I think everyone here knows that I’m not a Gryboski fan. He’s not exactly a bad pitcher, but he’s not actually good, either. He’s a mediocrity, and I don’t understand the commitment the Braves have made to him.
Gryboski put up a 2.84 ERA last year, but that is not at all indicative of the way he really pitched.
- He allowed 22 runs in 50 2/3 IP, but six of those runs were scored as unearned. Given a more normal ratio — say, 20 runs scored as earned — his ERA would have been 3.55, more in his normal range.
- He walked 23 men in those innings and struck out 24. The near 1:1 ratio is normal for him; his career K and BB are 89 and 83.
- He was often brought in with runners on base to theoretically get a double play, but 23 of the 60 runners he inherited scored, which isn’t very good.
That’s not to say that Gryboski doesn’t have some value. Righthanded hitters hit only .264/.303/.320 against him, which is superb, and he would be useful as a specialist against them. And nobody hit for much power against him — even lefties slugged just .397, and he allowed only two homers — which is very useful.
It is my feeling that Bobby misuses Gryboski. His double plays don’t make up for the additional runners he creates with his walks and comparable hittability, and his low strikeout rate puts a lot of pressure on the defense in situations when they’re already stressed. My feeling is that Gryboski should not be used in any situation where he’s likely to face a lefty with runners on base, and generally shoulnd’t be asked to pitch with a runner on second or third. The proper Gryboski situations are either to face one righthander with a runner on first, less than two out, or to start an inning with a number of righthanded power hitters due up.
Very nice breakdown Mac. I think this paints the clearest picture of Gyrb’s strengths & weaknesses that I have seen.
Your first point doesn’t make sense, and it’s not logical. You can’t take total runs aloud and then turn unearned runs into earned runs. If he allowed less unearned runs he would also have allowed less total runs. I don’t know how anyone who understands what makes a run unearned can even get confused as to think that this makes sense. If the run wasn’t unearned, it wouldn’t have scored in the first place, that’s the whole point of “unearned”.
However, the other two points, particularly the third, are good.
Grybo definately has his uses. It’s nice to have one guy in the pen that you can just throw out there anytime, anyplace and you don’t really care if his arm falls off or not.
I don’t really understand why he’s not a little better. His sinker bites and he throws in the low to mid nineties. If he could ever sharpen his control just a bit, he could be Ben Weber. As it is he’s stuck at the Al Levine level of effectiveness. Still, Levine is entering his 10th big league season.
Errors are an extremely marginal concept in modern baseball, and to count earned runs separately from unearned runs makes little sense. After all, Gryboski wasn’t getting ground balls that somehow turned into home runs. He would get a ground ball that a scorer decided that someone — usually Furcal — should have made a play on, but didn’t, and so it’s called an error. And then Gryboski, or a reliever following Gryboski, allowed at least two more events to occur to allow the run to score. Why shouldn’t he be charged with those runs?
Moreover, the charging of unearned runs often makes no sense. During the season, I mentioned on at least a couple of occasions that a run that certainly would have scored anyway was called “unearned” because an error allowed a runner to advance. Guy hits a single, the next guy hits another single, the runner scores on a throwing error, then the next guy hits a home run. The first run’s called unearned even though the error didn’t really cause the problem.
I believe the first run would still be an earned run in that scenario.
I don’t know, but it seems random. Here’s one I looked up… The Braves beat the Expos 1-0 in a game last year. The one run came on a bases-loaded, one-out grounder to deep first that Nick Johnson mishandled. You can’t assume a double play (even if the rules allowed it, Drew was running and never would have been doubled up) and the throw to the plate would have been very difficult, but the run was scored unearned. Leaving aside that question, the pitcher loaded the fricking bases! Who’s really at fault here?
The whole earned runs / unearned runs thing had its purpose … in the 1880s when players played without gloves and there were hundreds more errors than now. But these days its just a silly remnant.
If Ryan Klesko has the mobility of a lame turtle and a ball falls in just in front of him, its a hit and an earned run. If Charles Thomas runs just slightly less than the Boston Marathon and allows the balls to hit off the heal of his glove, its an error and an unearned run.
In both scenarios, the play wasn’t made and the run scores. In both scenarios the pitcher is partially resposible and the fielder is partially responsible. The black/white dichotomy of the earned and unearned run is just not an accurate view of the real world. A pitcher is not 100% responsible for all earned runs and isn’t 100% free from blame for all unearned runs.
