2009 Team Unearned Runs

On Friday I cross-posted something I wrote for Brew Crew Ball regarding earned and unearned runs in an inning with errors and pitching changes. One of the topics that came up was team unearned runs.

The concept of team unearned runs comes from Official Rule 10.16(i).

When pitchers are changed during an inning, the relief pitcher shall not have the benefit of previous chances for outs not accepted in determining earned runs.

You can read my Friday post for an example of this rule being applied. Team unearned runs come into play if a reliever allows earned (to him) runs to score after the team should have been out of the inning. They're relatively rare, since errors, pitching changes, and ineffective relief pitching all have to present for one to result.

Team unearned runs can have an effect on baseball statistics, though. Since they are charged as earned runs to relief pitchers but unearned runs to the team, summing up earned runs allowed by a team's pitchers can make a team's ERA look higher than it should. Baseball-Reference.com calculates team ER and team ERA by summing the earned runs allowed of the team's pitchers, so that's at least one often-used site that can be off. As noted Friday, Cleveland is one example of this. Unfortunately, there are ten other teams in the same boat through June 6.

• Arizona Diamondbacks
  Baseball-Reference ER/ERA: 252/4.48 - Correct ER/ERA: 250/4.45
  
  May 25 - 8th inning - Jon Rauch charged with two earned runs (both team unearned)
  
  
• Baltimore Orioles
  Baseball-Reference ER/ERA: 293/5.37 - Correct ER/ERA: 290/5.32
  
  April 15 - 4th inning - Radhames Liz charged with three earned runs (all team unearned)
  
  
• Chicago Cubs
  Baseball-Reference ER/ERA: 211/4.05 - Correct ER/ERA: 210/4.04
  
  June 2 - 8th inning - Carlos Marmol charged with one earned run (team unearned)
  
  
• Cleveland Indians
  Baseball-Reference ER/ERA: 298/5.27 - Correct ER/ERA: 297/5.25
  
  May 12 - 7th inning - Tony Sipp charged with one earned run (team unearned)
  
  
• Colorado Rockies
  Baseball-Reference ER/ERA: 252/4.71 - Correct ER/ERA: 250/4.67
  
  April 18 - 7th inning - Ryan Speier charged with one earned run (team unearned)
  May 17 - 7th inning - Manuel Corpas charged with one earned run (team unearned)
  
  
• Detroit Tigers
  Baseball-Reference ER/ERA: 217/4.09 - Correct ER/ERA: 209/3.94
  
  April 7 - 8th inning - Brandon Lyon charged with three earned runs (two team unearned)
  April 28 - 7th inning - Nate Robertson charged with four earned runs (all team unearned); Brandon Lyon charged with two earned runs (both team unearned)
  
  
• Florida Marlins
  Baseball-Reference ER/ERA: 259/4.51 - Correct ER/ERA: 257/4.47
  
  May 19 - 7th inning - Renyel Pinto charged with one earned run (team unearned)
  June 3 - 5th inning - Hayden Penn charged with one earned run (team unearned)
  
  
• Milwaukee Brewers
  Baseball-Reference ER/ERA: 223/4.07 - Correct ER/ERA: 222/4.05
  
  May 18 - 7th inning - Mark DiFelice charged with one earned run (team unearned)
  
  
• Oakland Athletics
  Baseball-Reference ER/ERA: 235/4.30 - Correct ER/ERA: 234/4.28
  
  April 28 - 6th inning - Andrew Bailey charged with one earned run (team unearned)
  
  This is further notable because of an error in the Baseball-Reference box score/play-by-play. Bailey entered the game due to injury with a 2-0 count on Michael Young. The box score/play-by-play lists Young's walk (and later run) as belonging to Bailey when in fact it should belong to Anderson per Official Rule 10.16(h). Curiously, the walk and run do not count against his total on his player page but they are still there in his game log.
  
  
• San Diego Padres
  Baseball-Reference ER/ERA: 245/4.49 - Correct ER/ERA: 244/4.47
  
  April 29 - 6th inning - Arturo Lopez charged with one earned run (team unearned)
  
  
• Tampa Bay Rays
  Baseball-Reference ER/ERA: 251/4.49 - Correct ER/ERA: 247/4.42
  
  May 25 - 9th inning - Grant Balfour charged with two unearned runs (both team unearned); Jason Isringhausen charged with two unearned runs (both team unearned)

Obviously the difference will mean less as the season goes on. Where one team unearned run might cause a difference of 0.02 in a team's ERA now, it might not cause an apparent difference in a team's ERA at the end of the season. However, that's no reason to ignore the fact there is a difference. Who knows, maybe those eight team unearned runs will be the difference between Detroit finishing first, second, or third in the team ERA standings.

A Few Innings Short

MLB Official Rule 10.22(b) covers the minimum innings pitched required to determine each league's ERA champion:

The individual pitching champion in a Major League shall be the pitcher with the lowest earned-run average, provided that the pitcher has pitched at least as many innings in league championship games as the number of games scheduled for each club in his club's league that season.

That means a pitcher needs to throw 162 or more innings to count himself among the league leaders. Thus the ERA title can today only go to starting pitchers who stay healthy. Plenty of guys stay on the mound enough to qualify for the ERA title each year: since 2004, the number of pitchers with 162+ IP has ranged between 77 and 92. But what about guys who just missed the cutoff, whether through injury, delayed callups, or any other reason? Surely there have been some guys who missed the cut by just a few innings. Of course there's no guarantee those guys wouldn't have been lit up in the innings they required, but let's take a look at them anyway:

• Mark Eichhorn (1986) - 1.72 ERA in 157 innings for Toronto
  
  Eichhorn made his major league debut in late 1982, but a shoulder injury doomed his career as a starter and forced him into a submarine-style delivery. In 1986, he appeared in 69 games for the Blue Jays, pitching a stunning 157 innings in relief. He wound up 14-6 with 10 saves and 166 strikeouts. Had he picked up five more innings during the season, his ERA would have easily beaten AL leader Roger Clemens' 2.48.
  
