49ers-Rams Statistical Preview: Singletary vs. Haslett

This week, the 49ers travel to the Central time zone to play the Rams. Although some have brought up the fact that it's yet another 10:00 a.m. EST start, the Niners haven't had as much difficulty in the Central generally, and in St. Louis specifically, than they have in the Eastern time zone. Thus, there's no reason to raise game location as a relevant issue for Sunday's game. Instead, I'm going to focus this week's statistical preview on how the teams' respective game performances have changed (if at all) since their interim head coaches were hired.

The most basic (but not very valid) way to go about this is to look at the team records. According to this metric, and in agreement with the conventional wisdom, Mike Singletary's 3-4 record is light years better than Jim Haslett's 2-8 record. It's even more of a disparity if we ignore the 49ers-Seahawks game, for which Singletary only had 3 days to prepare. In comparison, Haslett was hired at the beginning of the Rams' bye week, so he had 2 full weeks to prepare for his first game (a 19-17 win @ WAS).

So Singletary has a better winning percentage. For some, this is game over. However, for others like me who believe that team efficiency is a better predictor of long-term success, the game has only just begun. Below are the relevant stats showing how efficiently the 49ers and Rams have played under their interim head coaches as compared to their now-fired head coaches:

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Nolan	-14.5%	-19.1%	-1.9%	2.7%
Singletary	-22.5%	-11.8%	17.6%	7.0%
DIFFERENCE	-7.9%	7.3%	19.5%	4.3%

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Linehan	-69.2%	-25.1%	42.5%	-1.6%
Haslett	-41.4%	-24.7%	17.5%	0.8%
DIFFERENCE	27.8%	0.4%	-25.0%	2.4%

Based on these two tables, the conventional wisdom appears wrong on three counts. First, the 49ers have actually performed worse under Singletary. Second, Haslett's impact on Ram performance has been much greater than Singletary's impact on 49er performance (35.7% more positive impact). Finally, while Singletary and Haslett are both defensive-minded coaches, the Rams' defense has improved considerably, whereas the Niners' defense has gotten much worse. So what gives? This must be some kind of statistical magic wherein I can make the numbers say anything I want them to, right?

After the jump, I'll tell you whether Haslett over Singletary is reality, magic, or somewhere in between...kind of like David Blaine. I'll also detail the lonely SVW for this week's matchup, and serve up my game prediction...

OK, so is this statistical magic? Yes...and no. First, the "yes" part...

MAGIC

There are definitely two statistical phenomena at play in these DVOA comparisons: floor effect and regression to the mean. In stats, a floor effect occurs when the value for a variable is already so low that intervening events are increasingly unlikely to make it any meaningfully lower. It can also be thought of as a kind of law of diminishing returns. The improvement of a golfer's average score over time is one sports example of a floor effect. Getting from 100 to 90 is pretty easy with practice (and no cheating), getting from 90 to 80 is a little harder, getting from 80 to 70 is really hard, and getting from 70 to 60 is nearly impossible. Regression to the mean is related to a floor effect insofar as it's what is almost certain to follow if the ultra-low value is artificially ultra-low.

With these two concepts in mind, you realize that Haslett's Rams had to improve considerably given that he took over a team on pace to becoming the worst since 1995 by over 10%. Also, notice that Haslett took over a team that was 54.7% worse than the team Singletary took over!!! Clearly then, the Rams' "improvement" under Haslett has been, in part, statistically artificial.

Next, the "no" part...

REALITY

In my last 49ers-Rams preview, and to the dissent of many, I presented stats that showed Frank Gore to be a much more valuable running back to the 49ers than Steven Jackson is to the Rams. A corollary to the "Steven Jackson is not that valuable" argument applies to Jim Haslett's impact as well. Specifically, one might argue that, although the Rams' improvement is the result of some statistical chicanery, perhaps it's actually even more of an improvement (i.e., a real one) if you consider the fact that Linehan had Jackson in the backfield for all 4 of his games as head coach, while Haslett has only had Jackson for 6 of his 10 games as head coach. Well, here's the breakdown showing the Rams' game DVOAs under Haslett with Steven Jackson (SJ) and without him (NSJ), and comparing Linehan/Jackson to Haslett/Jackson:

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Haslett-NSJ	-77.9%	-43.4%	28.5%	-6.0%
Haslett-SJ	-17.1%	-12.3%	10.1%	5.4%
DIFFERENCE	60.8%	31.1%	-18.3%	11.4%

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Linehan	-69.2%	-25.1%	42.5%	-1.6%
Haslett-SJ	-17.1%	-12.3%	10.1%	5.4%
DIFFERENCE	52.1%	12.8%	-32.3%	6.9%

From these two tables, we find that the improvement under Haslett may be reality after all when you consider the huge effect Steven Jackson seems to have on the Rams' game DVOAs. Specifically, the Rams are playing about 60% more efficiently in games with Steven Jackson than in those without him. Also, the presence of Jackson does not seem to only impact the Rams' offense. Indeed, all three team units have played better when Jackson is in uniform. Perhaps this is why some believe he is invaluable to the Rams' despite his unimpressive DYAR. It makes sense given that DYAR only measures a running back's value in the running game.