I would love for them to do away with the whole earned/unearned run thing. I don’t like that an error completely absolves the pitcher of the responsibility for having put runners on to begin with.
We don’t give the pitcher credit for an imaginary earned run when a fielder makes a spectacular play, so let’s not pretend a run didn’t score when a fielder makes a bad play. I think it’d give us a clearer picture overall to use RA instead of ERA.
We could do away with earned runs, but that would create problems of its own — just as there are cases where unearned runs might mostly be attributable to the pitcher, there are other cases where a run is much more attributable to poor fielding (Jose Conseco’s Homerun off the head comes to mind). I think we’re looking at an imperfect statistic either way we approach the issue of runs scored.
It’s obvious who is not a pitcher here? Remove earned runs? ERA is already a bad enough indicator of ability, total runs would only be worse. Some idiot drops a fly ball and the pitcher gets fault? That doesn’t make sense.
Very very very rarely is a run scored unearned when it really shouldn’t have. Then again, this whole discussion reminds us why it’s best not to rely on ERA too heavily.
Neither shows the whole picture, which is why the error stat has some marginal value. But pure Run Average would give a better picture than ERA does.
Grst, leaving aside that most errors are on infield grounders and not fly balls… Very few unearned runs are scored without the pitcher allowing at least two other “events” — a hit, a walk, a stolen base, or a long fly ball. If it’s only one “event” other than the error, it’s always at least a double. The typical UER is something like a walk, a grounder to an infielder which he rushes (trying to get the DP) and botches, everyone safe, then a hit. Or the “third out” reaches on a ball through the first baseman’s legs, and then the pitcher gives up a homer.
Plus, I can’t say for certain, but I’d guess that half of the time the error is charged, it comes on a difficult play. (How many times a year do we hear Skip & Joe wondering if an error will be charged? And half the time, they think that the scorer got it wrong.) The scorer decides that Furcal “should” have made the play, instead of it tipping off his glove. Many shortstops wouldn’t have even gotten to the ball, or if they’d gotten there throwing him out is in question, but the scorer had an argument with his wife before the game and is being vindictive that day. It’s a judgment call, and there’s no real need for it.
Neither shows the whole picture, which is why the error stat has some marginal value. But pure Run Average would give a better picture than ERA does.
There is no evidence of this. I would say the RA is equally as inaccurate as ERA.
ERA makes an attempt to remove the impact of gross fielding errors from a pitcher’s performance record. This is a good thing to try to do, unfortuantely it is poorly done due to inconsistancy with which errors are scored. RA makes no attempt to remove fielding mistakes from the pitcher’s performance record and thus has a great deal of misrepresentation built in.
Neither does a great job of telling just how effective a pitchers was. At the end of the season, if you give the pitcher credit for half the unearned runs you probably get a lot better look at a pitcher’s performance than either ERA or RA would give you. This is (close to) what Mac has done with Grybo, and I think that is a good way to look at it.
So you’re saying his ERA was good but if he had given up more earned runs, it would be higher. Makes perfect sense…
Sorry for the double post but I just noticed that Horacio Ramirez had an even higher ERA to RA differential than Gryboski. Yet Ramirez’s DER was ninety points higher than Gryboski’s. So while the fielders were more on the ball behind Horacio, he still allowed more unearned runs than KG. Their walks and strikeouts were very similar. I know you’re also not much of a Ramirez fan, but what does this indicate of his effectiveness by your rationale?
I just can’t see how you can just remove unearned runs. Even if the pitcher is partially responsible, what you’re saying is making him fully responsible because he’s the one on the mound.
A pitcher walks a guy, the next guy bunts him over, the next guy hits a 62 hopper to the first baseman to get the runner over to third. Then, the next guy hits a pop-up to the second baseman. Boom, second baseman drops it, run scores. In Runs Scored Average Twilight Zone, the pitcher becomes fully responsible for that run and is charged with it. Here’s another: Furcal reaches on a bunt single with two outs and Giles hits a ball a million feet in the air to the warning track. Juan Pierre drops it with Furcal motoring around the bases to score. Pitcher fully responsible? Yeah right.
Here’s the epitome of this nonsense: a guy hits a ground ball to the shortstop, shortstop airmails the throw into the stands and the runner takes second. The next guy bunts the runner to third. The next guy hits a sac fly and the runner scores. The guy makes three pitches to get three outs, HIS JOB, and the defense can’t get the job done and you want to hold the pitcher responsible? Absolutely not.
