Why does CC Sabathia allow so many unearned runs?

Posted Monday, September 17th, 2012 at 10:00 am by Doug

Fielding in baseball (if not fielders, necessarily) is getting better all the time. Improvements in equipment, field conditions and, especially, defensive positioning, mean today’s players have a leg up on their predecessors. Indeed, today’s computer analysis of batter and pitcher tendencies is of a sophistication unimagined even as recently as a decade ago. And, to top it off, recent years have seen progressively fewer balls in play due to ever increasing numbers of strikeouts. The end result – fewer errors and fewer unearned runs.

Nevertheless, unearned runs haven’t disappeared entirely and some pitchers seem to allow more of them than others. As an example, these are the starting pitchers of 2012 (min. 150 IP) with the three highest ratios of unearned runs allowed to total runs allowed.

Rk	Player	R	ER	IP	Year	Age	Tm	G	GS	CG	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	C.J. Wilson	93	78	188.0	2012	31	LAA	31	31	0	12	9	.571	80	159	3.73	101	17
2	CC Sabathia	85	71	176.0	2012	31	NYY	25	25	2	13	6	.684	40	169	3.63	115	21
3	Aaron Harang	82	69	164.0	2012	34	LAD	28	28	0	9	9	.500	77	123	3.79	101	14

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/16/2012.

Surprised? I mean those are all pretty decent pitchers. Why would it be those guys? After the jump, I’ll explain why perhaps this shouldn’t be so surprising.

Here’s that same 2012 list expanded to the top 10 in unearned run ratio, those pitchers allowing more than 12% unearned runs.

Rk	Player	R	ER	IP	Year	Age	Tm	G	GS	CG	SHO	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	James Shields	100	86	205.1	2012	30	TBR	30	30	2	2	14	9	.609	53	193	3.77	100	24
2	C.J. Wilson	93	78	188.0	2012	31	LAA	31	31	0	0	12	9	.571	80	159	3.73	101	17
3	Homer Bailey	91	79	181.0	2012	26	CIN	29	29	1	0	11	9	.550	47	142	3.93	107	24
4	Wandy Rodriguez	90	77	189.2	2012	33	TOT	31	30	0	0	11	13	.458	51	126	3.65	108	18
5	Phil Hughes	88	77	175.0	2012	26	NYY	29	29	1	0	15	12	.556	40	148	3.96	106	33
6	CC Sabathia	85	71	176.0	2012	31	NYY	25	25	2	0	13	6	.684	40	169	3.63	115	21
7	Aaron Harang	82	69	164.0	2012	34	LAD	28	28	0	0	9	9	.500	77	123	3.79	101	14
8	Matt Moore	78	68	166.1	2012	23	TBR	28	28	0	0	10	10	.500	73	165	3.68	102	17
9	Justin Verlander	78	68	217.1	2012	29	DET	30	30	6	1	14	8	.636	55	218	2.82	148	18
10	Ross Detwiler	62	53	151.0	2012	26	WSN	30	24	0	0	9	6	.600	43	94	3.16	127	11

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/16/2012.

Pretty decent set of pitchers. None with ERA+ below 100 or ERA above 4. And, with Verlander, Shields and CC, some of the top echelon pitchers in the game.

Maybe you’re saying this is one season. Doesn’t mean anything. Okay, fair enough. How about 50+ years. Here are the starting pitchers since 1961 with seasons (min. 162 IP) with unearned runs allowed above 15% of total runs allowed.

Rk		Yrs	From	To	Age
1	Juan Marichal	8	1963	1973	25-35	Ind. Seasons
2	Gaylord Perry	7	1968	1980	29-41	Ind. Seasons
3	Jim Kaat	7	1961	1974	22-35	Ind. Seasons
4	Tommy John	6	1967	1979	24-36	Ind. Seasons
5	Kevin Brown	5	1992	2003	27-38	Ind. Seasons
6	Greg Maddux	5	1990	1999	24-33	Ind. Seasons
7	Phil Niekro	5	1967	1986	28-47	Ind. Seasons
8	Dennis Martinez	4	1981	1992	26-37	Ind. Seasons
9	Nolan Ryan	4	1974	1984	27-37	Ind. Seasons
10	Dock Ellis	4	1971	1977	26-32	Ind. Seasons
11	Jim Rooker	4	1970	1978	27-35	Ind. Seasons
12	Sonny Siebert	4	1966	1972	29-35	Ind. Seasons
13	Mike Cuellar	4	1966	1973	29-36	Ind. Seasons
14	Don Drysdale	4	1962	1968	25-31	Ind. Seasons
15	Dean Chance	4	1962	1967	21-26	Ind. Seasons
16	Bob Gibson	4	1961	1973	25-37	Ind. Seasons

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 5/16/2012.

