s not a lot to say other than the general trend is towards improvement for almost everyone. As it is a rolling average, it does hide some intriguing variations between individual years, but you can see that the conference as a whole is generally getting better.

INDIVIDUAL APRs:

In the table below, you’ll see the individual team APRs, some of which were found algebraically as I mentioned.

 photo APR957Table_zps58d58414.png

 

The average of the individual APRs for football for the conference (including Nebraska when appropriate) is 957, but I have shaded in this table the instance of APRs below 957 so you can see which teams have missed that mark and how often.  Northwestern, Indiana, Ohio State, Penn State and Wisconsin would come out as APR winners in this analysis if there were such a prize.

Here’s another way to look at it, however. This table shows team performance (by way of cool shading) the performance of individual teams against the yearly conference average of those individual APRs. The far right column is the conference average, and the bottom row is the school average in that period.

 photo APRAvgTable_zps409f67e7.png

 

The one that should immediately grab attention is Minnesota, of course, followed to a lesser extent by Purdue, Illinois and Michigan State. These schools seemed spend a majority of this period at or below the conference average for individual APR in a given year. Michigan had a bad stretch there but you can see the tremendous improvement in recent years. Northwestern should not shock anyone really. Ohio State does well in this analysis as well.

ANOTHER VIEW OF IT:

Here is a cleaner view of individual team performance versus the average:

 photo APRIndiana_zpscdc792b8.png  photo APRMichState_zps591afd48.png  photo APRNorthwestern_zps6a772800.png  photo APRPennState_zps0ad25e13.png  photo APRPurdue_zpsbb3f358e.png  photo APRIllinois_zps228805fe.png  photo APRIowa_zpsb590d891.png  photo APRMichigan_zps07e2b189.png  photo APRMinnesota_zpsf059484a.png  photo APRNebraska_zps139e3361.png  photo APRWisconsin_zpsfc57e0e3.png

TL;DR CONCLUSION:

Again, part of the analysis was actually trying to extract information through algebraic means, so if I did all that right and I am not just deluding myself with regards to my math skills, you should now have a somewhat clear view of where the Big Ten has been and where it is headed when it comes to the measure.  Whatever you may think of it as a tool, there has been a net increase of 5.06% in the Big Ten’s average yearly score over these last eight years. When you think of how many more student-athletes that may very well mean are completing their education, the effort inside the Big Ten to drive achievement is yielding results.

OH, AND OF COURSE...

Using Rivals' Star Ratings To Look At Big Ten Football Recruiting: 2002-2013

Using Rivals' Star Ratings To Look At Big Ten Football Recruiting: 2002-2013

Submitted by LSAClassOf2000 on May 20th, 2013 at 9:29 AM

USING RIVALS’ STAR RATINGS TO LOOK AT BIG TEN RECRUITING: 2002-2013 CLASSES

I decided to take a look at Rivals.com and the star ratings that they give recruits to come up with an approximation of relative recruiting success over twelve classes (2002-2013) in the Big Ten. To see where Nebraska may have fit in (we can’t really know how different it would be if they’d been in the conference for the whole period studied), I included them as well.

So, including Nebraska, Rivals had ratings on 3,160 recruits from the period 2002 to 2013. Here is what the relative distribution looks like:

 photo RatingDIstribution_zps1692db27.png

So, as you will note, 5-stars and to a lesser extent 4-stars are something of a rare commodity in the Big Ten historically, accounting for only about 1/5thof all recruits by their data. One thing that is interesting, and you can see this in more detail later, is just how far above the conference norm Michigan and Ohio State tend to sit.

 

STAR RATING

MICHIGAN

OHIO ST.

EVERYONE ELSE

MICHIGAN / OHIO ST. %

FIVE

13

17

20

60.00%

FOUR

125

130

344

42.57%

THREE

116

96

1370

13.40%

TWO

11

11

907

2.37%

 

The one thing that should jump out here is the percentage of five and four-star recruits that go to the traditional “Big Two”, if you will. In fact, about 44% of all players ranked four and higher end up at either Michigan or Ohio State. You can also see how sparse those same rosters tend to be when it comes to two-star recruits – by Rivals’ system, Big Ten teams not in Ann Arbor or Columbus attract nearly 98% of the two-star talent.