One could go all night about these scenarios, but it’s simple. The pitcher’s job is to get outs. He can’t physically get 27 outs, so that’s why he has 8 other guys. If he makes the hitter hit a ball that gets to a defender, and in the case of Charles Thomas running 25.9 miles to get to a ball and then booting it, then the pitcher did his job. If it hit leather, the pitcher did his job. If you aren’t going to compensate for counting total runs but putting some of the weight on the fielders, then you just can’t do it.
I should have noticed the Ramirez thing. And that accounts for a lot of the difference between his ERA and his K/BB and K/9.
Rob, I’m not saying that you should totally discount the difference, but I think that the majority of the fault in 90 percent of unearned runs will be due to the pitcher. Now, if something happens like sometimes did last season where the middle infield totally breaks down and Ramirez (I think this happened to him) needs to get six outs, that’s a different matter. But those things are relatively rare.
We’re in danger not of ignoring defense, but of double-counting. At one level, you pay attention to the ERA, but then you (this is a theoretical you, not any person) pay attention to the defense behind the pitcher.
Finally, I think unearned runs are going to happen more often to a ground-ball pitcher (like the three righthanded relievers I’ve grouped) than to a fly-ball pitcher like Russ Ortiz, who only had four unearned runs last year. That’s just simple math. The Braves’ outfielders committed eight errors as a group last season. Furcal committed 24 by himself. I think you have to take that into account, especially for an extreme ground-ball pitcher like Gryboski. I haven’t seen a detailed study of this, but it’s pretty simple and I’d bet it’s been done, if someone more up on the literature can point me to it.
Rob, the pitcher’s job is to get outs. The easiest way to do that is by a strikeout. The next easiest is by allowing infield and foul popups. Harder is by allowing flyballs, harder still by allowing ground balls, and hardest of all by allowing line drives. If the pitcher gives up a screaming liner that hits leather, he didn’t do his job. If he induces a popup to the second baseman, he did. There are lots of shades of gray in between.
Anyway, because it’s my nature when questions like these get asked, I whipped out my handy copy of the lahman database and did a quick study on year-to-year correlation of ERA and RA since 1990 for pitchers with at least 50 IP in consecutive years. There were over 2600 pairs of player-seasons that met such criteria (a player who threw at least 50 IP in both seasons since 1990) giving us a sufficiently robust sample size. I can’t properly insert the Excel correlation table or regression results, so I’ll describe what I found.
RA correlates better (if only slightly) year to year than does ERA: the R for RA_Year-1 and RA_Year0 is .3327, compared to .3265 for ERA_Year-1 and ERA_Year0. Additionally, RA has SLIGHTLY more predictive power than does ERA: R^2 of .108 compared to .106 (when used to predict future ERA). Both are vastly inferior to Voros’s DIPS.
Executive summary of above for non-economist or stats people: generally speaking, a pitcher’s Runs Allowed per nine innings in Year Zero correlates more strongly with his Runs Allowed per nine in Year One, than does a pitcher’s ERA in Year Zero compared with it in Year one. In addition to that, given the task of predicting a pitcher’s ERA next year, you do slightly better by using his RA/9 this year instead of his ERA.
All this goes to suggest what Mac says: while a pitcher may or may not deserve a particular earned run, it’s best to ignore them altogether when predicting future performance. Instead of RA, though, use an even better metric: strikeouts, walks, homers allowed, and ground ball ratio.
Just cause an unearned run requires an additional hit or anything doesn’t matter. The whole point is you can’t assume you know what would happen. As a pitcher, I can assure you, that there’s a difference pitcher from the windup or stretch, with a runner on base or bases empty. ERA is not accurate, RA would be worse. If you’ve played the game, it’s pretty damn easy to figure out when a play is or is not an error.
“Rob, the pitcher’s job is to get outs. The easiest way to do that is by a strikeout. The next easiest is by allowing infield and foul popups. Harder is by allowing flyballs, harder still by allowing ground balls, and hardest of all by allowing line drives. If the pitcher gives up a screaming liner that hits leather, he didn’t do his job. If he induces a popup to the second baseman, he did. There are lots of shades of gray in between. ”
Greg Maddux would laugh in your face from those statements. You can’t apply statistics to methodology. If you taught that to young pitchers you’d get a bunch of throwers and not pitchers, who don’t have the first clue what they are doing. Trying to blow the ball by everyone instead of pitching smart.
