Writers of Pro Football Prospectus 2008

27 Sep 2008

Tomlin Should Avoid Sophomore Slump

This week's Monday Night Football column explores the possibility that successful first-year coaches undergo a sophomore slump, and why such a slump might exist.

Posted by: Bill Barnwell on 27 Sep 2008

11 comments, Last at 30 Sep 2008, 9:28am by Arkaein

Comments

1
by Thunderbolt of ... :: Sat, 09/27/2008 - 9:32pm

I don't like the FO writing on mainstream sites nearly as much. Rather than the usual in-depth analysis that I had come to love FO for, the articles tend to be a bunch of fluff based around one small statistical piece.

6
by Ben Stuplisberger (not verified) :: Sun, 09/28/2008 - 12:24pm

I had been critical of the ESPN writing as well, but this article is better. It follows the "keep it simple" rule, which allows a fully realized argument in the space allotted by ESPN.

2
by Brian (not verified) :: Sat, 09/27/2008 - 9:36pm

I think there are more natural reasons for a sophomore slump, if one really does exist.

1. Coaches are usually fired and replaced after an unusually poor year for a team.
2. New coaches come into situations with fresh high draft picks and (usually) last place schedules.
3. Top players who were injured in the previous down year tend to return to health.
4. Those factors, combined with Guassian regression tendencies--including the fact that better division opponents will regress themselves, usually leads to a bounce-back.
4. Those conditions usually do not persist, but expectations are raised, and therefore 2nd year coaches appear to have slumps.

9
by ammek :: Mon, 09/29/2008 - 7:41am

1. Coaches are usually fired and replaced after an unusually poor year for a team.

Depends on the team. The Mike Tomlin example doesn't fit.

2. New coaches come into situations with fresh high draft picks and (usually) last place schedules.

Fresh high draft picks regularly contribute next to nothing in their first season. What value did Calvin Johnson, Reggie Bush or D'Brickashaw provide in Year One? Can you attribute the Bears' extraordinary 2001 campaign to the selection of David Terrell?

3. Top players who were injured in the previous down year tend to return to health.

And top players who were healthy in the down year may get hurt.

4. Those factors, combined with Guassian regression tendencies--including the fact that better division opponents will regress themselves, usually leads to a bounce-back.

Except if you're the Lions, Cards, Bengals, Raiders, Texans, Falcons, Niners....

4. 5? Those conditions usually do not persist, but expectations are raised, and therefore 2nd year coaches appear to have slumps.

Why would those conditions appear particularly in a coach's first year? Why are they not just completely random? And what about the exceptions? Why were the 2001 Bears or 1998 Falcons one-year wonders, while the 2000 Colts bounced back and the 2002 Pats begat a dynasty?

I think you under-emphasize the importance of opponents having film of a coach's system and studying how his players play in it.

I would be interested to see the correlation in first- and second-year improvement for teams hiring new general managers, especially when there's an enormous clearout (eg Parcells in Miami, Loomis in New Orleans, Thompson in Green Bay).

3
by DJ Any Reason (not verified) :: Sat, 09/27/2008 - 11:29pm

Ummm, Bill? While your namesake may have coined the phrase "plexiglass principle," its maybe a teensy bit of a stretch for crediting Bill James with inventing regression to the mean.

4
by Scott C :: Sun, 09/28/2008 - 12:47am

That comment has several 18th and 19th century mathematicians and physicists rolling in their graves.

5
by Jerry :: Sun, 09/28/2008 - 2:03am

While the question that the article asks is legitimate, even if the answer is just regression to the mean, the setup struck me as contrived. While, as with all teams, some fans want to see changes after any loss, the Pittsburgh fan base as a whole isn't particularly impatient, and the questions I've heard about Tomlin's second year have had much more to do with personnel (specifically the offensive line) than with any "sophomore jinx".

One other question the article brought to mind: Does the "Plexiglass Principle" disappear below a five-win swing, or is it just smaller than the 2.7 wins cited?

8
by Mystyc :: Mon, 09/29/2008 - 12:39am

Yes, I can definitely say that sophomore status has little or nothing to do with any questions being pointed Tomlin's way. They are more based on (as mentioned) the continuing offensive line issues and relative lack of in-game adjustments when a particular unit is being overwhelmed (see the O-line last week, or the D-line against JAX in the playoffs last year).

That said, he's also improved in other areas. He isn't throwing away challenges on a wing and a prayer anymore -- I think last week was his first actual won challenge. After making special teams an area of emphasis in training camp last year, they were terrible, but this year, he went hands-off and instead infused the coverage teams with actual talented players, and there's been a drastic improvement. If he can somehow whip up the same turnaround for the offensive line next year, he will be rightfully praised.

7
by Bill Barnwell :: Sun, 09/28/2008 - 12:43pm

I certainly wasn't implying that Bill James invented regression to the mean.

10
by Doug (not verified) :: Mon, 09/29/2008 - 10:28pm

Well Bill, you're getting a slightly better reception than this article did:

http://www.footballoutsiders.com/stat-analysis/bears-trend

Yet you're really talking about the same thing, you just used the existing term "Plexiglass Principle" to describe regression to the mean.

And I have to admit, it's amusing to see how accepted the idea of regression to the mean has become on this site only a few years after anyone mentioning it was ridiculed. Things just take time, I guess.

11
by Arkaein :: Tue, 09/30/2008 - 9:28am

No, the plexiglass principle is a variation of the standard regression to the mean. It says that the regression is greater following a large jump.

So if I understand it right, a 3-13 team that improves to 9-7 might actually be expected to regress somewhat towards 5 or 6 wins, which is below the NFL mean, though closer to that team's own recent historical mean.