Dewan on value of defense relative to offense
by rooster on Mar 13, 2009 2:08 PM CDT
Email
PrintThe most significant discovery of my careerI'm not sure where the Rangers rank in 2008, so consider two scenarios: ranked near the Royals, ranked in the middle.
March 12, 2009
About two weeks ago The Fielding Bible—Volume II went to print. Since then, as I've been studying some of the data in the book preparing for interviews, I came upon a discovery that was truly amazing to me. The most amazing, and significant, discovery of my 25 years in the baseball analysis business.
The key mission of the second volume of The Fielding Bible was to translate all of our new defensive methods into one common number that would be understandable by everyone. That number is Defensive Runs Saved. How many runs does a player save for his team defensively?
We look at each player individually. We then do a team summary by adding up all the individual players. How many runs does an above-average defense save compared to an average team? The team with the best defense in baseball in 2008 was the World Champion Philadelphia Phillies. By combining all of our defensive methods, eight different methods across the nine positions in baseball, we estimate that the Phillies defense saved 78 runs. Using the rule of thumb that 10 runs is equivalent to one win, that's eight wins. With an average defense, the Phillies wouldn't have had even a sniff of the playoffs.
The worst defensive team in baseball in 2008? The Kansas City Royals. Their defense cost them about 48 runs relative to the average team. Comparing the Phillies and the Royals, the difference between the best and worst defensive teams in baseball was about 130 runs.
Now, remember that number. 130.
The best run-scoring team in baseball was the Texas Rangers with 901 runs in 2008. The San Diego Padres were the worst with 637 runs. That's a difference of about 260 runs.
Here's the discovery, and I found it because the numbers just jumped out. The 130 difference in runs saved on defense is exactly half of the 260 difference in runs scored. That's exactly half. The implication is that defense is worth about half as much as offense.
That's a lot higher than I would have guessed, and a lot higher than I think most people would guess. But the numbers are remarkably consistent from one year to the next:
Year Best to Worst Offensive Difference Best to Worst Defensive Difference Defensive Spread as Percentage of Offensive Spread 2008 264 126 48% 2007 295 141 48% 2006 241 114 47%
Everyone realizes that defense is important, but it's never been quantified. Now we have the first way to quantify it. It's not necessarily the best way, and there will be more to come on this issue. The 50% figure is more of an indicator than an exact number, but it just jumped out at me and I wanted to share it with you.
If the Rangers in 2008 were near the top rather than near the Royals in defense runs saved, the 2008 run differential would have been something like +64 (901 scored minus 837 allowed) rather than -66 (901 scored minus 967 allowed). IOW, a pythagorean record of 87 wins and 75 losses
If the Rangers were near the top rather than middle, the 2008 differential would've been +9 runs (901 scored minus 892 allowed) for a pythagorean record of 82 wins and 80 losses.
Basically, the Rangers would've had to have had the best defense in the majors in 2008 to have had a winning record.
That's not exactly how Dewan intended that discussion to be used, as he appears to be focused on the relative value of scoring versus preventing runs rather than how it is applied to individual teams. In that vein, his numbers suggest that the top scoring teams have more to lose by replacing mashers with slick fielding slappers than poor defending teams have to gain by the same approach. If a team is poor at both runs scored and runs saved (ahem - KC), it better invest in a player's offense first.
I think the moves the Rangers have made on defense will be productive, but I don't think that those moves alone make it substantially more likely the Rangers finish as a .500 club, since it is hard to believe they will rise to the best run saved team in the league. The good thing is that the Rangers haven't really replaced poor-fielding mashers with top-fielding slappers, which appears to be an approach that would overemphasize defense.
BTW, if anyone has the Fielding Bible 2007 and 2008, I would really like to see the change in runs saved by the Rays. They allowed 273 fewer runs in 2008 compared to 2007. Dewan's observation suggests that at most about 140 of the 273 could be attributed to improved defense. Their team FIP went from 4.70 in 2007 to 4.22 in 2008, or a decrease of 80-85 runs. It doesn't all quite add up when thinking about pitching and defense independently. There may be some synergistic relationship as the Ray's team ERA went from much above the FIP ERA in 2007 to much below in 2008 (5.53 versus 4.72 in 2007, 3.82 versus 4.22 in 2008).