“All this goes to suggest what Mac says: while a pitcher may or may not deserve a particular earned run, it’s best to ignore them altogether when predicting future performance. Instead of RA, though, use an even better metric: strikeouts, walks, homers allowed, and ground ball ratio.”
That, on the other hand, is quite correct.
I think you misunderstood me. Striking people out is not easy; but if you can do it, you’ll get an out 99.9% of the time (the .1% saved for the instances that the catcher drops the third strike). Infield popups become outs almost that frequently; fly balls, less frequently; and on, and on, until we get to line drives, which are the toughest batted balls to turn into outs.
The ability to INDUCE a popup or ground ball or strikeout is wholly separate from its correlation with out percentage. I completely agree that striking someone out is incredibly difficult, and would never coach a young pitcher by telling them to go for the strikeout at all cost.
Glad you agree with using DIPS numbers instead of ERA to predict future performance. That’s a pretty revolutionary stance to take, and I didn’t expect a lot of support for it (although I should have known better at a sabermetric-friendly blog like this one!).
Using RA probably won’t make much of any difference at the major league level, but if you tried to use it with college or highschool pitchers the results would be horrible unreliable and won’t tell you anything. Thus, for those of you who want to use such analysis as predictive to future performance, you’d only be cutting your own legs out from under you.
“I think you misunderstood me. Striking people out is not easy; but if you can do it, you’ll get an out 99.9% of the time (the .1% saved for the instances that the catcher drops the third strike). Infield popups become outs almost that frequently; fly balls, less frequently; and on, and on, until we get to line drives, which are the toughest batted balls to turn into outs.
The ability to INDUCE a popup or ground ball or strikeout is wholly separate from its correlation with out percentage. I completely agree that striking someone out is incredibly difficult, and would never coach a young pitcher by telling them to go for the strikeout at all cost.”
Ok. I gotcha. A pitcher capable of high strikeouts certainly has an easier time than one that does not. However, groundballs are in fact the easiest way to get outs, and very effective. You very rarely run the chance of extra bases, and when people do get on you have a better chance of double plays. Less pitches means you can go longer per game, on average. For some reason I got the impression you were advocating a certain pitching style. Whereas a hitter can conciously increase his walk rate, thus resulting in better performance, it’s much harder for a pitcher to change his style and become a strikeout pitcher instead of a ground ball pitcher, or vice versa (though appearanlty Kolb has done just that, so not impossible).
I’m not sure about the groundballer vs flyballer thing. What you say is true about less XBH and more DPs. On the other hand, groundballers allow more singles. Also, non-homer flyballs are more frequently converted to outs than are groundballs. Of course, everything I just said is dependent on the quality of the defense behind you; inducing ground balls doesn’t do much good if Nomar Garciaparra is your shortstop. My inclination says you’re right (long live Dan Kolb), but the data are not conclusive.
Well, from an end result point of view you could be right. But I think you’re more likely to find a successful groundball pitcher than fly ball pitcher as I believe ground balls have more margin for error. Getting a lot of fly balls means your pitching up in the zone alot. If you can throw 100mph, then that’s fine. For everyone else that can be dangerous territory. A high fastball is far more likely to end up an XBH than a sinker. Just my observations, anyway.
RA has SLIGHTLY more predictive power than does ERA: R^2 of .108 compared to .106 (when used to predict future ERA).
This is not a statistically significant difference, and could just as easily be flipped if you changed the threshold to 51 innings. Thanks for doing the legwork, as it confirms what I said earlier that RA and ERA are equally inaccurate measures, and debunks Mac’s theory that unearned runs are 90% the pitcher’s responsibility. It’s pretty much 50%.
It’s obvious who is not a pitcher here? Remove earned runs? ERA is already a bad enough indicator of ability, total runs would only be worse. Some idiot drops a fly ball and the pitcher gets fault? That doesn’t make sense.
But it makes sense when the pitcher loads the bases and then gets zero blame when an error allows a run to score? It’s easy to talk about unearned runs in terms of the worst-case scenarios of when the fielder screws up, but what about the worst-cases when the pitcher sets up a bad situation and _then_ the fielder screws up once and then the pitcher screws up some more?
All I’m saying is that i think unearned runs shouldn’t be disregarded. Guys who have fewer unearned runs may be doing something better than those with lots of them – better damage control. Since errors are a fact of baseball life, how you handle them is important.
Consider the 1998 Cy Young race. Without getting into other issues like innings pitched, let’s just look at ERA – of the top contenders, Tom Glavine had a higher ERA than both Greg maddux and Kevin Brown. However, Tom had a better RA, because he allowed only 4 unearned runs, vs. 9 for Brown and 13 for Maddux. That had to count for something.