Wow! That’s quite a list of pitchers. Obviously tilted to the first part of the post-1961 period owing to the factors cited in the preamble. But, why such outstanding pitchers, as opposed to a more typical assortment of hurlers?

Here is my hypothesis. I welcome your feedback (not that I have to invite you, but go ahead and poke as many holes as you like). So, here goes.

1. This group will be allowing fewer runs (even a lot fewer runs) than “average” and “below average” pitchers in any era. Don’t think there should be much debate on this point.

2. The defense behind this group will play better than the defense behind an average or below average pitcher. Why? Fewer base runners, fewer pitches, fewer balls in play, balls in play not hit as hard, pitchers (probably) working faster – all these factors suggest less pressure on the defense and, ergo, less likelihood of committing errors that lead to unearned runs.

3. Factor #1 will be more pronounced than factor #2. That is, while the defense will play better behind this group, it’s not like they’re going to commit only half as many errors as they would behind a below average pitcher. However, a top rank pitcher may very well allow runs at a rate only half that of a below average pitcher, and perhaps 2/3 the rate of an average pitcher.

4. Taking the 3 points together, the reduced absolute number of unearned runs allowed that is expected for this group (point 1 and 2) will still be a larger proportion of their much lower total runs allowed (point 1 and 3), than would be the case for less skilled pitchers (i.e. the great majority of major league pitchers).

So, there’s the hypothesis. It would be great if I could get stats on errors committed behind pitchers (point 2), but I don’t know where those figures might be. I expect the rest of the argument should be fairly self-evident. But, if I’m wrong, please let me know.

Finally, to illustrate my point about unearned runs declining at ever increasing rates, here are lists identifying the pitchers with the highest unearned run ratios by period. Notice the decline in unearned run ratio thresholds needed to generate lists of comparable size.

For 1901-1919, here are careers (min. 1000 IP) with unearned runs more than 33% (yes, 33%) of runs allowed.

Rk	Player	R	WHIP	ER	IP	From	To	Age	G	GS	CG	SHO	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	Harry Howell	952	1.167	634	2230.0	1901	1910	24-33	289	245	214	18	110	133	.453	561	900	2.56	111	22
2	Jack Taylor	830	1.144	553	2008.0	1901	1907	27-33	237	217	210	17	119	101	.541	432	527	2.48	115	31
3	Addie Joss	730	0.968	488	2327.0	1902	1910	22-30	286	260	234	45	160	97	.623	364	920	1.89	142	19
4	Ed Siever	674	1.235	435	1507.0	1901	1908	26-33	203	174	136	14	83	82	.503	311	470	2.60	117	24
5	Joe Lake	671	1.260	417	1318.0	1908	1913	27-32	199	139	95	8	62	90	.408	332	594	2.85	99	19
6	Johnny Lush	571	1.276	369	1239.1	1904	1910	18-24	182	155	105	16	66	85	.437	413	490	2.68	97	17
7	Smoky Joe Wood	492	1.084	319	1432.1	1908	1919	18-29	224	158	121	28	117	57	.672	419	988	2.00	149	10
8	Phil Douglas	471	1.131	314	1103.0	1912	1919	22-29	189	131	63	13	54	69	.439	266	524	2.56	112	23
9	Andy Coakley	436	1.245	280	1072.1	1902	1911	19-28	150	124	87	11	58	59	.496	314	428	2.35	112	9
10	Jack Pfiester	365	1.089	240	1067.1	1903	1911	25-33	149	128	75	17	71	44	.617	293	503	2.02	128	6

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 5/16/2012.

For 1920-1949, let’s ratchet down that unearned run threshold to 18% of runs allowed.

