Study: Troy Polamalu worth 11.2 points per game

[cdcox]

An analysis at the link below shows that the Steelers were 11.2 points per game better when Polumalu was playing compared to the games where a replacement player filled his position.

http://www.advancednflstats.com/2009...plusminus.html

There are some limitations to the study. But let's say he was only worth 6 points per game. That is still eye-popping.

[Mecca]

Originally Posted by mylittlepony

"He's also a classy and selfless player who never showboats"

I stopped reading at this point.

Wait how is that not true?

There is nothing wrong with Polamalu at all he's not an asshole.

[tk13]

I'm all for taking Berry because he has a chance to be one of the great players taken from the draft, and there aren't any pass rushers of that caliber out there. Everybody's fallen in love with safeties... it's really important, but I'd still put it #2 behind pass rush. We should take Berry because he's the best player, not because we "have" to have a safety.

All of those recent Super Bowl teams had safeties, but pass rush is still more important. It's what allowed the Giants to control the best offensive team in recent history two years ago. I think having that Freeney/Harrison/Ware type pass rusher is the #1 priority.

Yeah, a guy like Dawkins has helped the Broncos, someone to make people pay for their mistakes. But I think the bigger factor is Dumervil's become one of the league's elite pass rushers.

[mylittlepony]

Originally Posted by Mecca

Wait how is that not true?

There is nothing wrong with Polamalu at all he's not an asshole.

He is probably my least favourite trojan ever after that big hit on Aaron Lockett.

[88TG88]

Originally Posted by mylittlepony

He is probably my least favourite trojan ever after that big hit on Aaron Lockett.

A safety hitting a receiver really hard. Thats unheard of and uncalled for.

[Mecca]

Originally Posted by mylittlepony

He is probably my least favourite trojan ever after that big hit on Aaron Lockett.

You guys are STILL hung up on that?

[RealSNR]

Originally Posted by tk13

I'm all for taking Berry because he has a chance to be one of the great players taken from the draft, and there aren't any pass rushers of that caliber out there. Everybody's fallen in love with safeties... it's really important, but I'd still put it #2 behind pass rush. We should take Berry because he's the best player, not because we "have" to have a safety.

All of those recent Super Bowl teams had safeties, but pass rush is still more important. It's what allowed the Giants to control the best offensive team in recent history two years ago. I think having that Freeney/Harrison/Ware type pass rusher is the #1 priority.

Yeah, a guy like Dawkins has helped the Broncos, someone to make people pay for their mistakes. But I think the bigger factor is Dumervil's become one of the league's elite pass rushers.

I don't see any pass rushers in the top half of this draft.

Berry it is.

[Mecca]

The pass rushers are 4-3 ends and our brilliant GM ruined that.

[tyler360]

It was a bad hit but i think the game being in the big 12 prolly has something to do with it. Ya he made a bad choice but its not like he does it all the time i have not seen that many bad hits by the guy. Every sports outlet also reports how awesome he is so I dont really see what the issue is.

[jAZ]

Originally Posted by SNR

It would be interesting to see some statistical analysis of the 2004 Colts with and without Bob Sanders. The result was pretty much the same.

And people say Eric Berry and other safeties are worthless

Well, 2004 was his rookie year. He didn't start much that season. But looking at his full career, most noteably, he's been an inury machine. He's only had 2 seasons where he started double digit games and has been out more than he's been in 47 of 96+ games.

But looking at games he started vs games he didn't:

Started: 16.8 points
Didn't: 17.3 points

Less than 1/2 a point.

By season:

2009
Started: 11.5
Didn't: 14.2

2008
Started: 18.4
Didn't: 19.2

2007
Started: 17.3
Didn't: 14.0

2006
Started: 18.8
Didn't: 22.9

2005
Started: 15.1
Didn't: 20.5

2004
Started: 21.5
Didn't: 21.7

[Third Eye]

The "analysis" is flawed to say the very least. Statistically insignificant is probably a better descriptor.

[jAZ]

Originally Posted by Third Eye

The "analysis" is flawed to say the very least. Statistically insignificant is probably a better descriptor.

Yeah, Bob Sanders has missed just over 1/2 of his 96+ games, so we have some valid data to work with. I didn't look at the TP data, but if you were to cherry pick one season (2005) where BS's absense was dramatic (5 points), it's largely because he only missed 2 games that year. BS had his best seasons in 2005 and 2007 when he was healthy for all but 1 and all but 2 games respectively. Of course in 2005 he was "worth" 5 points and in 2007, he actually "cost" them 3 points.

[Coogs]

Whitout pouring through the statistics, it appears Peyton Manning is worth about 24-27 points per game for the Colts.

[cdcox]

jAZ -- The analysis that Brian Burke does is significantly more sophisticated than the way you've looked at it. I've asked him to run the numbers of Bob Sanders. I agree that it is a good case to consider.

Third Eye -- WPA is calculated based on play-by-play data. So you have hundreds of raw data points for with and without TP. The difference between with and without TP is large. I think it is a mistake to say that the study has no statistical significance.

[jAZ]

Originally Posted by cdcox

jAZ -- The analysis that Brian Burke does is significantly more sophisticated than the way you've looked at it. I've asked him to run the numbers of Bob Sanders. I agree that it is a good case to consider.

I admit I didn't read the article, and after reading it, I think it's pretty cool analysis, but going back and doing the same thing for TP that I did for BS.

TP, 2009
Started: 13.8
Didn't: 23.1

That's your 10 point difference right there. Pretty straight forward.

[Third Eye]

Originally Posted by cdcox

jAZ -- The analysis that Brian Burke does is significantly more sophisticated than the way you've looked at it. I've asked him to run the numbers of Bob Sanders. I agree that it is a good case to consider.

Third Eye -- WPA is calculated based on play-by-play data. So you have hundreds of raw data points for with and without TP. The difference between with and without TP is large. I think it is a mistake to say that the study has no statistical significance.

There are simply too many omitted variables IMO. I honestly wouldn't have the slightest clue how to go about testing his system, but I'm fairly confident that if it were transformed into a linear regression or something similar the results wouldn't be conclusive. I could do it, I've got the software, but I don't feel like spending my day off getting it together.