For all 3,160 players whose ratings I dumped from Rivals’ database, the grand mean star rating is 2.93, but individual teams obviously have had varied success. Here are the team means for the period from 2002 to 2013.

 photo AvgRatingBar_zps17a6c302.png

Including Nebraska as a point of some comparison, they would have been quite competitive in Big Ten recruiting circles regardless, so it seems. Here, they are in possession of the third-highest average. In fact, four teams have managed to recruit at or above a “three-star” in those twelve classes. Five teams have managed to stay above the grand mean of 2.93 – Ohio State, Michigan, Nebraska (had they been part of the Big Ten for the whole period), Penn State and Michigan State. Michigan State’s numbers spike with the Dantonio years, as you might expect, or otherwise this doesn’t happen, I think.

Here is the conference mean for each year as well:

 

YEAR

CONF. AVERAGE

2002

2.84

2003

2.88

2004

2.70

2005

2.91

2006

2.89

2007

2.99

2008

2.89

2009

3.01

2010

2.98

2011

3.01

2012

3.03

2013

3.05

 

So, the trend is actually upwards, ever so slightly. The net increase in the average star rating over this period is about 7%, which doesn’t seem like much. It would be interesting to compare this to other conferences in a further diary. I suspect some of this – as it is subjective – has to do with perhaps a steady-state perception of the Big Ten.  I really don’t know – that’s speculation on my part.

A more meaningful comparison between each team and the conference mean for a given year will appear shortly, but this now allows us to look at another intriguing phenomenon.

 

TEAM

TEAM AVG. (2002-2013)

CLASSES ABOVE CONF. AVG

CLASSES BELOW CONF. AVG

CLASSES ABOVE TEAM AVG

CLASSES BELOW TEAM AVG

OHO STATE

3.60

12

0

7

5

MICHIGAN

3.54

12

0

8

4

NEBRASKA

3.20

12

0

7

5

PENN STATE

3.16

9

3

4

8

MICHIGAN STATE

2.95

7

5

6

6

WISCONSIN

2.86

3

9

7

5

IOWA

2.79

1

11

6

6

ILLINOIS

2.79

2

10

5

7

PURDUE

2.69

2

10

6

6

MINNESOTA

2.65

2

10

7

5

NORTHWESTERN

2.57

0

12

6

6

INDIANA

2.42

0

12

5

7

 

So, here again we see the relative inequities in where the talent tends to go in the Big Ten, with two schools having never experienced a year above the conference mean rating, one school having managed this feat only once and three schools having achieved this only twice. Of course, on-the-field success is a different story from year to year with some of these teams too, but it seems to illustrate that several teams in the conference do indeed get along with less, if these ratings are any indication. Obviously, there is overlooked or underrated talent, so it is ultimately subjective and not 100% accurate by any means.

So, how did each team fare against the conference mean each year? Here’s what that looks like for each team:

 

 photo MichiganAvgRating_zps2d3d9c1e.png  photo OhioStateAvgRating_zps9cdefacb.png  photo MichiganStateAvgRating_zps2ffe370f.png  photo NorthwesternAvgRating_zps3739196a.png  photo PennStateAvgRating_zpse2e8c6fd.png  photo IowaAvgRating_zps54f06287.png  photo MinnesotaAvgRating_zps7bb25385.png  photo IllinoisAvgRating_zps1c391a29.png  photo IndianaAvgRating_zps2dd94c9e.png  photo PurdueAvgRating_zps673dc7ff.png  photo WisconsinAvgRating_zpsc114df79.png  photo NebraskaAvgRating_zpsc9f66085.png

 

Here’s the whole conference on a rather non-descriptive but somewhat telling chart. You can see Michigan and Ohio State flying comfortably above the rest of the conference for the most part:

 photo TheWholeConference_zps62245149.png

Here are some comparisons with select teams. Why Indiana? I was inspired somehow:

 photo MichOhioStateComparison_zps2413fa24.png  photo MichMichiganStateComparison_zps18ed1295.png  photo MichIndianaComparison_zpsb6f5ad4a.png

TL;DR CONCLUSION:

Like many of these diaries that I do, the driver is for the reader to draw their own conclusion about what they see. Rivals’ data was the easiest to categorize, which is why I used it here, but they aren’t the only ranking service, nor are stars the sole measure of who is in fact the better overall player. What was intriguing to me is how these ratings make the Big Ten appear when you dig into them a little, and the trends seem reasonably accurate to me.