Rk	Player	R	WHIP	ER	IP	From	To	Age	G	GS	CG	SHO	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	Dutch Leonard	1309	1.251	1059	2933.1	1933	1949	24-40	474	373	192	30	172	169	.504	634	1048	3.25	118	136
2	Jimmy Ring	1172	1.543	948	1944.0	1920	1928	25-33	312	251	126	3	96	128	.429	819	713	4.39	97	96
3	Joe Shaute	1043	1.447	838	1818.1	1922	1934	22-34	360	208	103	5	99	109	.476	534	512	4.15	99	76
4	Elam Vangilder	1010	1.513	812	1702.2	1920	1929	24-33	364	186	89	13	98	102	.490	697	468	4.29	100	93
5	Stan Coveleski	851	1.319	688	1933.2	1920	1928	30-38	274	253	138	21	133	89	.599	510	558	3.20	127	53
6	Hooks Dauss	813	1.425	653	1521.2	1920	1926	30-36	269	169	89	6	97	90	.519	476	462	3.86	104	62
7	Bill Doak	596	1.351	477	1268.2	1920	1929	29-38	228	172	66	14	82	68	.547	384	385	3.38	109	42
8	Joe Heving	559	1.460	450	1038.2	1930	1945	29-44	430	40	17	3	76	48	.613	380	429	3.90	108	64
9	Rube Marquard	541	1.386	437	1000.1	1920	1925	33-38	176	130	47	6	52	60	.464	256	341	3.93	94	43

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 5/16/2012.

For 1950-1992, we’re down to 15% unearned runs.

Rk	Player	R	WHIP	ER	IP	From	To	Age	G	GS	CG	SHO	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	Curt Simmons	1368	1.284	1158	3038.0	1950	1967	21-38	499	424	153	36	181	160	.531	894	1519	3.43	114	240
2	Juan Marichal	1329	1.101	1126	3507.0	1960	1975	22-37	471	457	244	52	243	142	.631	709	2303	2.89	123	320
3	Sonny Siebert	907	1.213	767	2152.0	1964	1975	27-38	399	307	67	21	140	114	.551	692	1512	3.21	110	197
4	Randy Jones	875	1.251	735	1933.0	1973	1982	23-32	305	285	73	19	100	123	.448	503	735	3.42	101	129
5	Dean Chance	832	1.212	697	2147.1	1961	1971	20-30	406	294	83	33	128	115	.527	739	1534	2.92	119	122
6	Hoyt Wilhelm	773	1.125	632	2254.1	1952	1972	29-49	1070	52	20	5	143	122	.540	778	1610	2.52	147	150
7	Al Jackson	725	1.336	614	1389.1	1959	1969	23-33	302	184	54	14	67	99	.404	407	738	3.98	91	115
8	Jim Hearn	709	1.350	597	1410.0	1950	1959	29-38	308	191	56	9	88	73	.547	534	567	3.81	105	137
9	Zane Smith	692	1.317	583	1485.1	1984	1992	23-31	281	216	31	14	75	86	.466	483	828	3.53	107	85
10	Bill Krueger	571	1.492	484	1024.0	1983	1992	25-34	233	143	8	2	57	57	.500	431	516	4.25	92	85
11	Mel Parnell	567	1.422	481	1194.2	1950	1956	28-34	200	167	69	15	81	57	.587	507	510	3.62	121	88
12	George Stone	524	1.361	441	1020.2	1967	1975	20-28	203	145	24	5	60	57	.513	270	590	3.89	96	122
13	Ed Halicki	509	1.262	428	1063.0	1974	1980	23-29	192	157	36	13	55	66	.455	334	707	3.62	102	82
14	Ron Perranoski	442	1.332	364	1174.2	1961	1973	25-37	737	1	0	0	79	74	.516	468	687	2.79	124	50

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 5/16/2012.

And, for 1993-2011, all the way down to 11% unearned runs allowed.