  
• John Denny (1984) - 2.45 ERA in 154 1/3 innings for Philadelphia
  
  One year after taking home the NL Cy Young Award and Comeback Player of the Year Award, Denny came within one start of taking home the NL ERA title. A midseason injury cost him the chance as he missed two months from late May to July. Denny wound up 7-7 over the 22 starts he was able to make in 1984. Alejandro Pena's 2.48 ERA in 199 1/3 innings took the title instead.


• Tom Hall (1970) - 2.55 ERA in 155 1/3 innings for Minnesota
  
  The 1970 AL ERA belongs to Diego Segui for putting up a 2.56 ERA in exactly 162 innings. Segui was able to just barely sneak onto the ERA leaderboard, but Twins swingman Tom Hall wasn't quite as lucky. Appearing in 52 games (11 starts) and putting up an 11-6 record with four saves, Hall wound up 6 2/3 innings short. If he had been allowed to start a couple more games before September, he might well have taken Segui's title.
  
  
• Roger Craig (1959) - 2.06 ERA in 152 2/3 innings for Los Angeles
  
  Called up in June, Craig wound up four outs away from the ERA title (the 154-game schedule was still being used). Utilized as a swingman, Craig started 17 games and appeared 12 times in relief. He wound up 11-5 and even picked up some MVP votes. Much like Mark Eichhorn thirty years later, he easily outpaced the competition in the ERA race: Sam Jones wound up edging Stu Miller and Bob Buhl for the title with a 2.83 ERA.
  
  
• Jim Konstanty (1950) - 2.66 ERA in 152 innings for Philadelphia
  
  Sal Maglie took the crown with a 2.71 ERA in 206 innings, but Jim Konstanty was right there with him. The relief ace for the Whiz Kids appeared in 74 games, finishing 62 of them, going 16-7 with 22 saves (by modern definition). He was a surprise pick to start Game 1 of that year's World Series and lost 1-0. Konstanty picked up the MVP award for his season, but was two innings away from claiming the ERA title as well.

Konstanty was the first pitcher since 1920 to get close to the title without qualifying. I didn't look beyond 1920 because of the varying schedule length from year to year and drastically different style of pitcher use. I'm also not sure if the ERA title was determined in the same way back then. Did I miss someone in the last 90 years? Let me know.

Major League Debuts as a Starting Pitcher

Hey, look at that, it's been a week again since I last posted. Oops. During the past week, major league active rosters expanded from a limit of 25 to a limit of 40. There are a few different general types of September call-ups: major league veterans who have spent the season at AAA, fringe major league players who get called up to give the big club more flexibility in games, guys who have no business being on a major league roster except for being great at one of the five tools (usually speed), or young prospects.

In the first three days of September, twenty-two call-ups made their major league debuts. Take a gander at the list:

• Matt Antonelli, 2B, SDP
• Wilkin Castillo, C, CIN
• Phil Coke, LHP, NYY
  
• Luis Cruz, SS, PIT
• Alcides Escobar, SS, MIL
  
• Dexter Fowler, CF, COL
• Mat Gamel, 3B, MIL
• Greg Golson, CF, PHI
  
• Mike Hinckley, LHP, WSN
• Wade LeBlanc, LHP, SDP
  
• Casey McGehee, 3B, CHC
• Jim Miller, RHP, BAL
• Jason Motte, RHP, STL
  
• Brad Nelson, 1B, MIL
• Jonathon Niese, LHP, NYM
• Josh Outman, LHP, OAK
• Jason Pridie, CF, MIN
  
• Rich Rundles, LHP, CLE
• Angel Salome, C, MIL
  
• Justin Thomas, LHP, SEA
  
• Luis Valbuena, 2B, SEA
• Josh Whitesell, 1B, ARI

I think Outman is a good last name for a pitcher. Anyway, as minor league playoffs end and a second, smaller wave of call-ups makes it to the big leagues, this list will grow a little more. I really want to focus on two debutants, though: Wade LeBlanc and Jonathon Niese. Those two recently became part of a large group that's still rather small in the grand scheme of baseball history. I'm talking about pitchers making their major league debut by starting a game.

Since 1956, 4439 different players have pitched in a major league game. Of those, 1270 (28.6%) debuted as a starting pitcher. So far in 2008, thirty-four different pitchers have made their debut as a starter. Usually, unless the pitcher in question is a mega-prospect known around baseball, fans of a team going up against one of these rookie pitchers figure the guy should get roughed up in his first start. After all, the thinking goes, he might have pitched well in AAA or AA, but he's in the majors.

It turns out that guys making their debut do tend to get roughed up more than the average pitcher making a start. This makes sense when you consider the guys who get called up, get shelled, and get sent back to the minors shortly after. This is also probably most of the reason why fans figure almost any debuting pitcher should get lit up on the mound.

I've put together a chart showing the combined ERA of all starters making their debut in a given season against the ERA of all major league starters in that season.


(click image to enlarge in a new window)

Unsurprisingly, most of the time the rookies allow more earned runs. Interestingly, however, there are a few seasons where they actually had a lower ERA than the average starter. The most recent time that happened was in 2005 when guys like Paul Maholm, Kyle Davies, Dustin McGowan, and J.P. Howell had great debuts. Sure there were clunkers, too, like Sean Henn and Ervin Santana's first games, but they were more than balanced by good games that year.

The prior season was the worst since 1956 for debuting starters. Thirty-four starters combined to put up an awful 7.68 ERA and an abysmal .990 OPS against. There were eleven starts with a game score of 30 or below, including one of -7 (belonging to Arnie Munoz). Ten guys couldn't even get through the fourth inning. It wasn't all bad: Chris Saenz, Daniel Cabrera, and Tyler Yates (great combo) each went six scoreless while Scott Kazmir, Chris Young, and Zack Greinke, among others, performed capably. All in all, however, it wasn't a great year for the newbies.

