Mirror, Mirror on the Wall,
Who's the Luckiest Pitcher of All?

Dave Paisley

As the regular season winds down, it's interesting to speculate about who will win the the post season hardware, both on the field and as a result of a bunch of old gits voting. But one of the joys of my life is delving into the background behind the numbers. Mind you, it isn't as much a joy for me as it is for some serious stat geeks. No, some people aren't happy till they've melted a Cray supercomputer with a plethora of complex calculations only slightly more complicated than nuclear physics. No, for me, a simple Excel spreadsheet is plenty, at least until I decide to do that Ph.D. in Economics.

"So what the heck is he babbling about?" you're asking yourself already. Well, I'll tell you what I'm babbling about: there's a lot more to pitching stats than just Wins and ERA. Being the astute strikethree reader you are, you already knew that. You already knew that Aaron Sele isn't the next coming of Randy Johnson, because you know he gets a ton of run support that masks the fact that he's a pretty darned average pitcher. In fact, you're probably wondering to yourself how come he's managed to be so lucky two years in a row.

But there's more to the story than just Aaron Sele,. And that's why I'm here today -- to expose the numbers behind the numbers that are inside the numbers. One of those supercomputer-melting stat geeks that you may want to check out if you're not familiar with his work is Michael Wolverton, of Baseball Prospectus fame.

So without further ado, here's the info on the top AL pitchers.

Name	Tm	W	L	RS	RA	Avg Team R/G	W Pyth	W LA Pyth	W - W Pyth	W - W LA Pyth
P Martinez	Bos	22	4	6.08	2.38	5.58	23	22	-1	0
B Radke	Min	12	14	4.44	4.10	4.32	14	16	-2	-4
O Olivares	Oak	15	10	5.40	4.30	5.58	15	15	0	0
O Hernandez	NYA	16	9	5.37	4.61	5.58	14	14	2	2
M Mussina	Bal	16	7	7.00	4.16	5.35	17	14	-1	2
F Garcia	Sea	16	8	6.28	4.41	5.44	16	14	0	2
S Erickson	Bal	15	11	5.08	4.92	5.35	13	14	2	1
K Appier	Oak	15	13	5.86	5.41	5.58	15	14	0	1
C Nagy	Cle	17	9	7.19	5.05	6.19	17	14	0	3
D Burba	Cle	15	8	6.11	4.51	6.19	15	13	0	2
J Moyer	Sea	14	7	5.13	4.08	5.44	13	13	1	1
A Sele	Tex	17	8	7.32	5.04	5.79	17	13	0	4
D Mlicki	Det	13	12	5.02	5.07	4.52	12	13	1	0
S Ponson	Bal	12	11	5.59	4.73	5.35	13	13	-1	-1
M Sirotka	ChA	10	13	4.55	4.82	4.81	11	13	-1	-3
B Colon	Cle	16	5	7.17	4.36	6.19	15	13	1	3
J Rosado	KC	8	14	5.37	4.66	5.39	13	12	-5	-4
B Moehler	Det	9	16	4.44	5.39	4.52	10	12	-1	-3
D Wells	Tor	14	10	6.07	5.33	5.39	14	12	0	2
R Helling	Tex	13	10	5.58	5.19	5.79	12	12	1	1
C Finley	Ana	11	11	4.97	5.01	4.32	11	12	0	-1
J Baldwin	ChA	10	13	5.24	5.54	4.81	11	11	-1	-1
J Suppan	KC	10	10	5.87	4.81	5.39	12	11	-2	-1
G Heredia	Oak	13	7	6.39	5.20	5.58	12	10	1	3
E Milton	Min	7	11	4.75	5.03	4.32	8	9	-1	-2

The first few columns in the table are self explanatory until RS, which is Run Support, the average runs scored by his team for that pitcher per nine innings. RA is the Runs Allowed by that pitcher per nine innings. The Avg Team R/G is the run scoring by the pitcher's team in all games. From that you can tell that Boston has scored 5.58 runs per game overall, but 6.08 for Pedro Martinez. Not that he needs the extra, of course. After that there's W Pyth, the expected wins for that pitcher based on the Pythagorean approximation using actual run support and runs allowed. Then there's W LA Pyth, the expected number of wins if that pitcher had only league average support. The last two columns are the difference between actual wins and the two different Pythagorean projections. Clear? I doubt it somehow.

Let's follow Pedro's line. In real life, he's allowed a mere 2.38 runs per nine innings while getting 6.08 in support. That translates to 23 wins using the Pythagorean rule (using the same amount of decisions) meaning he's a win shy of what the numbers say is reasonable. With league average support (around 5.50) he would only be expected to have 22 wins, exactly what he does have. Thus the differences are -1 and 0. So Pedro has pretty much earned what he has, and he has earned so much more than any other pitcher this year.

I've ordered the table by the wins predicted with league average support. the first thing it tells you is that there's Pedro, and then there's the rest. The second thing it shows is that uncelebrated names like Brad Radke and Omar Olivares are pretty useful commodities if you can get them into a situation where they get league average support. Another interesting quartet are the pairs of Baltimore and Seattle pitchers, Jamie Moyer, Freddie Garcia, Mike Mussina and Scott Erickson, all of whom are doing pretty well in bad situations. Another point of note is the relatively poor performance of the three Indians pitchers (no real surprise there.) In fact, Bartolo Colon is widely being hailed as the new Cleveland ace, while the truth is the Indians have just scored a truckload of runs for him this year. Sorry, pardon me, I was forgetting that it's his "whammy" powers that cause the Tribe to score the runs for him. Sheesh, Omar Olivares wouldn't be getting anywhere near the hype.