Rk	Player	R	WHIP	ER	IP	From	To	Age	G	GS	CG	SHO	W	L	W-L%	BB	SO	ERA	ERA+	HR
1	Tim Wakefield	1765	1.355	1560	3134.1	1993	2011	26-44	614	450	29	5	192	179	.518	1170	2105	4.48	104	415
2	Mike Hampton	1157	1.442	1024	2268.1	1993	2010	20-37	419	355	21	9	148	115	.563	901	1387	4.06	107	200
3	Kevin Brown	942	1.176	828	2380.2	1993	2005	28-40	359	349	47	14	155	101	.605	597	1920	3.13	135	155
4	Bobby Jones	833	1.351	735	1518.2	1993	2002	23-32	245	241	11	4	89	83	.517	412	887	4.36	94	194
5	Julian Tavarez	808	1.498	696	1404.1	1993	2009	20-36	828	108	2	0	88	82	.518	563	842	4.46	101	113
6	Dustin Hermanson	675	1.360	600	1283.0	1995	2006	22-33	357	180	4	2	73	78	.483	460	874	4.21	105	160
7	Wandy Rodriguez	598	1.346	532	1176.0	2005	2011	26-32	206	197	2	2	73	75	.493	422	1004	4.07	102	134
8	Tom Candiotti	595	1.339	514	1116.2	1993	1999	35-41	206	175	12	1	56	71	.441	359	705	4.14	97	122
9	Felix Hernandez	564	1.224	500	1388.1	2005	2011	19-25	205	205	18	4	85	67	.559	424	1264	3.24	128	116
10	Ramon Martinez	564	1.375	499	1156.0	1993	2001	25-33	186	185	16	10	83	51	.619	527	841	3.88	104	108
11	Brandon Webb	557	1.239	479	1319.2	2003	2009	24-30	199	198	15	8	87	62	.584	435	1065	3.27	142	92

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 5/16/2012.

Subscribe to: RSS feed

32 Comments

1

Ed says:

September 17, 2012 at 10:16 am

Doug – I wonder what role official scorers play in this. Just an example. Justin Verlander. For his career, 11.5% of his runs at home have been unearned whereas on the road only 4.3% have been unearned. A pretty big difference. Obviously hard to say exactly what causes that but I wonder if official scorers are more likely to give big name pitchers the benefit of the doubt on a close call, particularly when pitching at home. Kind of like how NBA superstars always get the foul calls. Anyway, just something to think about.

Reply
- 3
  
  Doug says:
  
  September 17, 2012 at 11:07 am
  
  Good thought, Ed.
  
  Also, my sense, from observation, is that on a debatable play as to whether to call a hit or error, the scorer tends to be far more likely to call it a hit for the home team, but an error if the away team puts that ball in play. Presumably that washes out with players playing as much at home as on road, except that there’s obviously much less variation in scorers for the home games as those on the road, and the degree to which the postulation above is true very likely varies somewhat (or, possibly, more) according to individual scorers.
  
  Reply
  - 7
    
    Ed says:
    
    September 17, 2012 at 11:20 am
    
    Obviously another possibility is that a pitcher might be more likely to have a melt down on the road following an error. That’s at least an alternative explanation for the Verlander data I posted.
    
    Reply
- 6
  
  John Autin says:
  
  September 17, 2012 at 11:20 am
  
  I took a Q&D look at this question, using pitchers with 100+ IP in 2011.
  
  – The 20 pitchers with the highest UERAs (avg. 0.61) had an avg. ERA of 3.89 and an avg. RA of 4.50. The 20 with the lowest UERAs (avg. 0.15) had an avg. ERA of 3.41 and an avg. RA of 3.55.
  
  – The 20 lowest ERAs (avg. 2.78) had an avg. UERA of 0.27. The 20 highest ERAs (avg. 4.80) had an avg. UERA of 0.38.
  
  Small sample, and admittedly there’s a certain circularity to the method. But it doesn’t discourage the notion that official scorers may be unconsciously influenced by a pitcher’s reputation.
  
  Reply
2

Bryan O'Connor says:

September 17, 2012 at 10:17 am

Doug, if you’re just looking for an answer to the titular question of this post, I’ll submit this: Derek Jeter and Curtis Granderson. By fangraphs’ FRAA, they’re the worst infielder and worst outfielder, respectively, in the AL this season. Then again, neither gets charged with absurd amounts of errors, so their ineptitude goes farther toward explaining Sabathia’s ERA underperforming his FIP again this year (though, oddly, the team’s 3.89 ERA is better than its 3.99 FIP) than his unearned runs.