For fun, I decided to look at the numbers since 1998, the final round of expansion. In the past almost eleven seasons, 387 different pitchers have made their debut as a starter. In their debuts, they combined for the following pitching line:
1991.7 IP, 2093 H, 1265 R, 1175 ER, 934 BB, 1368 K, 309 HR, Average Game Score: 46

In the more familiar rate stats, that all turns out to be:
5.31 ERA, 1.52 WHIP, 6.18 K/9, 4.22 BB/9, 1.40 HR/9, .271/.354/.464 opponent batting line

It turns out there are a couple individual pitchers who put up similar numbers during the period between 1998 and 2008. The closest comparisons to all the rookies mashed together are these two:

• Rocky Biddle, 378.2 IP between 2000 and 2004:
  5.47 ERA, 1.53 WHIP, 6.20 K/9, 4.04 BB/9, 1.40 HR/9, .274/.357/.462
• Victor Santos, 602.1 IP between 2001 and 2007:
  5.21 ERA, 1.58 WHIP, 6.35 K/9, 4.17 BB/9, 1.29 HR/9, .284/.362/.467
  

Rocky Biddle is spot on. Besides Biddle and Santos, Jason Hammel and Paul Rigdon are also kind of close, Allen Levrault was a few strikeouts away from the group, and if you toss out Bill Pulsipher's first season, he's in the mix, too. Or, if you're a normal baseball fan: one current fringe guy, two forgotten guys and an ex-top prospect join the three listed above.

If you're curious about the actual year-to-year numbers or just want to see how your favorite starter(s) debuted, check out the Google spreadsheet I made listing all 1270 debuts as a starting pitcher since 1956.

EDIT: Of course, I post this and James Parr of the Braves debuts the next night with six scoreless innings against the Nationals. The Nats countered with rookie Shairon Martis, who went five innings while allowing two runs.

Playoff Rotations, 2003-2007

Anyone who's been a regular reader here knows how I've already looked at rotation spots throughout MLB in terms of FIP and ERA. The numbers for each league in 2007 and data for 2003-2007 using both statistics can be found here. The methodology is also explained there.

I've noticed some people disparaging the notion of a pitcher putting up league average numbers having much value. After all, the reasoning goes, league average includes all the bad teams whereas the average rotation to make the playoffs would be more telling (because no mediocre teams make the playoffs). Snark aside, there's a certain logic in that line of thinking. Making the playoffs with a below average rotation is possible (see: 2007 Phillies), but rare.

With that in mind, I decided to crunch the numbers and find out the rotation numbers of playoff teams over the past five seasons. Treating all four teams each year as one giant staff starting approximately 648 games, I then figured out the FIP and ERA for each rotation spot. Below are the results compared to the relevant league average for each season. The numbers (#1, #2, etc.) correspond to each spot in the rotation.

FIP: AL Playoff Teams vs. AL Average, 2003-2007

Year	Team	FIP	#1	#2	#3	#4	#5
2003	AL Playoff Teams	4.10	3.18	3.77	4.23	4.43	5.15
	AL Average	4.73	3.57	4.30	4.78	5.17	6.30
2004	AL Playoff Teams	4.38	3.42	3.93	4.34	4.95	5.55
	AL Average	4.80	3.75	4.40	4.77	5.19	6.11
2005	AL Playoff Teams	4.33	3.60	4.05	4.30	4.56	5.43
	AL Average	4.56	3.65	4.18	4.56	4.92	5.77
2006	AL Playoff Teams	4.49	3.31	4.19	4.47	4.81	6.03
	AL Average	4.66	3.55	4.18	4.63	5.08	6.31
2007	AL Playoff Teams	4.18	3.22	3.80	4.12	4.50	5.67
	AL Average	4.47	3.47	4.01	4.45	4.88	6.02
Last 5	AL Playoff Teams	4.29	3.31	3.96	4.28	4.61	5.63
	AL Average	4.64	3.60	4.21	4.64	5.05	6.13

Only in one spot in on year did the league out-perform the playoff teams. In 2006, the league average #2 starter had an FIP of 4.18 compared to the playoff teams' 4.19. Otherwise the playoff squads handily outpaced the rest of the league. The average playoff team's starter FIP was between 92.6% that of the league's. Note, however, that a league average starter (by FIP) still makes a good #4 pitcher for a playoff team based on the last five seasons.

ERA: AL Playoff Teams vs. AL Average, 2003-2007

Year	Team	ERA	#1	#2	#3	#4	#5
2003	AL Playoff Teams	4.15	2.66	3.66	4.30	4.61	6.04
	AL Average	4.66	3.18	4.05	4.56	5.37	6.97
2004	AL Playoff Teams	4.47	3.22	4.03	4.58	4.99	6.27
	AL Average	4.83	3.57	4.35	4.84	5.22	6.93
2005	AL Playoff Teams	4.15	3.21	3.59	4.00	4.52	6.17
	AL Average	4.52	3.21	3.85	4.40	5.07	6.73
2006	AL Playoff Teams	4.38	3.16	3.80	4.19	4.81	6.58
	AL Average	4.73	3.36	4.17	4.62	5.10	7.06
2007	AL Playoff Teams	4.29	3.07	3.70	4.26	4.73	6.27
	AL Average	4.61	3.29	3.88	4.42	5.23	7.00
Last 5	AL Playoff Teams	4.29	3.02	3.77	4.26	4.74	6.29
	AL Average	4.67	3.32	4.06	4.56	5.20	6.93

Again the playoff teams sink to league average in only one spot: #1 pitchers on playoff teams in 2005 were league average. Of course, when the rest of your rotation outpitches everyone else, you can afford that. The average playoff team ERA was 91.8% that of the league, or about the same as the difference in FIP. Not surprisingly, then, a league average pitcher slots in at #4 on a playoff team using ERA again.

So now that it turns out a league average starter is generally a good #4 starter on a playoff team in the American League, let's turn to the senior circuit. Pitching in the NL is considered to have a positive effect on a pitcher's statistics since opposing pitchers bat for themselves and the NL FIP has been below that of the AL in four of the past five seasons (2007 is the exception). Similarly, the NL ERA was higher than the AL ERA only in 2007. Perhaps, then, this skews the importance of league average starters to NL playoff teams.