Pitching relies on fielding, but during any given year it both benefits from occasional great plays and suffers from occasional bad plays. i figure those balance each other out in the end.
“All I’m saying is that i think unearned runs shouldn’t be disregarded. Guys who have fewer unearned runs may be doing something better than those with lots of them – better damage control. Since errors are a fact of baseball life, how you handle them is important.”
You say they should be disregarded then you turn around and saw how they should be regarded. They should be disregarded, as in ignored, not projected through flimsy reasoning into potential earned runs.
Grst,
I’ll say it: I think that we should 100% ignore errors when judging a MLB pitcher’s record. Absolutely, completely, totally.
I am not a 100% devotee of DIPS. But I do buy the logic that once a ball is put into play, whether it is turned into an out or the batter reaching is largely a function of the defense. Whether a ball in play is a hit is first a function of the batter, second a function of the fielders, and lastly and least significantly a function of the pitcher.
If you trade leadfooted Jeff Blauser for Ozzie Smith, the pitchers’ ERAs will go down not because Oz made fewer errors, but because he got to more balls.
Taking the subsection of balls in play and dividing them further into ones that fell in without a fielder touching it and those that were touched first, without likewise dividing it into ones that a player with average range would or would not have gotten to is misleading to the point that it distorts the record.
Baseball record keeping is amazing. We know what batters from 150 years ago did with great precision. Similar data has been kept for pitchers for nearly as long. But fielding is woefully inadequate, judging successful plays and only one very small subset of unsuccessful plays. Errors (together with the paired statistic unearned runs) has long since outlived its purpose.
And just to get what seems to be an oft repeated refrain from you out of the way, my playing career up to American Legion ball many years ago, included substantial time on the mound.
Looking back, I see that people are using the phrase “ignore errors” in multiple ways in this thread. To make sure I’m clear, I think that there is no reason to apportion runs allowed when a pitcher is on the mound into earned and unearned and that this dichotomy is false and misleading (and that I used and too many times in one sentence and period.)
“All I’m saying is that i think unearned runs shouldn’t be disregarded.
You say they should be disregarded then you turn around and saw how they should be regarded.
Huh? I said they shouldn’t be disregarded. Please note the “n’t”. I want the distinction between earned and unearned removed.
They should be disregarded, as in ignored, not projected through flimsy reasoning into potential earned runs.
Projected? The act of ignoring them involves projecting some version of what a potentially biased hometown scorekeeper thinks “should have happened”. My approach involves no projection – just count the things.
Why, exactly, are you so passionately against counting them? What do you think would be the negative consequence of doing so? It’s not like it would dramatically reorder the rankings of who’s considered a good or a bad pitcher.
Wow, this discussion is great. Nice entry Mac, to inspire so much analysis.
Doesn’t have a lot to do about Gryboski at this stage, but I’m enjoying it.
To make sure I’m clear, I think that there is no reason to apportion runs allowed when a pitcher is on the mound into earned and unearned and that this dichotomy is false and misleading (and that I used and too many times in one sentence and period.)
It sure is, but giving a pitcher 100% blame for runs scored because they are stuck on a team that routinely screws up routine plays is false and misleading as well. There is no right answer here. We’ve seen that RA is no better of a predictor than ERA. If official scoring were 100% accurate and uniform the unearned run route would be the way to go, but it’s not and it’s only getting worse. As it is, RA and ERA both are not that great and we would all be better of ignoring both numbers and looking at DIPS. While DIPS isn’t perfect, it’s much closer than RA and ERA.
nyb,
Jeeze, did you have to quote the parenthetical too? 😉 We’re getting rather esoteric but what the heck, I’ll go along.
RA provides and exact record of something that happened on the field of play: how many runs scored when the pitcher was pitching. (With the obvious exceptions of inherited runners from both starter and reliever’s POV.)
ERA does not. It provides a mixed record of what happened and what an official scorer thought should have happened. It blends fact and opinion but is labeled in the common parlance as a fact. To that degree it is distorting.
McCracken’s DIPS (or my prefered version of Tangotiger’s FIPS) has strong predictive value. By regressing BABIP to league (or team) norms, it is a better predictor of future runs allowed (or earned runs allowed if you wish) than simple RA. I repeat, DIPS predicts future ERA better than does ERA itself. It provides a solid framework to establish what — all other things being equal — the pitcher is likely to do in the future.