As to the question of all those great pitchers giving up high percentages of unearned runs, could official scorers be assuming that these great pitchers are inducing weak contact and assigning errors to fielders on plays they might have called hits behind lesser pitchers? Kind of like superstars drawing more fouls in the NBA? Just a thought.

Reply
- 4
  
  Ed says:
  
  September 17, 2012 at 11:18 am
  
  Bryan – Looks like you and I were posting similar thoughts at the same time!
  
  Reply
- 11
  
  Joseph says:
  
  September 17, 2012 at 1:59 pm
  
  Bryan: My understanding of why Jeter is rated as a poor infielder is not because of errors he makes (his fielding % as a SS is 27th all time, so that’s pretty good, no?), but balls he doesn’t get to in time to make a play; i.e., hits. Hits count towards earned runs, no?
  
  Reply
  - 28
    
    mosc says:
    
    September 18, 2012 at 1:36 pm
    
    I agree. The consensus is Jeter’s range is bad. He’s got a strong and accurate arm and a solid glove too. He’s not going to give you a lot of unearned runs even if he’s not saving you as many runs as a guy with better range could.
    
    Reply
5

no statistician but says:

September 17, 2012 at 11:20 am

In other words, you’re saying that while earned runs are variable depending in large part on a pitcher’s skill in a particular year, unearned runs are going to happen within a small range that impacts on all pitchers fairly equally. Ergo, the fewer earned runs, the higher the percentage of unearned runs. Sounds reasonable, but are there any pitchers whose records run contrary to this phenomenon?

A stray comment: I would have guessed Hoyt Wilhelm to be at the top of any list for unearned runs for an era. Only 6th? I demand a recount.

Reply
- 8
  
  Doug says:
  
  September 17, 2012 at 11:41 am
  
  Yes, that’s my hypothesis, nsb. But, not stated as succinctly and elegantly as you did. Thanks.
  
  The list where you found Wilhelm is ranked by total runs allowed. I can generate that list base on pitchers exceeding a % threshold of unearned runs (derived from relationship between R and ER), but I can’t rank according to that percentage because it’s not a variable in the P-I database.
  
  At a glance, I see Wilhelm, Niekro, Wakefield and Candiotti are all present on these lists, an indication, perhaps, that knuckleballers may experience more errors behind them, with catchers rushing throws to try to catch base-stealers or runners moving up on ball that gets away from the catcher.
  
  Reply
9

BryanM says:

September 17, 2012 at 1:12 pm

I have a couple of thoughts; but first, when i search for pitchers having at least 12.5% more R than ER this year, I also get Kevin Millwood (86/76) , Rick Porcello (96/85) , and clayton Richard (95/84) , none of whom is tearing up the league, so really good pitchers are not particularly dominant on the list this year
however, your career data for 1993-2011 is more compelling, since single-season blips have had a chance to even themselves out — there you have 3 really good guys, and 8 who are about average, so there is at least some eyeball evidence that good pitchers give up more unearned runs , probably for the reasons cited above and maybe for

a) both good and bad pitchers are sometimes pulled in mid-inning, after which the runner(s) they left sometines score — if this happens equally to good and bad pitchers ; it obviously affects the ratios of good pitchers more.

b) more generally , If unearned runs are distributed truly randomly, you would expect UR/IP to be pretty evenly around a mean, with no bias to good or bad pitchers, but simply variation due to luck – let’s say the league URA (unearned run average ) is 0.32 (8% of a 4.00 ERA) this .32 is obviously much more significant to a 3 era pitcher than one with an ERA of 5

that said , the “scorer ” argument (protecting the ERA of stars) makes sense to me and is probably a factor.

Reply
- 10
  
  Doug says:
  
  September 17, 2012 at 1:36 pm
  
  Thanks Bryan,
  
  My selection criteria for the 2012 list was ER < 0.88 * R, which is a bit different than R > 1.125 * ER, which, I gather, is the query you used to get the three guys you mentioned.
  