OBLIGATORY:

A Few Historic Trends: 25 Years Of Wolverine Basketball

A Few Historic Trends: 25 Years Of Wolverine Basketball

Submitted by LSAClassOf2000 on February 4th, 2013 at 1:29 PM

HISTORIC TRENDS: 25 YEARS OF WOLVERINE BASKETBALL

This week’s diary is mainly designed to graph out some of the trends in box score statistics (and a few derived from the box score) that we have seen over the last 25 seasons of Michigan basketball. What I hoped to achieve here was merely to provide a visual for the increases and decreases in these statistics over time, but also to look at the averages in this period  to get some idea of how well the current team is performing compared to the last, well, several teams. It is important to note that all totals for the current team are to date – I didn’t want to do a projection of future totals simply because of the potential error that would be part of that projection.

TABLE 1 – COMPARISON OF CERTAIN AVERAGES

 

 

1988-2012

Current Team

FG%

45.80%

50.50%

3PT%

35.10%

40.20%

FT%

70.80%

70.50%

Rebounds / Game

35.6

36.4

Assists /Game

14.2

15.4

Points Per Game

71.9

77.8

Offensive Eff.

1.008

1.181

Defensive Eff.

0.993

0.915

Off. Rebound %

31.80%

31.40%

Def. Rebound %

68.80%

74.60%

It should be noted (hence the italics) that the data for efficiency was only available for the past 16 seasons, so bear this in mind when looking at the above figures. That being said, you can see from this quick comparison that the current team is performing markedly above the combined average of the 24 teams which came before it, and it is arguably on track to fall under the historic average for statistics such as turnovers and personal fouls, and about on average for blocks and steals.

The charts below are admittedly exaggerated via some varied scaling in order to show clearly some of the cycles and trends in some of the statistics. I will apologize in advance for this, but it seemed like a good way to show this.

For example, you can see average rebounds per game on a trend of general decline for years, bottoming out over the last few seasons before a definite recovery in this season:  photo 25YearsAvgRebounds_zps08e4f959.jpg

You can see with average assists per game several cycles of varied length, but another generally steady downward turn from the 90s through the last few seasons:

 photo 25YearsAvgAssists_zpsd186b53b.jpg

In average points per game, there was a very protracted trend of general decline (with noted spikes) that seemed to get reasonably flat at time wore on. Again, this year, the trend is reversing.

 photo 25YearsAvgPoints_zps04ad32e2.jpg

Looking at offensive and defensive efficiency, you will see that, most years, we hung out somewhere around 1.000 for offensive efficiency, but have had widely varied performance on defense. A similar tale can be told for offensive and defensive rebound %, but here, it is defensive rebounding % which is relatively stable in comparison to its counterpart. This is less obvious when looking at rebound totals for the season, of course.

 photo 25YearsEfficiency_zps0c5bd9bc.jpg  photo 25YearsReboundPct_zpsc0a0aec9.jpg  photo 25YearsAvgRebounds_zps08e4f959.jpg

On this chart, you’ll find field goal percentage, three-point percentage and free throw percentage trends so you can see the interplay between the three. Free throw shooting and field goals seemed to actually be relatively stable over time, with our three-point percentage being by far the most variable aspect of our shooting offense.

 photo 25YearsShooting_zps7de2ac91.jpg

Here are the trends for turnovers, blocks, personal fouls and steals.

 photo 25YearsTurnovers_zps916ba150.jpg  photo 25YearsBlocks_zps45071a91.jpg  photo 25YearsPersonalFouls_zps173e0bee.jpg  photo 25YearsSteals_zpsa91ad158.jpg

TL;DR CONCLUSION:

I am not sure there is a fixed conclusion here only because the current season is not yet finished, but no it is even clear in the numbers that, at least in a few important regards, Michigan basketball is coming back to levels of production and play that it has not seen in quite some time. We can see it on the court, but the statistics definitely do back up the so-called “eye test” comparisons we tend to make of this team.

