The Coors Field Effect On Defensive Metrics

Dexter Fowler is a below average defensive outfielder. At least, that's what the advanced defensive metrics have to say. While playing most of his games in center field for the Rockies, Fowler had a -40.6 UZR and a -24 DRS. last year, he posted much better results, but was still considered slightly below average. However, the Rockies' fans I talked to think that the numbers are slightly off and that Fowler is probably an average to an above average defender in center field. The thinking is that outfielders play farther back in Coors Field than they do in any other park. Which could cause havoc on defensive metrics that use a grid-based evaluation.

Prior to starting my research, I searched for any previous research on the effects Coors Field has on defensive statistics. I cam up with exactly zero studies. Most of research has been done on the pitching and offensive side of the ball. Which means I was essentially left to fend for myself.

Advanced defensive metrics only go as far back as 2003, so right off the bat I'm limited in the amount of data I could collect. Since, the data is limited I decided to look at all outfielders instead of just centerfielders. I found 35 outfielders with at least 100 innings played at Coors Field, which isn't a bad start. Unfortunately, I had to throw 10 out because they had never played outside of Coors Field. And, as I'm sure you sharp readers noticed, a 100 innings is hardly enough innings to make a determination about a players defense so there are issues.

Sample size isn't the only issue. There's also age, ball park factors, the fact that other parks can't be parsed out of a players statistics, when they played for the Rockies (going to Coors Field or coming from), position in the outfield and all the usual factors that make defensive metrics shaky in the first place. Which is probably why no one has studied defensive metric differences between Coors Field and other ball parks.

For the players and sample sizes I did have, I collected UZR, DRS and Tango Tigers fan statistic. Since, Rockies fan perception seems to be different about the defensive aptitude of it's players I decided to include the fan scouting statistic, which should represent fan perception of those players. Tango Tiger has only been doing the fan scouting reports for only a few years so again issues with sample size are present.

UZR and DRS are averaged for each player (metric average) to get a single defensive metric number. I totaled all three numbers and also averaged those numbers out as well to get an average defensive metric number per player with which to compare. I hope that's not too confusing. The spreadsheet will be available at the end of this article to take a look at. Hopefully, with that you'll be able to see what I did.

Once I had the data of players who have played in the Rockies outfield, I was then able to begin looking at each individual player and collecting their other team data. Here are the overall numbers:

As a Rockies' outfielder

	Innings	UZR	DRS	Metric Average
Total	35,411.1	-155.1	-169	-162.05
Average		-6.20	-6.76	-6.48

As another team's outfielder

	Innings	UZR	DRS	Metric Average
Total	59,802	-9.60	3.50	-3.05
Average		-.38	0.14	-0.12

Hey look! Better numbers outside of Coors Field. Massively different numbers.

The five players that I had fan scout ratings on averaged out to a 43 in Coors Field and a 41.4 outside Coors Field, which is a little interesting. Maybe there's a bit of bias in over estimating how much of an effect Coors Field has on a players defensive metrics. Most fans contributing to the scouting report are, I imagine, are well aware of the advanced metrics and how they rate their players. With only five players to compare there's not really any assertions we can make.

Next, I decided to refine the pool of players by eliminating all players who didn't have a comparable sample size or had less than 600 innings on either side of the defensive data set.

This left me with 11 players: Michael Cuddyer, Jeromy Burnitz, Preston Wilson, Larry Walker, Matt Holliday, Seth Smith, Eric Young, Tyler Colvin, Carlos Gonzalez, Jay Payton and Willy Taveras.

As a Rockies' outfielder

	Innings	UZR	DRS	Metric Average
Total	24,771.1	-48.8	-94	-71.4
Average		-4.44	-8.55	-6.49

As another team's outfielder

	Innings	UZR	DRS	Metric Average
Total	31,738	17.8	2.5	10.15
Average		1.62	0.23	0.92

The gap has closed a little bit for the total metric average of UZR and DRS but actually widened a little bit looking at the overall average grades. Also, the numbers for the other team has improved in each category. The fan rating has now shifted from an average of 46 for Rockies' players and 49.33 when that player is with another team.

Again, per player these numbers aren't clean. Carlos Gonzalez has played 4903 innings in Coors Field and 689 with another team. Jay Payton played 1350 innings with the Rockies and 4999 innings with other teams. each player has they're own little nuance in regards to these numbers. The only player who truly has enough innings with the Rockies and another team is Matt Holliday. He has 5914 innings with the Rockies and 6148 innings with another team.

Matt Holliday as a Rockies' outfielder

Innings	UZR	DRS	Metric Average
5,914	-1.2	-13	-7.1

Matt Holliday as another team's outfielder

Innings	UZR	DRS	Metric Average
6,148	-10	3.5	-3.25

The data again leans toward Holliday being perceived as a better defender with another team outside of Coors Field.

As I began to work with the collected data, I decided to shoot our stats gurus, Clack and Chris, an email asking for their feedback. Clack suggested I standardized the numbers to help with the sample size issue. This required me to take the defensive metric, divide by innings and then multiply by 100.

Here are the numbers from the 26 players with at least 100 innings played at Coors Field:

As a Rockies' outfielder

	Innings	UZR Standardized	DRS Standardized	Metric Average
Total	35,411.1	-13.68	-13.97	-13.82
Average		-0.55	-0.56	-0.55

As another team's outfielder

	Innings	UZR Standardized	DRS Standardized	Metric Average
Total	59,802	2.72	4.29	3.51
Average		0.11	0.17	0.14

With the numbers standardized we see an overall positive difference in defensive metrics for players who play with another team.

So with that, can we conclusively say that Coors Field has a negative perception of a players value in the outfield? No. There are too many issues right now with some of the factors, particularly the sample size. I do think it is plausible that defensive metrics, more than any other park, struggle to identify a Rockies outfielder's defensive value.

What this means for Fowler is that we could see a small uptick in his defensive numbers moving from Coors Field to Minute Maid Park. It's not a guarantee but the data above does indicate that he might just be better than what he's show so far as a Rockies outfielder. Either way it'll be interesting to watch this season and the seasons to come.

The link to the Google Docs with all the numbers I used is below. I need to make one note about the spreadsheet. The extra outs line is me trying to not pull my hair out with decimal points. Spreadsheets, calculators and computers don't account for three outs in innings numbers being a whole number. Therefore, after grabbing all the data I counted the extra outs by hand and added them to that line.

https://docs.google.com/spreadsheet/ccc?key=0AmVtXqDI2xDudGlnNXY0RGltSXdHZTV6ZkVWVi10ZXc&usp=sharing

As always thoughts and feed back are encouraged.