  I like your point about inherited runners scoring. I know relievers have a stat for inherited runners they prevent from scoring. Seemes that starters should also have a stat – say Bequeathed Runners Scoring that could be used to calculate OERA (Own ERA). Just to see if some starters are more unlucky than others in this regard.
  
  Reply
  - 22
    
    bstar says:
    
    September 18, 2012 at 12:41 am
    
    Doug, Baseball Prospectus does have stats for this. You called it-it’s Bequeathed Runners & Bequeathed Runners Scored.
    
    Reply
    - 24
      
      Doug says:
      
      September 18, 2012 at 1:12 am
      
      Thanks, bstar.
      
      I’ll take a look. I think that difference between ERA and what I called OERA (if there is such a thing) would be something interesting to look at.
      
      Reply
    - 29
      
      Evan says:
      
      September 18, 2012 at 5:58 pm
      
      BB-R tracks this too. It’s under the Starting Pitching section of the More Stats player page. BQR and BQS (Runners left and runners left that scored).
      
      Reply
- 13
  
  Evan says:
  
  September 17, 2012 at 2:38 pm
  
  I wouldn’t expect unearned runs to be distributed randomly. I might expect errors to be evenly distributed (excepting official scoring biases discussed in the other comments), but I would expect superior pitchers to have lower UERA. Superior pitchers should have fewer runners on base when those errors are committed and should be better at run prevention after the error is committed as well.
  
  Reply
  - 15
    
    Doug says:
    
    September 17, 2012 at 3:03 pm
    
    Good distinction, Evan, between errors and unearned runs.
    
    Also, the top pitchers are probably more likely to be given the rope to get out of a tight spot, as opposed to being relieved and having the runners then on base score after his departure. Question is, whether being given that rope was a good idea by the manager.
    
    Reply
    - 20
      
      BryanM says:
      
      September 17, 2012 at 4:19 pm
      
      And another thing ; not sure whether this is statistically valid , but I divided my good group into strikeout pitchers, (Unit, Clemens, J Santana, Pedro ( 8.73-11.04 SO/9)), and guile guys (Brown,Webb, Maddux, Halladay, (6.14-7.26) the ERAs were virtually the same (3.15 to 3.17) but the guile guys gave up 28% more unearned runs, over ten thousand IP ; in fact the guile guys were the same URA (0.40) as the replacement -level group I mentioned above.
      
      Unsurprising conclusion; pitchers who rely on their defense give up more unearned runs than the guys who only need a catcher
      
      Reply
    - 23
      
      bstar says:
      
      September 18, 2012 at 12:46 am
      
      I guess arguing whether giving that rope to an elite pitcher is a good idea or not is a question for another thread, but there certainly is mounting evidence that’s it’s not. It has to do with that rope being given during a pitcher’s third or fourth time thru the order and/or the effectiveness of one-inning-or-less relief pitchers.
      
      Reply
  - 17
    
    BryanM says:
    
    September 17, 2012 at 3:25 pm
    
    In fact this intrigued me so much that I ran a little study – Apologies in advance since I have no idea how to insert tables and such into comments . But I took all the pitchers that pitched at least 1000 innings in the 19 year period 1993 -2011,starting at least 60% of their games, the same period used by Doug above and ranked them by ERA+.
    The top 8 guys (Pedro, Big Unit, B Webb, G Maddux, Rocket, J Santana, K Brown, Doc) pitched a total of 20895 innings, which is a lot, and gave up 7371 earned runs, ( ERA=3.16) and 819 unearned runs (URA=0.35) ,
    The ERA+ of the group ranged from 154-135. At the bottom of the same list, I took the worst ERA+ guys to add up to roughly the same innings . Since worse pitchers don’t get the same innings ,it took me 16 guys to get up to 20905 IP; These guys gave up 11661 ER and 936 unearned runs (ERA= 5.02, URA =0.40)
    
    Conclusions — As Evan says above, the worst pitchers gave up slightly more unearned runs than the best ones , although this effect is small (.40 to .35), probably because even Pedro is not going to get the run back if an outfielder drops a routine fly ball with a man on third and two out, and because unearned runs are mostly luck. Also there is a lot of variation within the groups
    but , as a percentage of total runs, unearned runs are 10% of the total for the best pitchers, and only 7.4% for the Josh Fogg group — just an artifact of the vast difference in earned runs
    
    Reply
    - 18
      
      BryanM says:
      
      September 17, 2012 at 3:29 pm
      
      should have noted the worst pitchers had an ERA+ of between 83and 90 , so just better than replacement (which is probably why they kept getting handed the ball…
      
      Reply
    - 19
      
      Doug says:
      
      September 17, 2012 at 3:34 pm
      
      Good work, Bryan.
      