Getting back to Haslett, his 27.8% overall impact on game DVOA nearly doubles when you ignore games in which Steven Jackson did not play. Statistically speaking, a 52.1% change is not regression to the mean. Rather, it is a progression to meaningful. Therefore, it looks like Haslett really has had a positive impact on the Rams. Whether or not Singletary's game DVOAs have been affected by factors similar to "Steven Jackson OUT" is more difficult to measure because there are far more moving parts (e.g., starting lineup changes, Frank Gore injury, lame Letterman bits, etc.).

Finally, the "somewhere in between" part...

DAVID BLAINE

So far it seems pretty clear that Jim Haslett has had a positive impact on the Rams, much more so than the harder-to-measure impact Singletary has had on the 49ers. All is not lost, loyal Niner fans. There is one stat that gives a hint about why Singletary's 49ers appear to be more improved than Haslett's Rams. It's called consistency (or lack thereof). Below are three final tables (poker pun intended) comparing the standard deviations (SD) of performance for the two teams under their interim head coaches, as well as to their previous coaches, after adjusting Haslett's stats to reflect Steven Jackson's availability:

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Nolan	28.7%	21.4%	17.0%	7.3%
Singletary	19.1%	17.2%	9.2%	15.2%
DIFFERENCE	-9.5%	-4.2%	-7.8%	8.0%

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Linehan	18.4%	15.1%	16.7%	9.7%
Haslett-SJ	47.8%	29.3%	24.3%	7.5%
DIFFERENCE	29.3%	14.3%	7.6%	-2.3%

Head Coach	TOT DVOA	OFF DVOA	DEF DVOA	ST DVOA
Singletary	19.1%	17.2%	9.2%	15.2%
Haslett-SJ	47.8%	29.3%	24.3%	7.5%
DIFFERENCE	28.6%	12.2%	15.0%	-7.8%

In statistics, the standard deviation (SD) is a measure of variability. It tells you how different a set of numbers are from the average value. When applied to change statistics, it tells you how inconsistent a set of numbers is over time for the same individual/group. In the case of Singletary vs. Haslett, the SDs above measure how inconsistent the 49ers and Rams have been under their interim head coaches, and how inconsistent they were under their previous head coaches. Because consistency is better than inconsistency when it comes to performance, lower SDs are what we're looking for.

With that knowledge in hand, we can see from the tables that, as compared to their previous head coaches, the 49ers have become more consistent under Singletary, whereas the Rams have become less consistent under Haslett. In addition, the Niners have been over twice as consistent under Singletary as the Rams have been under Haslett. These stats confirm what our eyes have seen for the past 6 games, and are an affirmation of Singletary's quest to establish a "49er identity" through his "Formula for Success." The main feature of an identity, whether we're talking about people or NFL teams, is consistency over time. Likewise, the hallmark of a team that executes well (Formula Variable 2) is consistency from play to play.

Bottom line: Both interim head coaches have been successful at improving their teams, they've just addressed two different problems. Jim Haslett inherited a team that was engaging in epic fail performances, and got them to perform better. Mike Singletary inherited a team that wasn't performing that badly, but had no identity or ability to execute, and got them to perform more consistently. Haslett's task was about competing, whereas Singletary's task was about winning. That's why their records are so different.

49ERS-RAMS DVOA MATCHUP

Here's how the 49ers stack up against the Rams in terms of DVOA:

Team	Total DVOA	Rank	Offense	Rank	Defense	Rank	Special Teams	Rank
SF	-19.0%	26	-15.2%	28	8.4%	18	4.6%	2
STL	-49.0%	32	-23.8%	32	25.3%	31	0.1%	18

Regardless of the fact that the game is on the road, this is definitely a game the Niners should win. If they can't beat the no-doubt worst team in the league, a team that should be more concerned with offseason tee times than beating the 49ers, then I'll take back everything I just said in the preceding Singletary vs. Haslett analysis. Luckily, the 2008 Niners have shown a propensity to wipe the floor with really bad teams (See games vs. DET and vs. STL). Among the three team units, the 49ers have clear advantages on special teams (+4.5%) and when their defense is on the field (+15.4%).

THE DEVIL IS IN THE DETAILS

The one SVW to focus on during this week's 49ers-Rams matchup is detailed below:

Situation	SF Rank	STL Rank
SF DEFENSE VS. STL OFFENSE ON 3RD AND MID	3	26

Not very juicy stuff here, but there's a clear reason for it. Namely, neither of these teams are top 8 in much of anything. Of the 50 statistical matchups I go through each week, I could identify just this one as an SVW. That's compared to the 15 weakness vs. weakness matchups I found. Two situations where both of these teams are seriously lacking just happen to be two of the most important when it comes to winning: red zone and late/close. Luckily, the Rams are epically failing, whereas the 49ers are simply failing: The Rams' are 27^th or worse in overall red zone offense (32^nd) and defense (29^th), red zone passing offense (32^nd) and defense (27^th), red zone rushing offense (31^st) and defense (30^th), second half offense (32^nd) and defense (29^th), and late/close offense (32^nd) and defense (29^th). Wow, if that's not a useful way to end up with the #1 pick in the draft, I don't know what is.

"WITH HIGH HOPE FOR THE FUTURE, NO PREDICTION IS VENTURED." - ABRAHAM LINCOLN

Last week, I picked the 49ers on general principle. The stats predicted a 12-point 49er loss (seems like my stats are always a touchdown off), but I picked a 2-point 49er win. With that said, I'd do it again. I hate the Dolphins that much. This week, it's back to the stats.

San Francisco 49ers	27
St. Louis Rams	19

**DVOA statistics used to produce this article were obtained from Football Outsiders.