OBLIGATORY:

An Attempt At A Big Ten Hoops Scorecard: (Approximately) Midseason Review

An Attempt At A Big Ten Hoops Scorecard: (Approximately) Midseason Review

Submitted by LSAClassOf2000 on January 22nd, 2013 at 11:50 AM

“ATTEMPT AT A BIG TEN HOOPS SCORECARD”

Over this past weekend, I began to frame out the hoops version of the football scorecard that I had advanced last month on the board. Indeed, I have already made some enhancements to that one thanks to the input of board members. This one, however, is indeed another rough attempt at correlating various productivity measures into a relative measure of success.

I took eighteen typical boxscore statistics and gathered the season-to-date totals for each team in the conference. When the season is complete, it might be intriguing to do a “conference vs. overall” sort of analysis with this, but for now, I am presenting merely the overall card to this point. There were a few things that I admittedly waffled on leaving here, but did for purposes of discussion. Most notably, the actual totals of FGs, 3PTs, and FTs both made and attempted are analyzed along with success rate. My thought here was – among other things - that it indicates potentially where teams could be doing a lot of work for very little return and could show inefficiency.

Obviously, it is not by any means an exact science, and although I show that the results are rather highly correlated to actual win percentage, they obviously do not take into account the whole set of variables active in a game, and indeed, in a season, such as strength of schedule. This is merely an attempt to apply a similar scoring approach to basketball as I had done with football. It is fairly similar to the football scorecard in that there are certain statistics where being above the conference mean is decidedly better, and a few where actually being below the mean is preferred.

The larger table is linked below – there is no way to compress the overall scorecard to make it fit easily into the diary setup on the blog, so the link should take you to the relevant images in my Photobucket account.

OVERALL CARD (THUMBNAIL):

 photo BigTenScorecardMidway_zps9603e073.jpg

SUMMARY CARD:

 photo BigTenScorecardMidwayCOrrel_zps8659b80c.jpg

DISCUSSION:

The idea of "grading" teams on these 18 metrics seems to bear itself out fairly well if we correlate it to current win percentage. One of the noted anomolies here is Ohio State, which falls into the "red" zone, but yet has a fairly decent win percentage. This might be an example of one of the thing which mere productivity does not catch, which is a team with erratic shooting success that plays decent defense, enough so to win a majority of the time. If anyone has some suggestions on ironing such things out, I would love to hear them as I want to make these as accurate as possible.

Comments and suggestions are welcome as always. My intent here is to develop tools which can be used perhaps by the whole community here to gauge relative performance among our fellow conference members.

FOR THOSE WHO MERELY WANT THE CAT PHOTO:

Week #11 Statistics and FEI Prediction for Iowa

Week #11 Statistics and FEI Prediction for Iowa

Submitted by Enjoy Life on November 14th, 2012 at 1:14 PM

Prediction for Iowa: The FEI Forecast for this Saturday is Michigan 28 – Iowa 16 with a 78% Probable Win Expectation for Michigan. Michigan's offense continues to be excellent (4.81 PPPo) against poor teams (AFA, UMass, Purdue, Illinois, Minnesota, Northwestern) but has struggled (0.90 PPPo) against every good team (Alabama, ND, MSU, Nebraska). Iowa is ranked #88 in OFEI and #30 in DFEI. Looks like FEI has it just about right this week.

Strength of Schedule: Michigan's SoS for Out of Conference games is much harder than the B1G games. This is quite unusual and because of the OOC Strength of Schedule, M is actually doing better in B1G games versus OOC for both offense (2.7 vs. 2.5 PPPo) and defense (1.3 vs. 2.0 PPPo). The defense had their worst game (2.8 PPPo) since Alabama.

imageFremeau Efficiency Index: Michigan improved in both overall and offense FEI with defense basically unchanged. In the detailed chart below, GE represents the raw data for FEI before adjustments for opponents.

The S&P Ratings (Also from Football Outsiders) is a play based analysis (rather than possession based) and M is ranked #19 overall, #18 in offense, and #27 in defense. The S&P ratings DO include games against non-FBS opponents (go figure).