FIP: NL Playoff Teams vs. NL Average, 2003-2007

Year	Team	FIP	#1	#2	#3	#4	#5
2003	NL Playoff Teams	4.13	3.08	3.82	4.14	4.55	5.42
	NL Average	4.55	3.42	4.13	4.52	4.95	6.15
2004	NL Playoff Teams	4.50	3.38	4.08	4.47	4.92	6.17
	NL Average	4.61	3.37	4.14	4.63	5.09	6.47
2005	NL Playoff Teams	4.20	3.11	3.68	4.36	4.72	5.48
	NL Average	4.45	3.35	4.05	4.43	4.95	5.85
2006	NL Playoff Teams	4.65	3.56	4.16	4.72	5.12	5.97
	NL Average	4.66	3.62	4.21	4.63	5.07	6.25
2007	NL Playoff Teams	4.66	3.64	4.34	4.70	5.01	5.90
	NL Average	4.60	3.48	4.17	4.66	5.04	6.04
Last 5	NL Playoff Teams	4.42	3.28	4.03	4.46	4.89	5.83
	NL Average	4.57	3.45	4.14	4.57	5.02	6.15

The past two seasons have been interesting in the National League. The general mediocrity of the starting staffs of the Cardinals and Mets in 2006 contributed the most to the #3 and #4 spots on playoff teams doing worse than league average (the Cardinals #3 starters had a 5.31 FIP!). Of course, each of those teams had better than average results from their fifth starters which no doubt helped them win enough games to get into the playoffs.

Last season is a different story. The Rockies (4.67) and Phillies (4.86) both had below average starting staffs overall. The Diamondbacks (4.58) and Cubs (4.54) were above average, but not by much. Cumulatively, this means the average playoff rotation was worse than league average overall. Only in the fourth and fifth starter spots did the playoff teams manage to outperform the league. There might be a moral in there (depth! depth! depth!). In any event, the National League in the past two seasons has shown that if you can run a half-decent staff out there and mash the ball, you've got a real shot.

The average NL playoff team rotation over the past five years put up an FIP that was 96.8% that of the league. Again, a league average starter slots in as a #4 overall, but in the past two seasons, playoff teams could use such a pitcher as a #3 guy. Maybe ERA will clear up the picture: perhaps these FIP-challenged teams were ERA-lucky.

ERA: NL Playoff Teams vs. NL Average, 2003-2007

Year	Team	ERA	#1	#2	#3	#4	#5
2003	NL Playoff Teams	3.90	2.75	3.45	3.90	4.24	5.66
	NL Average	4.40	2.97	3.81	4.24	4.92	7.04
2004	NL Playoff Teams	4.12	3.03	3.61	4.04	4.35	6.76
	NL Average	4.43	3.00	3.75	4.22	4.87	7.55
2005	NL Playoff Teams	3.79	2.42	3.09	3.78	4.45	5.66
	NL Average	4.23	2.81	3.70	4.23	4.70	6.23
2006	NL Playoff Teams	4.46	3.31	3.74	4.21	4.90	6.17
	NL Average	4.65	3.25	4.00	4.57	5.12	7.07
2007	NL Playoff Teams	4.47	3.43	4.00	4.30	4.78	6.51
	NL Average	4.64	3.29	3.97	4.50	5.12	7.03
Last 5	NL Playoff Teams	4.14	2.86	3.63	4.07	4.54	6.29
	NL Average	4.47	3.07	3.85	4.35	4.95	6.98

ERA makes recent playoff teams look a little better. Only the #1 and #2 (barely) starters last season were substandard, while only 2006 aces were below average. As long as most of your rotation outperforms their equivalents around the league, you can deal with a slightly below average top dog. The #1 pitchers in 2004 didn't get lucky in this table either, still finishing a little below the league average.

The playoff team ERA is 92.7% that of the league average over the past five seasons. That's close to the American League's numbers for FIP and ERA. League average starters are, yet again, good #4 starters on playoff teams.

I can see some people complaining that numbers like FIP and ERA don't get you into the playoffs: it's wins that count! With that in mind, I averaged the wins by each starting staff in each league in each season. In the American League, playoff teams could expect about 70 wins from their rotation each year. In the National League, that number drops to 65 wins. Either way, I think the moral of the story is it takes good defense combined with great offense to overcome a league average pitching staff and make the playoffs, but it's certainly possible.

Revisiting Rotation Spots

Earlier this offseason I had a few posts talking about the average ERA and FIP for each rotation spot in 2007. Now, one season isn't a whole lot of context in baseball: depending on what you're looking at, there can be a lot of fluctuation from year to year. That's why I thought it would be interesting to look at FIP and ERA by rotation spot over the past five seasons.

The methodology is the same as in the previous posts. A team's starts are broken up into quintiles, based on the FIP or ERA of their pitchers, and a weighted average is taken in each quintile based on the starts and FIP of each pitcher in it. For example, a team with a normal 162-game schedule would have their rotation spots looking like: #1 - 33 starts, #2 - 33 starts, #3 - 32 starts, #4 - 32 starts, #5 - 32 starts. If Pitcher A made 20 starts with a 3.30 ERA and Pitcher B made 15 starts with a 3.40 ERA and everyone else on the staff had higher numbers, Pitcher A's 20 starts would be combined with 13 of Pitcher B's starts to yield an average ERA of 3.34 in the top 33 starts. The remaining two starts from Pitcher B would be combined with Pitchers C, D, etc., to figure out the average ERA of the other rotation spots.

The numbers below in the tables are found by combining all the starts in the league for each season (~2490 for the NL and ~2268 for the AL), separating them into fifths, and using the process described above. The "Average" row is the average of the five numbers above. Also remember that rather than adjusting the constant added to the FIP number every season like The Hardball Times does, I just use 3.2 across the board.

