Mike Martz

[TroyF]

Last night I picked up the "Pro Football Prospectus"

It's a pretty good read if you're a stat geek. Reminds me of some of the old Bill James books for baseball. It has articles for each team and analyzes statistics in some new ways.

Flipping through the book, I came across an interesting article.

"Research: In Defense of Mike Martz"

As someone who has never really liked Martz a lot, this was a surprising read. For instance, the top ten coaches winning% when having any lead at any time in the fourth quarter - 1970-2004 (minimum 20 4th Q leads):

#1 on the list? Mike Martz. 91.4% of the time.

They also talk about the Pythagorean projection. Just as in baseball, this projection shows how many games a team should have won based on its overall statistics.

The Rams are the only NFL team to outperform their projected win total in each of the last five seasons. Martz has the highest average number of wins over expected wins of any coach currently in the league. He also has the lowest standard deviation, meaning he's the most consistent at getting more wins.

They do take into account that maybe the Rams style plays something into that, but note that Vermeil won fewer games than projected all three years with the Rams and three of the past four with the Chiefs.

I'm not sure it'll sway a lot of people. For me, the numbers really don't lie. I think I'll start giving him a little more credit.

The book has an $18.95 list price and is a very good read for those interested.

[ice4277]

Interesting stuff. Out of curiosity, which team most underperformed compared to their projected wins?

[TroyF]

Interesting stuff. Out of curiosity, which team most underperformed compared to their projected wins?


I'm not sure. They don't have a single table that shows all of the teams. In the Martz article they only listed the best coaches projections. (Martz, Gibbs, Green, Sherman, Haslett)

It does have the last five years of each team shown in each teams 5 year performence chart. It looks like the last two years show the Bucs with the worst margain. (keep in mind this is a cursory glance)

Last year the pythagorean formula showed they should have 7.9 wins. They had 5. The year before that the split was 9.2 and 7.

Personally, I still think the style of play can have something to do with it. But I was stunned at the fourth quarter lead statistic. I can't see any reason for that happening.

[Noop]

Where have you been Troy?

Since it looks like we have the college football group(you know those of us who really enjoy the sport) maybe we should start a discussion.

[clintl]

I'm not sure that the fourth quarter stat is that stunning. When the Rams have had leads, they often have had big leads. If the stat was broken down more to account for that, you might have something. Say, leads of less than 7 or less than 10 points.

[TroyF]

I'm not sure that the fourth quarter stat is that stunning. When the Rams have had leads, they often have had big leads. If the stat was broken down more to account for that, you might have something. Say, leads of less than 7 or less than 10 points.


I don't buy that. We are talking about a statistic covering 35 years and well over 100 coaches.

He hasn't had two score leads in the fourth quarter 90+% of the time.

Just a quick glance at last year shows the Rams were 5-1 in games they led in the fourth quarter by 10 or fewer points. (the loss was a game against the Saints where they trailed, briefly took a three point lead and then lost in OT)

The statistical varience there just isn't that great. Simply put, if the Rams are leading in the fourth under Martz, they are more likely to win the game than ANY team in football over a 35 year period.

We can try and make excuses for that all we want, but it seems pretty silly to do so IMO.

[clintl]

That's not what I'm saying. I'm saying that not all 4Q leads are equal, and that the Rams likely have proportionally more big 4Q leads than a team that wins primarily with defense is likely to have because of its style of play. I'm not saying that Martz doesn't have the best or one of the best records with 4Q lead in a close game, just that I don't think considering all games with a 4Q lead proves it.

The Pythagorean stat is a more compelling one.