The FEI is a drive based analysis considering each of the nearly 20,000 drives each year in FBS vs. FBS college football. The data is filtered to eliminate garbage time (at the half or end of game) and is adjusted for opponent. A team is rewarded for playing well against good teams (win or lose) and is punished more severely for playing poorly against bad teams than it is rewarded for playing well against bad teams.

imageNational Rankings: The rankings for offense and defense are based on scoring (yardage statistics are inherently flawed). These are simply raw numbers without any adjustments for opponent, garbage time, or anything else. The data is from TeamRankings and includes only games between two FBS teams.

FEI Details: Here are the FEI numbers for Michigan and their opponent ( Football Outsiders FEI ).

imageimage_thumb15_thumb1_thumb_thumb

imageimagePoints Per Possession: Cumulative PPPo is 2.6 for the offense and 1.6 for the defense. M finished 2011 outscoring opponents by almost a 2:1 margin with PPPo for offense of 2.8 and defense of 1.4. The 2 charts show the raw data for offense and defense with the number of possessions adjusted for "kneel downs" at the half or end-of-game (maximum deduction = 2).

Using Scoring Offense and Scoring Defense National Rankings for the past 5 years (FBS AQ teams only), this table shows the percentage of teams that finish the season with a +WLM and a +5 WLM. For example, teams that finished in the Top 40 in both offense and defense had a 100% chance to be +WLM and an 82% chance to be +5 WLM (9-4 or better).

image_thumb34_thumb

Week #10 Statistics and FEI Prediction for Northwestern

Week #10 Statistics and FEI Prediction for Northwestern

Submitted by Enjoy Life on November 6th, 2012 at 11:47 AM

Prediction for Northwestern: The FEI Forecast for this Saturday is Northwestern 22 – Michigan 21 with a 52% Probable Win Expectation for Northwestern. Another toss up game? NSFMF! FEI is wrong yet again and M wins this one 31 —13. M has the #3 ranked Strength of Schedule and NW is #69 in SoS. In National statistics, M & NW are ranked about equal in scoring offense (#50 & #53) but M is ranked #13 in scoring defense with 16.8 PPG and NW is ranked #43 allowing 24.2 PPG.

Michigan's offense continues to be excellent (4.81 PPPo) against poor teams (AFA, UMass, Purdue, Illinois, Minnesota) but has struggled (0.90 PPPo) against every good team (Alabama, ND, MSU, Nebraska). Northwestern would be classified as a poor defensive team.

imageFremeau Efficiency Index: Not much movement in the FEI. In the detailed chart below, GE represents the raw data for FEI before adjustments for opponents. M is ranked #28 in GE and overall FEI is #34. This seems about right since M has lost 3 of the 4 games to their highly ranked opponents.

The S&P Ratings (Also from Football Outsiders) is a play based analysis (rather than possession based) and M is ranked #20 overall, #19 in offense, and #20 in defense. The S&P ratings DO include games against non-FBS opponents (go figure).

The FEI is a drive based analysis considering each of the nearly 20,000 drives each year in FBS vs. FBS college football. The data is filtered to eliminate garbage time (at the half or end of game) and is adjusted for opponent. A team is rewarded for playing well against good teams (win or lose) and is punished more severely for playing poorly against bad teams than it is rewarded for playing well against bad teams.

imageNational Rankings: The rankings for offense and defense are based on scoring (yardage statistics are inherently flawed). These are simply raw numbers without any adjustments for opponent, garbage time, or anything else. The data is from TeamRankings and includes only games between two FBS teams.

FEI Details: Here are the FEI numbers for Michigan and their opponent ( Football Outsiders FEI ).

imageimage_thumb15_thumb1_thumb

imageimagePoints Per Possession: Cumulative PPPo is 2.5 for the offense and 1.5 for the defense. M finished 2011 outscoring opponents by almost a 2:1 margin with PPPo for offense of 2.8 and defense of 1.4. The 2 charts show the raw data for offense and defense with the number of possessions adjusted for "kneel downs" at the half or end-of-game (maximum deduction = 2).

Using Scoring Offense and Scoring Defense National Rankings for the past 5 years (FBS AQ teams only), this table shows the percentage of teams that finish the season with a +WLM and a +5 WLM. For example, teams that finished in the Top 40 in both offense and defense had a 100% chance to be +WLM and an 82% chance to be +5 WLM (9-4 or better).

image_thumb34_thumb