NL Rotations by FIP, 2003-2007

Year	FIP	#1	#2	#3	#4	#5
2003	4.55	3.42	4.13	4.52	4.95	6.15
2004	4.61	3.37	4.14	4.63	5.09	6.47
2005	4.45	3.35	4.05	4.43	4.95	5.85
2006	4.66	3.62	4.21	4.63	5.07	6.25
2007	4.60	3.48	4.17	4.66	5.04	6.04
Average	4.57	3.45	4.14	4.57	5.02	6.15

NL Rotations by ERA, 2003-2007

Year	ERA	#1	#2	#3	#4	#5
2003	4.40	2.97	3.81	4.24	4.92	7.04
2004	4.43	3.00	3.75	4.22	4.87	7.55
2005	4.23	2.81	3.70	4.23	4.70	6.23
2006	4.65	3.25	4.00	4.57	5.12	7.07
2007	4.64	3.29	3.97	4.50	5.12	7.03
Average	4.47	3.07	3.85	4.35	4.95	6.98

Note: NL defenses must have taken some sort of hit in the past two seasons to have ERA's jump the way they did without FIP being similarly affected. 2005 is a kind of fluky season, but the FIP from before then are similar to 2006-2007 and yet ERAs are lower. Maybe it's an injury issue, as well.

AL Rotations by FIP, 2003-2007

Year	FIP	#1	#2	#3	#4	#5
2003	4.73	3.57	4.30	4.78	5.17	6.30
2004	4.80	3.75	4.40	4.77	5.19	6.27
2005	4.56	3.65	4.18	4.56	4.92	5.77
2006	4.66	3.55	4.18	4.63	5.08	6.31
2007	4.47	3.47	4.01	4.45	4.88	6.02
Average	4.64	3.60	4.21	4.64	5.05	6.13

AL Rotations by ERA, 2003-2007

Year	ERA	#1	#2	#3	#4	#5
2003	4.66	3.18	4.05	4.56	5.37	6.97
2004	4.83	3.57	4.35	4.84	5.22	6.93
2005	4.52	3.21	3.85	4.40	5.07	6.73
2006	4.73	3.36	4.17	4.62	5.10	7.06
2007	4.61	3.29	3.88	4.42	5.23	7.00
Average	4.67	3.32	4.06	4.56	5.20	6.93

Now for the fun part: finding an average team rotation (using ERA) over the last five seasons. Players included must have started 60% of their games and had 500+ innings pitched in their league over the last five seasons.

First, the National League:

1. Average #1, 3.31 ERA: Carlos Zambrano (3.30 ERA)
2. Average #2, 3.91 ERA: Tom Glavine (3.97 ERA)
3. Average #3, 4.40 ERA: Chris Capuano (4.39 ERA)
4. Average #4, 4.96 ERA: Claudio Vargas (4.96 ERA)
5. Average #5, 6.68 ERA: Eric Milton (5.54 ERA) - he was the worst qualifier under my criteria. You'd have to go down to a 100 IP minimum to find guys who actually put up a 6.68 ERA in the NL, which tells you how often guys are shuffled in and out of the #5 spot due to injuries.

Second, the American League:

1. Average #1, 3.57 ERA: C.C. Sabathia (3.61 ERA)
2. Average #2, 4.19 ERA: Kenny Rogers (4.18 ERA)
3. Average #3, 4.61 ERA: Jose Contreras (4.57 ERA)
4. Average #4, 5.13 ERA: Joel Piniero (5.05 ERA)
5. Average #5, 6.66 ERA: Sidney Ponson (5.37 ERA) is the worst qualifier. You again have to go down to a minimum of 100 IP to find guys at 6.66 or worse.
   

It's Better To Be Lucky Than Good

In my last post I looked at pitchers who were exceedingly unlucky. The guys I listed all had an ERA+ of 120 or above but won one-third or fewer of their (at least 15) decisions. Today I want to look at pitchers who did the opposite: had a low ERA+ but won two-thirds or more of their decisions. The criteria again are 15+ decisions, a winning percentage of .667 or above, and 50% or more of appearances as a starter.

Lowest ERA+, .667 or above W-L%, min. 15 decisions, 1901-2007

Name	Year	ERA	ERA+	W-L Record	Run Support*	League Runs Per Game**
Santo Alcala	1976	4.70	75	11-4	6.34	3.98
Rip Collins	1921	5.44	78	11-5	6.20	5.12
Mike Lynch	1905	3.79	79	17-8	4.46	4.11
Ed Wells	1930	5.20	83	12-3	6.90	5.41
Jim Coates	1960	4.28	83	13-3	7.81	4.39
Rube Benton	1916	2.87	85	16-8	3.85	3.45
Storm Davis	1989	4.36	85	19-7	5.81	4.29
James Baldwin	1998	5.32	86	13-6	5.19	5.01
Guy Bush	1931	4.49	86	16-8	5.31	4.48
Tommie Sisk	1966	4.14	86	10-5	5.78	4.09
Chuck Dobson	1971	3.81	88	15-5	5.96	3.87
Roxie Lawson	1937	5.26	89	18-7	6.03	5.23
Johnny Allen	1933	4.39	89	15-7	6.10	5.00
Lefty Williams	1917	2.97	89	17-8	4.21	3.65
Ralph Works	1911	3.87	90	11-5	5.40	4.61
Hooks Dauss	1919	3.55	90	21-9	4.41	4.09
Jack Coombs	1911	3.53	90	28-12	5.66	4.61
Kevin Tapani	1998	4.85	90	19-9	5.37	4.60
Adam Eaton	2005	4.27	90	11-5	5.49	4.45
Rube Foster	1916	3.06	90	14-7	3.53	3.68

*Runs scored by the pitcher's team's offense in games in which he pitched; if prior to 1957, the team's runs per game. If more than one team, a weighted average based on innings pitched.
**League refers to the league (American/National) the pitcher was in. Any pitcher in both leagues gets a weighted average of the two league values based on innings pitched.

These were the only twenty pitchers who had an ERA+ of 90 or less. The complete list of all 88 seasons with an ERA+ of 100 (league average) or below can be found here.