      Thanks for sharing that.
      
      Reply

Bruce Reznick says:

September 17, 2012 at 2:16 pm

I know it’s small sample size, but this season, Jason Berken has appeared twice.

In 1 IP with Baltimore he gave up 2 ER and 5 UER.
In 4 IP with the Cubs, he gave up 2 ER and 4 UER.

In his previous 229 IP, he gave up 136 ER and 9 UER, so he’s doubled his career numbers. His ERA this year is 7.20, which is pretty good considering his WHIP is 3.20. Batters have a cozy .500/ .531/ 1.000
against him. He is expected to get another start on Thursday.

Doug says:

September 17, 2012 at 2:58 pm

Nice find Bruce,

Assuming lightning doesn’t strike 3 times (or Berken allows no runs in that next start), he will probably fall out of this group of players allowing over 3 times as many runs as earned runs (min. 4 ER).

Rk	Player	R	ER	Year	Age	Tm	Lg	G	GS	CG	W	L	IP	H	BB	SO	ERA	ERA+	HR
1	Ollie Johns	22	7	1905	25	CIN	NL	4	1	1	1	0	18.0	31	4	8	3.50	96	1
2	Crazy Schmit	20	5	1901	35	BLA	AL	4	3	1	0	2	22.2	25	16	2	1.99	197	0
3	Jason Berken	13	4	2012	28	TOT	ML	2	1	0	0	1	5.0	14	2	2	7.20	65	3
4	Harry Kimberlin	13	4	1937	28	SLB	AL	3	2	1	0	2	15.1	16	9	5	2.35	211	2
5	Nick Altrock	13	4	1902	25	BOS	AL	3	2	1	0	2	18.0	19	7	5	2.00	181	0

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/17/2012.

Crazy numbers indeed for Crazy Schmit, in his final ML season.

16

Doug says:

September 17, 2012 at 3:12 pm

I’ll throw another log on the fire as to why unearned run rates have fallen – the end of scheduled double-headers. Up until the early 1970s, 25 or 30 (or more) double-headers in a season wasn’t uncommon. That’s a significant number of games being played by fielders probably not as sharp, physically or mentally, as they would be otherwise. Of course, the flip side is lots of double-headers also means lots of off days, though traveling and exhibitions probably meant a lot of those off days really weren’t.

Would be great to get some fielding stats added into the P-I.

Reply
21

Jeff Angus says:

September 18, 2012 at 12:07 am

It would take a lot of work to isolate out what I suspect is a major SUB-group of these high-UER ratio hurlers:
Runs Ruled Unearned Because of a Passed Ball.

You have as already mentioned here a mess ‘o knuckleballers (add Dutch Leonard), but also a pile ‘o spitballers (from the post-ban era, Phil Douglas, Coveleski, Spitting Bill Doak, from before ban… .well you may know the names of the infamous ones from before the ban, but I don’t).

Anyway, IF you had the resources to isolate out the UER attributable to PBs, I think you could parse these tables into paired ones, clustering the PB engenderers, and then the residual, bigger (I’m guessing) pile might have clearer factors that tied them together as a cluster.

Interesting study, thanks.

Reply
- 25
  
  Doug says:
  
  September 18, 2012 at 1:19 am
  
  “It would take a lot of work …” Absolutely, Jeff. Thanks for your thoughts.
  
  I was interested in just getting some feedback on my rationalization for what I was seeing in the lists. And, I got that … and more. Bryan M in comment #17 has provided some empirical evidence supporting the basic idea, using a fairly simple method that I’m sure I probably wouldn’t have thought of anytime soon. (BTW, I would have been just as happy if Bryan’s little study provided evidence refuting the basic idea here).
  