At the other end of the scale there's Brian Moehler of the Tigers, with his 9-16 record that could be more like even money if the Tigers could actually score for him. Maybe he should brush up on his motivational skills.

So if Pedro doesn't win the CY, just call me Grumpy. Otherwise I'll be Happy. With that, it's on to the NL, where the competition is much fiercer at the top.

Name	Tm	W	L	RS	RA	Avg Team R/G	W Pyth	W LA Pyth	W - W Pyth	W - W LA Pyth
R Johnson	Ari	15	9	4.89	2.82	5.53	18	18	-3	-3
K Millwood	Atl	18	7	5.59	3.15	5.12	19	18	-1	0
J Lima	Hou	20	9	5.45	4.02	5.14	19	18	1	2
S Reynolds	Hou	16	12	5.34	4.08	5.14	18	17	-2	-1
K Brown	LA	17	8	4.53	3.55	4.91	15	17	2	0
M Hampton	Hou	20	4	6.81	3.43	5.14	19	16	1	4
G Maddux	Atl	18	8	6.28	3.96	5.12	19	16	-1	2
O Daal	Ari	15	9	5.88	3.97	5.53	16	15	-1	0
A Ashby	SD	14	10	4.50	4.05	4.41	13	15	1	-1
K Bottenfield	StL	18	7	6.05	4.30	5.05	17	15	1	3
S Hitchcock	SD	12	12	5.12	4.28	4.41	14	14	-2	-2
R Ortiz	SF	17	9	5.72	4.78	5.38	15	14	2	3
I Valdes	LA	9	14	4.78	4.29	4.91	13	13	-4	-4
J Guzman	Cin	11	11	5.43	4.25	5.34	14	13	-3	-2
J Schmidt	Pit	13	10	5.00	4.59	4.94	12	13	1	0
T Glavine	Atl	12	11	4.82	4.62	5.12	12	13	0	-1
P Byrd	Phi	14	10	7.21	5.09	5.28	16	12	-2	2
A Leiter	NYN	11	11	5.16	4.69	5.34	12	12	-1	-1
W Williams	SD	10	12	4.48	4.76	4.41	10	12	0	-2
P Astacio	Col	15	11	6.61	5.69	5.63	15	12	0	3
D Hermanson	Mon	9	12	3.92	4.61	4.41	9	12	0	-3
An Benes	Ari	12	12	5.94	5.42	5.53	13	11	-1	1
S Trachsel	ChN	7	17	3.82	5.73	4.61	7	11	0	-4
D Springer	Fla	6	15	4.26	5.22	4.33	8	10	-2	-4
S Estes	SF	11	9	6.34	5.00	5.38	12	10	-1	1
J Lieber	ChN	8	11	5.34	5.05	4.61	10	10	-2	-2
S Karl	Mil	10	11	5.06	5.59	5.03	9	9	1	1
L Hernandez	SF	7	11	4.95	5.04	5.38	9	9	-2	-2
B Bohanon	Col	12	11	6.02	6.30	5.63	11	9	1	3
D Kile	Col	8	13	5.52	7.08	5.63	8	7	0	1

Here we have a tight race between Randy Johnson, Kevin Millwood and Jose Lima. Not much to choose between them, except for the strikeouts and name recognition. Who should get the CY here? I'm not real sure, so you could call me Bashful, or maybe Dopey if you're in a bad mood.

"Where's Mike Hampton?", you might ask. "Is he really the eighth dwarf?" Well, he's almost right there, having passed his motivational skills test, somehow inducing the Astros to rattle off 6.81 runs per game for him, almost 2 runs higher than their average (encouraging Bags and the boys to pick up the pace when he's on the mound). Poor Lima has to put up with barely better than average performance. Shane Reynolds also has got the slightly short end of the stick in Houston. Note that Kent Bottenfield has been the recipient of copious Cardinal runs, one whole run more than average. I suspect his early season run was fueled at an even higher rate, but I'm too lazy to look that one up.

Pity poor Steve Trachsel, languishing last of this group with a paltry 3.82 runs per nine. While his ERA is hardly anything to write home about (and trust me, I'm not) with average support he becomes an average pitcher. Do you think Cubs fans are thinking that right now? I can almost smell the bonfire being lit right now.

I should point out that the Colorado pitchers deserve a special mention here, as league average support hardly does them justice. They get beat up pretty badly in Coors, and so they should be allowed a few more runs in support. Still, don't expect a rush of GMs clamoring for Rockies pitching this winter.

[Note: Michael Wolverton is the guru of what he calls Support-Neutral Win-Loss record (or SNW/L for short.) I'm not going to get into the details here, but it figures out what a pitcher's W-L record would be if he got league average run support. You can catch his work in Baseball Prospectus 1999 as well as the BP website.]

about the author

There's nothing Dave Paisley loves more than the smell of burning spreadsheets in the morning. Tell him where to find extra napalm at drdjp@strikethree.com.

Web Strikethree.com