Anything for a Win

In 1987, Nolan Ryan placed fifth in the NL Cy Young voting while pitching for the Houston Astros. He had an impressive 270 strikeouts in 211 2/3 innings and won the league ERA title by 27 points (2.76 vs. Mike Dunne's 3.03). Unfortunately for Ryan, the Astros as a team scored 648 runs that season, or 4 per game. When he took the mound, he was only received 3.28 runs of support from his offense -- the league average runs scored by a team per game was 4.52. Despite his many strikeouts and ERA title, Ryan only went 8-16 on the year. Imagine being possibly the most dominant pitcher in the league and only winning one-third of your decisions. Sometimes life just isn't fair. I should note that Orel Hershiser, who finished fourth in Cy Young voting could complain as well - he went 16-16 with a 3.06 ERA and 190 strikeouts for the hitting-challenged Dodgers.

With Ryan's season in mind, I want to find starting pitchers (i.e., 50+% of appearances were starts) who pitched well but had no luck in gathering wins. To do this, I'm going to use ERA+, defined as "the ratio of the league's ERA (adjusted to the pitcher's ballpark) to that of the pitcher. -- found by lgERA/ERA" Thus since Nolan Ryan played in a ballpark that favored pitchers in 1987, the league ERA used to derive ERA+ will be slightly higher than the actual average to reflect the slightly easier conditions (the exact process behind park adjustments is described here). Since Ryan's 2.76 was still much better than the adjusted league ERA, his ERA+ is an excellent 142. Using ERA+ also allows you to make judgments across time since it compares pitchers to others in their same league in the same year. For example, in 1918 Allen Sothoron put up a sterling 1.94 ERA in 209 innings. Though Sothoron's ERA looks much better than Ryan's, using ERA+ we see they were about equal in relation to the league they played in (Ryan's ERA+ was 142, Sothoron's was 141).

The second part of finding these unlucky guys is deciding which winning percentage to look under. Looking for guys with a losing record makes sense, but first I want to see guys who were really unlucky. To that effect, I'm going to look for pitchers who had a .333 or below winning percentage. Finally, a guy that goes 1-2 in 5 excellent starts is certainly unlucky, but I want to find pitchers who were unlucky for most of a season. Thus I'm going set the minimum for decisions at fifteen. It would be hard to start games all season and not get fifteen decisions somehow: only 35 pitchers since 1901 have started 25 or more games in a year and wound up with fewer than fifteen decisions. The only pitcher who might be excluded by this is John Dopson: he went 3-11 with a 118+ ERA for Montreal in 1988.

Highest ERA+, .333 or below W-L%, min. 15 decisions, 1901-2007

Name	Year	ERA	ERA+	W-L Record	Run Support*	League Runs Per Game**
Ned Garvin	1904	1.72	158	5-16	3.27	3.89
Jim Abbott	1992	2.77	144	7-15	2.48	4.32
Nolan Ryan	1987	2.76	142	8-16	3.28	4.52
Harry Brecheen	1953	3.07	137	5-13	3.60	4.46
Dummy Taylor	1902	1.72	132	5-16	3.19	4.11
Brandon Webb	2004	3.59	129	7-16	4.08	4.64
George Mogridge	1916	2.31	125	6-12	3.70	3.68
Turk Farrell	1962	3.02	124	10-20	2.67	4.48
Buster Brown	1910	2.67	123	9-23	3.15	4.03
Matt Cain	2007	3.65	122	7-16	3.20	4.71
Dennis Lamp	1978	3.30	122	7-15	3.65	3.99
Howard Ehmke	1925	3.73	121	9-20	4.20	5.20
Ed Durham	1932	3.80	120	6-13	3.68	5.23
Bill Piercy	1923	3.41	120	8-17	3.79	4.78
Erv Kantlehner	1915	2.26	120	5-12	3.57	3.62
Jumbo Elliott	1927	3.30	120	6-13	3.51	4.58
Eddie Smith	1937	3.94	120	4-17	4.54	5.23

*Runs scored by the pitcher's team's offense in games in which he pitched; if prior to 1957, the team's runs per game. If more than one team, a weighted average based on innings pitched.
**League refers to the league (American/National) the pitcher was in. Any pitcher in both leagues gets a weighted average of the two league values based on innings pitched.

Poor Matt Cain, stuck on a team that couldn't score for him. Of course, it's perhaps harder to feel for him with his contract that makes him set for life than it is to feel for a guy like Erv Kantlehner who was out of baseball at the age of 24 and who probably didn't get paid real well when he was in the game. But from a strictly baseball sense: poor Matt Cain. I always figured Nolan Ryan was the best recent example of good pitching wrecked by anemic offense, but Jim Abbott's 1992 is even better: Abbott's Angels even got a designated hitter to help spur offense. I cut the chart off at 120 simply for brevity's sake, but you can find the list of all 126 better than average losers meeting my criteria here.

A final note about George Mogridge and Kantlehner: as noted, their "Run Support" number reflects their team's average runs per game. They very well might (and probably did) have had even fewer runs scored behind them.

AL Rotations by ERA for 2007

EDIT: Upon further review, I don't like the way I originally came about these numbers. For some reason, I had simply averaged each team's FIP in each spot for the league numbers. A smarter way to do it would be to treat the entire league as one giant rotation and determine the top 20% of starts (generally around 518 for the NL and 454 for the AL) for spot #1, etc. It makes no sense to penalize the league average rotation because the top 15 pitchers are clustered on, say, 8 teams. This tends to lower the value for #1 starters and raise the value for #5 starters while leaving the middle guys generally unchanged. I've changed the numbers in the tables to reflect the new method. The "average" rotations at the end are the same as before. Sorry for the mistake.

This post is similar to the last one using FIP. Using the ERA for each team's starters in games they started, I took the top 33 starts to determine the ERA for the #1 spot in the rotation, then the next 33 for spot #2, the next 32 for spot #3, the next 32 for spot #4, and the final 32 to find the #5 spot ERA for each team and then averaged those across all teams to get the league numbers. The STDEV column again gives an idea of how "even" a team's rotation was across all spots.

