Baseball Musings: Probabilistic Model of Range, 2005, Runs Created Against Centerfielders

February 20, 2006

Probabilistic Model of Range, 2005, Runs Created Against Centerfielders

One question that came up recently is what the RCA number represents. Basically, imagine that every batter put a ball in play that was catchable by the particular fielder. For a centerfielder, imagine players just keep hitting line drives and fly balls to centerfield. No homers, no walks, no strikeouts. That's how many runs you'd expect the team to score before the CF made 27 outs.

Here are the numbers for centerfielders.

Probabilistic Model of Range, 2005, Runs Created Against Centerfielders, Original Model (minimum 200 fieldable balls in play)
Player	Fieldable Balls In Play	Actual Outs by Fielder	Predicted Outs by Fielder	RCA	Predicted RCA	RCA/27 Outs	Predicted RCA/27	Runs Saved/27 Outs
Joey R Gathright	375	181	167.23	23.76	48.36	3.54	7.81	4.263
Jerry Hairston	200	90	84.03	22.04	31.70	6.61	10.19	3.574
Jason Ellison	403	197	178.25	36.49	54.52	5.00	8.26	3.257
Andruw Jones	800	365	337.56	99.20	131.84	7.34	10.55	3.208
Jim Edmonds	673	319	297.21	69.83	98.88	5.91	8.98	3.072
Gary Matthews Jr.	586	258	242.31	71.66	91.31	7.50	10.17	2.675
Grady Sizemore	796	373	370.07	69.49	103.55	5.03	7.56	2.525
Jason Michaels	334	161	150.73	28.36	40.42	4.76	7.24	2.485
Willy Taveras	771	332	322.83	81.34	107.70	6.61	9.01	2.393
Shawn Green	212	81	84.00	30.23	38.16	10.08	12.26	2.188
Aaron Rowand	811	388	362.99	70.62	95.24	4.91	7.08	2.170
Nook P Logan	581	282	270.92	45.57	63.75	4.36	6.35	1.990
Curtis Granderson	237	119	110.91	31.15	37.00	7.07	9.01	1.940
Mark Kotsay	691	299	306.87	73.41	95.91	6.63	8.44	1.810
Tike Redman	355	158	143.77	52.38	56.29	8.95	10.57	1.621
Jeremy T Reed	834	384	384.13	63.96	86.41	4.50	6.07	1.576
Corey Patterson	526	240	232.53	59.95	71.17	6.74	8.26	1.520
Vernon Wells	832	351	356.22	91.64	112.47	7.05	8.52	1.476
Luis Terrero	268	121	121.69	21.72	28.23	4.85	6.26	1.416
Laynce Nix	355	160	159.88	40.08	48.39	6.76	8.17	1.409
Brady Clark	823	399	380.69	75.00	91.32	5.07	6.48	1.402
Damon J Hollins	472	198	197.37	56.59	66.43	7.72	9.09	1.371
Jason Repko	240	97	105.29	20.41	27.29	5.68	7.00	1.314
Luis Matos	642	299	286.93	82.44	91.75	7.44	8.63	1.189
Johnny Damon	878	396	402.01	106.64	122.95	7.27	8.26	0.986
Randy Winn	382	184	182.71	39.69	44.62	5.82	6.59	0.769
Torii Hunter	510	218	220.35	56.90	62.46	7.05	7.65	0.606
Carlos Beltran	806	378	372.03	76.84	83.59	5.49	6.07	0.578
Kenny Lofton	458	201	207.17	47.87	53.71	6.43	7.00	0.570
Cory Sullivan	438	172	179.90	67.78	74.36	10.64	11.16	0.519
Lew Ford	348	140	150.24	46.02	51.86	8.87	9.32	0.446
Juan Pierre	790	332	337.90	107.44	111.48	8.74	8.91	0.170
Brad Wilkerson	509	234	230.76	61.26	60.46	7.07	7.07	0.006
Dave Roberts	579	234	240.18	73.35	73.60	8.46	8.27	-0.190
David DeJesus	672	306	313.16	87.76	86.92	7.74	7.49	-0.250
Milton Bradley	416	181	183.19	56.16	54.76	8.38	8.07	-0.307
Chone Figgins	296	131	134.46	34.32	32.73	7.07	6.57	-0.502
Bernie Williams	556	226	245.61	63.33	63.29	7.57	6.96	-0.609
Steve Finley	598	266	279.55	82.21	78.10	8.34	7.54	-0.801
Preston Wilson	652	267	283.89	95.84	93.25	9.69	8.87	-0.822
Ken Griffey Jr.	695	286	321.33	114.51	101.19	10.81	8.50	-2.308
Jose Cruz	224	87	96.22	37.26	32.86	11.56	9.22	-2.343

Notice Curtis Granderson and Nook Logan are very close. However, the balls in play vs. Granderson seem to have a higher run value than the balls in play vs. Logan. Logan played twice as much, so maybe it's just sample size. Looking at the opponents Granderson faced vs. the opponents Logan faced, I'd say a higher proportion of Granderson's opponents were stronger teams. Logan faced all the NL West teams, Granderson none. Maybe Granderson was in behind worse pitching?

Posted by David Pinto at 03:11 PM | Probabilistic Model of Range | TrackBack (0)

Comments

I know that the season is getting close when I see the Probabilistice Model of Range tables at Baseball Musings....

Posted by: stat man at February 20, 2006 07:00 PM

As a Rays fan, I have to admit being surprised at seeing Gathright #1. He had alot of struggles in the early part of the year (in fact, when Piniella sent him down to Durham, he said Gathright needed to work on his defense). Admittedly, he looked 10x better when he came back up, and I can see him being above-average thanks to his incredible speed. But it's hard for me to believe he's the league best.

But I'll take it, and please forward this onto the Marlins brass. They need to see this if we are going to get a very good pitching prospect from them. Love your stuff, David!

Posted by: Jay at February 20, 2006 07:15 PM

Seems to be an affirmation of Jones' talents... when a lot of people have been saying he's been slipping -- which perhaps he has, but still looks pretty good by this reckoning... thanks for printing this!

Posted by: Subtle guy at February 20, 2006 07:20 PM

Where would Chris Duffy fall in if you normalized his opps to 200?

Posted by: Pirate at February 20, 2006 07:45 PM

This is an excellent metric for measuring the defensive value of a player. I do wonder how you arrive at predictive values, though.
I'm looking for a measurement of the total player. I now feel the need for complementary offensive stats for the total offense. How about a stat allowing one to view a player's defensive performance in the context of how many runs he produces. It could be made up of number of on base runners he scored, his RBI's and his runs scored. I leave it to you to figure out how you would develop predictive values for offensive production.

Posted by: Paul E Keane at February 22, 2006 02:53 PM