  Reply
  - 26
    
    BryanM says:
    
    September 18, 2012 at 8:20 am
    
    Doug; Your idea that good pitchers give up a lot of unearned runs is a very important question , and gets to the heart of baseball’s double -entry bookkeeping system ; every “event” is attributed to the batter or baserunner and simultaneously to the pitcher and/or defense. Unearned runs are attributed in the traditional statistics only to the defense, so if we find a systemic way in which pitchers ” cause ” unearned runs , say by PB @21 above, we could contribute to pitcher evaluation. The evidence I found @17 , which could clearly be overturned by more and better data was evidence for 4 things, IMHO:
    
    1) there is variation between pitchers in unearned runs, allowed just as for earned runs, in MLB in 2011 , 8.37% of runs were unearned , expressed as a rate stat (R-ER)/R, better pitchers have a slight tendency to have a higher rate.
    
    2) this fact is (more than ) entirely due to the denominator being lower for better pitchers, they actually give up fewer unearned runs than weaker hurlers, but only a little ;
    
    3) the 0.05 run advantage for better pitchers (.35 to .40) translates into 1 run per year for a 180 inning season, not worth worrying about ON AVERAGE
    
    4) however , there is a large variation among individual pitchers, masked by the group averages,
    the difference between the best ( Johan Santana 0.23) and the worst ( Jaime Navarro 0.60) in my study is 55 runs over Navarro’s career , which , if it is attributable to the players, means that
    Santana’s 136-84 ERA+ advantage in the study period significantly understates how much better he was.
    
    there is definitely more meat on this chicken bone that you have uncovered in baseball’s back yard; I suspect this thread is about to die . Is there any way we can follow up with each other directly if i do a little more work on this?
    
    Reply
    - 27
      
      BryanM says:
      
      September 18, 2012 at 10:50 am
      
      DId a little more work : looking at Good pitchers only 2000-2011, min 110 ERA+ ,
      1000 IP bears out overwhelmingly that strikeout pitchers give up significantly fewer unearned runs than finesse guys — there were 40 pitchers in my study , and the top 8 in SO/9 gave up 0.25 Unearned runs per 9 innings, whereas the lowest strikeout guys (still good pitchers, Glavine , Maddux, Tim Hudson, Andy Pettitte….) gave up 50 percent more, or about the same as the league as a whole. Outs are not all created equal: strikeouts are better. In doing the work I came across the amazing Curt Schilling ; The guy was just off the charts in the whole study period he gave up 12 TWELVE! unearned runs in 1569 innings — he was a little more inclined to give them up earlier in his career in Philly, but even then he was stingy.
      in a two year stretch 2005-2006 he gave up ZERO unearned runs.
      
      conclusion; unearned runs are at least partly attributable to the pitcher, RA+ would be a better indication of pitcher Value than ERA+
      
      Reply
30

Andy says:

September 18, 2012 at 7:24 pm

Great post, Doug. I wonder if the results would make more sense looking at unearned runs allowed per inning pitched instead of as a ratio of total runs allowed. You could eliminate the effect of lower number of earned runs skewing the ratio that way.

I’d also like to see the general splits of unearned runs per inning allowed at home vs on the road–that should be easy to get–as I suspect it will strongly back up your claim of home-cooking by the official scorer.

Reply
- 31
  
  Doug says:
  
  September 19, 2012 at 12:08 pm
  
  Thanks Andy,
  
  Yes, this is just something curious I noticed. I’m musing over what to do with it next, based on the feedback you and others have provided.
  
  Reply
- 32
  
  BryanM says:
  
  September 19, 2012 at 12:57 pm
  
  Andy , I totally agree – should not be a % of ER stat – If we define URA so that RA/9 =ERA + URA ,then I encountered URAs varying from .07 to .60 in the work i did for the posts above, and a lot of eyeball-quality evidence that at least some of the variation was attributable to pitcher skill. This is in addition to any home/away bias – since most pitchers pitch just a few fewer innings on the road than at home.
  If Unearned runs are partly attributable to the pitcher, then understanding the link can help with pitcher evaluation ; Doug is really onto something here
  
  Reply

High Heat Stats

Why does CC Sabathia allow so many unearned runs?

32 Comments

Leave a Reply

Recent Comments

Recent Posts

Top Commenters