2007 AL Rotations and Rotation Spots by ERA

Team	SERA	#1	#2	#3	#4	#5	STDEV
Cleveland Indians	4.19	3.06	3.21	4.37	4.58	6.25	1.29
Boston Red Sox	4.21	3.14	3.88	4.42	4.75	5.15	0.78
Los Angeles Angels	4.22	3.00	3.42	3.97	4.98	6.19	1.29
Toronto Blue Jays	4.24	3.72	3.77	3.88	4.14	6.42	1.15
Oakland Athletics	4.29	2.98	3.77	4.11	4.52	7.24	1.62
Minnesota Twins	4.33	3.33	4.00	4.24	4.69	5.62	0.85
Chicago White Sox	4.47	3.51	3.77	4.28	5.38	5.77	0.99
New York Yankees	4.57	3.52	4.07	4.19	5.00	7.11	1.41
Detroit Tigers	4.68	3.47	4.57	4.82	4.99	5.93	0.88
Baltimore Orioles	4.86	3.20	3.65	4.91	5.60	9.04	2.31
Kansas City Royals	4.87	3.65	3.82	4.53	5.60	7.88	1.73
Seattle Mariners	5.16	3.96	4.31	4.45	5.99	7.95	1.66
Tampa Bay Devil Rays	5.20	3.48	3.90	5.78	5.96	7.75	1.73
Texas Rangers	5.50	4.45	5.11	5.33	5.70	7.06	0.97
AL	4.61	3.29	3.88	4.42	5.23	7.00	1.44

Despite the traditional thought that moving to the AL entails an increase in ERA, American League starters managed to post an ERA that was 0.03 lower than their NL counterparts (click here for the NL data in a new window). The American League total ERA was a little bit higher, but that was because AL relievers weren't quite as good as NL relievers, at least in terms of ERA. Interestingly, the #1 starters in each league came out to a 3.46 ERA. The #3 starters' ERA was also the same and the other three spots were very close.

Now to give the numbers context by seeing which pitchers fit in each spot:

AL #1 Starters (~3.46 ERA or below), 20+ starts

• John Lackey, 33 starts, 3.01 ERA
• Fausto Carmona, 32, 3.06
• Dan Haren, 34, 3.07
• Erik Bedard, 28, 3.16
• C.C. Sabathia, 34, 3.21
• Josh Beckett, 30, 3.27
• Johan Santana, 33, 3.33
• Jeremy Guthrie, 26, 3.44
• Scott Kazmir, 34, 3.48 (it didn't feel right to exclude him since the next 20+ game starter had a 3.63 ERA)

AL #2 Starters (~3.95 ERA), 20+ starts

• Brian Bannister, 27 starts, 3.87 ERA
  
• Curt Schilling, 24, 3.87
• Shaun Marcum, 25, 3.91
• Jered Weaver, 28, 3.91
• Felix Hernandez, 30, 3.92
• Joe Blanton, 34, 3.95
• Dustin McGowan, 27, 4.08
• Andy Pettitte, 34, 4.09

AL #3 Starters (~4.52 ERA), 20+ starts

• Daisuke Matsuzaka, 32 starts, 4.40 ERA
• Chad Gaudin, 34, 4.42
• Steve Trachsel, 25, 4.48
• Paul Byrd, 31, 4.59

AL #4 Starters (~5.14 ERA), 20+ starts

• Jeremy Bonderman, 28 starts, 5.01 ERA
• Mike Mussina, 27, 5.16
• Kevin Millwood, 31, 5.16
• Julian Tavarez, 23, 5.22

AL #5 Starters (~6.81 ERA), 15+ starts

• Cliff Lee, 16 starts, 6.38 ERA
• Bartolo Colon, 18, 6.41
• Robinson Tejeda, 19, 6.61
• Horacio Ramirez, 20, 7.16

Not really any surprises, though it is worth noting the Orioles actually had two ace-level pitchers and a #3 pitcher, but not much else (see their #5 ERA above). The Mike Mussina-Chris Capuano analogy from yesterday seems pretty apt, since Capuano was also a #2 in FIP but fell to #4 in ERA.

I'll finish this up, as I did yesterday, with a recap of the "league average" rotation:

1. Jeremy Guthrie/Scott Kazmir
2. Joe Blanton
3. Steve Trachsel
4. Mike Mussina/Kevin Millwood
5. Robinson Tejeda
   

NL Rotations by ERA for 2007

EDIT: Upon further review, I don't like the way I originally came about these numbers. For some reason, I had simply averaged each team's FIP in each spot for the league numbers. A smarter way to do it would be to treat the entire league as one giant rotation and determine the top 20% of starts (generally around 518 for the NL and 454 for the AL) for spot #1, etc. It makes no sense to penalize the league average rotation because the top 15 pitchers are clustered on, say, 8 teams. This tends to lower the value for #1 starters and raise the value for #5 starters while leaving the middle guys generally unchanged. I've changed the numbers in the tables to reflect the new method. The "average" rotations at the end are the same as before. Sorry for the mistake.

A few days ago I had a post detailing rotations and average starters in the National League using FIP. Today I want to change gears, take off the beginner-sabermetrician's hat, and return to the commonly cited ERA. I admit, when I hear the phrase, "he's a #4 starter," I quantify that in terms of ERA and I would guess most of you do, too. ERA was the statistic used in the article that started me thinking about this "analysis" and it's common enough that I figured it would be worth a look today. I think the most interesting thing to take away from this exercise is the difference between perception of back of the rotation starters and their actual numbers.

As I did for the FIP table, each rotation spot is figured out the same way for each team (33 starts for #1, then 33, 32, 32, 32). Obviously I use ERA instead of FIP so for the Brewers, the #1 spot calculation looks like:

(6*2.06+17*3.74+10*3.82)/33 = 3.46 -- Carlos Villanueva, Yovani Gallardo, and 10 of Sheets' starts combine for the #1 slot.