19 comments
|
6 recs |
Do you like this story?
Comments
This is fascinating.
Sure, its only a corollary supposition at this point, but the numbers do seem to back up his hypothesis.
Again, fascinating.
P.S. How could I sign up for his newsletter, rooster?
"...my balls are really like a veiny flesh color" blueballlefty on Jun 4, 2008 7:44 PM EDT
"you gonna lose your horse. seriously." FX2
Yes we can! November 04, 2008
Send an email to 'info@actasports.com' and make the first line of the email body 'opt in'.
JD: Adamant about 78 wins in 2009. Go Rangers!
Where did your 140 come from?
The Rays allowed 273 fewer runs. OK. The 50% number that Dewan puts out there is the relationship of the value of offense to defense. It doesn’t imply anything about why the rays allowed 273 fewer runs. It could the improvement in pitching in 2008 that accounts for all the difference (though I’m sure that’s not the case).
by NorCalRangersFan on Mar 13, 2009 3:13 PM CDT reply actions
Here, I guess:
They allowed 273 fewer runs in 2008 compared to 2007. Dewan’s observation suggests that at most about 140 of the 273 could be attributed to improved defense.
That would presume a rate of about 51%, which may be a scoch high, but close enough.
"...my balls are really like a veiny flesh color" blueballlefty on Jun 4, 2008 7:44 PM EDT
"you gonna lose your horse. seriously." FX2
Yes we can! November 04, 2008
Yeah. I rounded up.
It’s kind of a lucky circumstance the Rays run differential divided by 2 is about 140. I really got it the way Dewan did by looking at the top scoring team minus the low scoring team and dividing by two. It’s fairly consistent from 2006 through 2008.
JD: Adamant about 78 wins in 2009. Go Rangers!
Interesting
Thanks for posting this!
BTW, if anyone has the Fielding Bible 2007 and 2008, I would really like to see the change in runs saved by the Rays. They allowed 273 fewer runs in 2008 compared to 2007. Dewan’s observation suggests that at most about 140 of the 273 could be attributed to improved defense. Their team FIP went from 4.70 in 2007 to 4.22 in 2008, or a decrease of 80-85 runs. It doesn’t all quite add up when thinking about pitching and defense independently. There may be some synergistic relationship as the Ray’s team ERA went from much above the FIP ERA in 2007 to much below in 2008 (5.53 versus 4.72 in 2007, 3.82 versus 4.22 in 2008).
I’m not sure if you can necessarily say that the 140 is the maximum change.
Remember this is all relative. This is the the best runs saved in the majors for that year vs. worst runs saved.
That does NOT mean that the league in general could not have gotten better in general at saving runs.
In fact, if you take a look at defensive efficiency for 2007 to 2008, the Rays seem do exactly that in going from league worst at 0.656 vs. a league best of 0.705 to a league best at 0.710 vs. a league worst of 0.670 (TEX). That’s an additional swing of 0.020 from a diffference of ~0.040-0.050 which could easily make up the additional ~50 runs saved.
Other possible explanations are that possibly FIPS doesn’t give us accurate information in runs prevented or that the Rays were unlucky in 2007 and lucky in 2008.
R
I'm just not sure how to interpret runs saved.
The 140 is the typical difference between the best and worse teams when tallying runs saved. At least, it’s been the typical difference the past few years.
So, if the Rays were bottom of the league in 2007, then nearly top of the league in 2008, that’s something like a 140 runs saved swing.
I just don’t know if runs saved corresponds 1:1 with actual runs prevented. Dewan writes like it is a 1:1 correspondence, or nearly so.
I think one thing that stands out among all of the statistics is that teams that rate highly on defense nearly always have ERA < FIP.
JD: Adamant about 78 wins in 2009. Go Rangers!