The columns are similar to my last post: SERA is the cumulative starter ERA for the team and STDEV is the standard deviation of the starting rotation spots. Who would have thought, going into the year, the two NL teams with the most "even" rotations would be Chicago and Milwaukee? Ah, but enough of the preview, take a look at the table:

Team	SERA	#1	#2	#3	#4	#5	STDEV
San Diego Padres	4.11	2.54	3.15	4.12	4.62	7.16	1.78
Chicago Cubs	4.19	3.80	3.91	3.95	4.13	5.49	0.70
Arizona Diamondbacks	4.23	3.01	4.08	4.30	4.72	6.48	1.27
San Francisco Giants	4.24	3.38	3.86	4.09	4.47	5.70	0.88
New York Mets	4.40	3.47	3.72	4.05	4.48	6.98	1.42
Los Angeles Dodgers	4.43	3.03	3.58	4.24	5.35	6.63	1.44
Atlanta Braves	4.45	3.11	3.32	4.18	5.52	7.23	1.72
Milwaukee Brewers	4.55	3.46	4.24	4.81	5.16	5.20	0.73
Colorado Rockies	4.58	3.80	4.21	4.33	4.87	6.06	0.87
Houston Astros	4.71	3.21	4.44	4.75	5.26	6.29	1.13
Cincinnati Reds	4.86	3.73	4.21	4.58	5.29	8.08	1.72
Philadelphia Phillies	4.91	3.35	4.22	5.08	5.80	6.61	1.28
Pittsburgh Pirates	5.02	3.76	3.95	5.11	5.85	7.41	1.50
St. Louis Cardinals	5.04	3.57	4.11	4.86	5.69	7.53	1.55
Washington Nationals	5.11	3.56	4.46	4.68	5.77	7.82	1.63
Florida Marlins	5.58	4.64	5.13	5.21	5.81	7.67	1.18
NL	4.64	3.29	3.97	4.50	5.12	7.03	1.43

A #4 starter in the NL this past season would have been above-average if he'd put up a 5.00 ERA. Similarly, a team would have to count its lucky stars if their #5 starter did the same. I mentioned it last time, but this table really points out how hapless Florida's starters were: only 37 starts all year were made by a pitcher ending up with an ERA (as a starter) of 5.00 or under.

As in the FIP post, here are the pitchers who threw ace-level or better in 20 or more starts (~3.46 ERA):

• Jake Peavy, 34 starts, 2.54 ERA
• Brandon Webb, 34, 3.01
• Brad Penny, 33, 3.03
• John Smoltz, 32, 3.11
• Chris Young, 30, 3.12
• Roy Oswalt, 32, 3.19
• Tim Hudson, 34, 3.33
• Chad Billingsley, 20, 3.38
• Cole Hamels, 28, 3.39
Moving on to the #2 starters (~4.04 ERA):

• Noah Lowry, 26 starts, 3.92 ERA
  
• Carlos Zambrano, 34, 3.95
• Tim Lincecum, 24, 4.00
• Aaron Cook, 25, 4.12
FIP liked Lincecum, placing him among the aces, but he slides to #2 status here. You can't be too disappointed for the rookie, though.

Continuing the journey to #3 starters (~4.52 ERA):

• Jason Marquis, 33 starts, 4.43 ERA
• Tom Glavine, 34, 4.45
• Jay Bergmann, 21, 4.45
• Barry Zito, 33, 4.55
• Justin Germano, 23, 4.56
• Wandy Rodriguez, 31, 4.58
It's far too easy, but I'm guessing this is not what the Giants signed Barry Zito for.

Now we're into the hinterlands: #4 starters (~5.17 ERA):

• Woody Williams, 31 starts, 5.06 ERA
• Jamie Moyer, 32, 5.15
• Dontrelle Willis, 35, 5.17
• Chris Capuano, 25, 5.20
• Dave Bush, 31, 5.20
Not good news for Brewers fans here. At least you can say Moyer and Williams were showing their age, but it's never good when twenty-something starters struggle so much.

The final stop on the ERA track, those disturbingly bad #5 starters (~6.77 ERA). I've got to go down to 15+ starts for this, as most teams won't put up with such mediocrity for long.

• Kip Wells, 26 starts, 6.27 ERA
• Adam Eaton, 30, 6.29
• Tony Armas, 15, 6.40
• Jason Jennings, 18, 6.43
• Rick Vanden Hurk, 17, 6.81
Those are some unpleasant numbers. His poor season didn't stop Vanden Hurk from being a celebrity guest at the clinching game of the Dutch Championship Series.

To recap, the average NL starting rotation would be akin to this list:

1. Cole Hamels
2. Tim Lincecum
3. Barry Zito
4. Dontrelle Willis
5. Rick Vanden Hurk
Fold the Marlins into the Giants and you're 80% there!

Highest Season ERA, 10+ Wins

Here's a list of the only twenty-three players ever to win ten or more games in a season with an ERA 6.00 or higher. These guys may have won 10+ games, but no one really won when they were on the mound.

1. Colby Lewis, 2003, 10 wins, 7.30 ERA
2. Harry Staley, 1894, 12, 6.81
   
3. LaTroy Hawkins, 1999, 10, 6.62
4. Ryan Drese, 2002, 10, 6.55
5. Ray Benge, 1929, 11, 6.29
6. Wes Ferrell, 1938, 15, 6.28
7. Chris Carpenter, 2000, 10, 6.26
8. Mike Morgan, 1999, 13, 6.24
9. Pedro Astacio, 1998, 13, 6.23
10. Pat Caraway, 1931, 10, 6.22
11. Dave Burba, 2001, 10, 6.21
12. Brian Bohanon, 1999, 12, 6.20
13. Guy Bush, 1930, 15, 6.20
14. Dave Mlicki, 2001, 11, 6.17
15. Ramon Martinez, 2000, 10, 6.13
16. Red Donahue, 1897, 10, 6.13
17. Dan Daub, 1894, 10, 6.11
18. Bill Hutchison, 1894, 14, 6.06
    
19. Jason Dickson, 1998, 10, 6.05
20. Jason Marquis, 2006, 14, 6.02
21. Darren Oliver, 2001, 11, 6.02
22. Ad Gumbert, 1894, 15, 6.02
23. John Harkins, 1887, 10, 6.02

I find it interesting that 1894 had so many guys (4) on the list. The NL averaged a jaw-dropping 7.36 R/G (the highest since 1900 was 5.68 in 1930) - from 1893-1897 the league spiked in runs before dropping down to normal pre-1900 levels again.