Sure
I guess what I’m trying to emphasize is that even if it is a 1:1 correspondence, you can’t just say that the maximum difference is 140 runs , because the runs saved is relative to the league (IOW, it’s a spread).
And if you take a look at the actual plays turned to outs, you’ll see that in absolute terms, Tampa Bay improved beyond the “maximum” possible from 0.656 to 0.710 which is a difference of 0.055 where usually the difference is around 0.040-0.050.
I think one thing that stands out among all of the statistics is that teams that rate highly on defense nearly always have ERA < FIP.
Isn’t that always going to be the case, a priori? :D
R
Conceptually, I agree, but the stats are not really derived from the same set of data, so
it is possible the numbers could show inconsistencies.
JD: Adamant about 78 wins in 2009. Go Rangers!
Hey, just checked out your sky doc
page. Very clever analogy in your Dec. 31 post, good read.
G G G E-flat_______ F F F D__________....
Thanks.
A few climatologists/meteorologists I’ve tried it on weren’t too excited about it, but most who are not in this field seem to like it.
I’m going to update that site this weekend with the stimulus and omnibus spending going to climate science via NOAA/NSF/NASA and earmarks.
JD: Adamant about 78 wins in 2009. Go Rangers!
Stats not from same source?
Isn’t it though?
I mean defensive efficiency is turning outs from plays.
And Dewan or UZR is based on play-by-play data where outs are an important part of determining the rating (the other part is “playable balls”).
Maybe I’m not viewing it in the right light.
R
I always find it a bit surprising when these complicated normalizations of baseball data
show consistency.
In this case, I think FIP takes into account BB, K, HR, HBP, IBB, and IP, and I think but don’t know for certain that runs saved involves watching tape of every ball in play and some measure of the direction/speed of the ball, which, to me, is independent of BB/K/HR/HBP/IBB/IP. So, I would view these as nearly independent measures.
UZR, defensive efficiency, I think are less independent.
The relationship between FIP and ERA, I need to think about a bit longer, just because I have a meeting to get to in a few minutes.
JD: Adamant about 78 wins in 2009. Go Rangers!
Another interesting thing
The Rays improved in UZR at EVERY SINGLE POSITION.
That’s amazing. Not only that, they improved even though some of the roster stayed the same.
1B: Pena had 1168.7 Innings in 2008 and 1221 Innings in 2007
LF: Crawford had 920.7 INN in 2008 and 1186.3 INN in 2007
Yet both 1B and LF improved in UZR:
1B: -1.4 to 6.3
LF: -7.8 to 23.1 (!)
Here are the other positions:
2B: -6.2 to 4.2 (Upton to Iwamura)
3B: -3.7 to 18.7 (Iwamura to Longoria)
SS: -12.6 to -2.0 (Harris to Bartlett)
CF: -15.3 to 13.3 (1/2 Upton and hodgepodge to full-year experienced Upton)
RF: 2.7 to 10.6 (Young to Gross)
Some interesting questions and thoughts are raised:
1) Did the Rays bring in some defense instructor who vastly improved the defense or some more emphasis on defense? I can understand significant improvements in defense, but not in EVERY SINGLE POSITION especially when the players were essentially the same in some of the positions.
2) Was Crawford injured or overweight in 2007? That’s really the only explanation I can give for the extreme drop in fielding for 2007 (averaged ~ 15 UZR before 2007).
3) Iwamura improving on Upton for 2B means either A) Upton was REALLY bad at 2B, B) Iwamura works much better at 2B, C) or Iwamura improved. It also means that 2B defense and 3B defense might not be that significantly different…..
R
My thoughts
Where as the defense might have gotten better, the picthing staff made strides as well. Adding Garza, plus having Shields and company with more expierence under there belt could have led to balls not being hit as well. Balls that are being smashed are harder to field than softer hit balls.
I’m betting it was a drastic change in pitching staff plus some defensive improvement.
My two cents anyways.
Maybe all that and
they got lucky as well. I expect some regression from both the defenders and pitchers in Tampa this year. The AL East will be an amazing race to watch, I hope injuries don’t decide it.
G G G E-flat_______ F F F D__________....
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