But unlike RA, it does not establish what happened. By saying that a team’s DIPS indicated ERA is X.YZ, a writer postulates what that ERA should have been had the defense been average. RA is an exact measure; it tells one exactly what actually happened. It doesn’t attempt to apportion credit or blame, it just simply tells exactly what occured. DIPS is a projection. Based on solid logic and empirical data, but a projection none-the-less.
So the two are complementary. Depending on your purposes, either may be the appropriate measure. ERA, as a mixed hybrid of fact and opinion doesn’t tell us what actually happened and isn’t a good predictor of future performance. Its only present day purpose is a weak correlation to moral culpability for the runner scoring. In this context, I find such apportionment of blame to be silly and, as such, serves no meaningful purpose.
It’s pretty simple. Runs allowed is about as usefull as wins when it comes to measuring INDIVIDUAL performance. It relies too much on the performance of other people. ERA is a simple, though not very effective, method to correct this. It’s not perfect, but it’s much better than runs. If you’d ever been on the mound with shitty defense behind you, we wouldn’t be having this discussion.
Grst: You slut. Calm that ego down!
In regard to the situation in the Expos game you mentioned and Nick Johnson’s error, the bases were loaded with one out. I doubt that the scorer pre-supposed a double play. They never do. The likely scenario is that the Expos were probably playing in at the corners, and the scorer judged that Johnson was coming home on the play to get the force at the plate. That would have been out #2, and the next batter struck out. It’s not possible to tell from the game log or the write-up on the game.
If, for example, the SS and 2B were playing back for the DP, and one of them bobbled it, the run would have been earned. It is hard to judge from the game log what actually happened on a given play and what the scorer saw. And it is not uncommon for the scorer to change a ruling later, even after the game.
I remember the play a little. Johnson couldn’t have made the throw home. Actually, I wasn’t sure it shouldn’t have been a hit.
Grst, I’m curious. You’ve trotted out the “If you’d ever been on the mound …” card several times. Please feel free to let me know when you played and at what level. MLB? High minors? low minors? Division I college program? Small college? This past season? The 1980s? The 1950s?
Mac, I’m getting old … is that 1-0 game against the Expos that you’ve been referring to the game played in San Juan two (?) seasons ago?
No, it was this past season, because Drew was the hitter… Here’s the recap.
“Grst, I’m curious. You’ve trotted out the “If you’d ever been on the mound …” card several times. Please feel free to let me know when you played and at what level. MLB? High minors? low minors? Division I college program? Small college? This past season? The 1980s? The 1950s?”
No card, just observations. Actually, that comment was more about the joys of having a shitty defense than just having pitched. An unpleasureable experience, for sure. Collegiate level, recently.
If you’d ever been on the mound with shitty defense behind you, we wouldn’t be having this discussion.
Well, that feels a little dismissive and largely emotional. But again, keeping things at a rational level – we’re perfectly willing to let a pitcher benefit in his stats when defense is good, but why does he also get to benefit when it’s bad? Bear in mind here that ERA/RA is a two-sided coin. When that scorekeeper makes a bad decision, he affects both hitter and pitcher. Right now it seems to me that the equation is tilted completely toward favoring the pitcher.
And if you’d ever been at-bat facing a stellar defense, you wouldn’t be having this discussion 😉
(that was a joke, please don’t overreact.)
It’s pretty simple.
Clearly not, else we wouldn’t be having such a lengthy thread in February.
Runs allowed is about as usefull as wins when it comes to measuring INDIVIDUAL performance. It relies too much on the performance of other people. ERA is a simple, though not very effective, method to correct this. It’s not perfect, but it’s much better than runs.
I just don’t see that it’s better than runs. Both seem to have equally predictive value, so the discussion may be overblown. But if we have two essentially equal methods, but one is based on simple counting and the other is based on a fallible and often biased human being piecing together in his mind what he thinks should have happened, often disregarding things like relative fielder range and whether an errant throw would even have been attempted by a lesser fielder, but also factoring in whether the player in question is on the home team or the visiting team, then I’ll go with the one based on simple counting of runs. At least it takes politics out of the thing (I’m looking at you, Juan Gonzalez, the guy who tried to intimidate a scorekeeper while chasing an RBI record).
” Right now it seems to me that the equation is tilted completely toward favoring the pitcher.”
How so? Since, as so many people have pointed out, error giving is not perfect, isn’t it just as likely that pitchers might get hurt by things not counting as errors? How many times can you recall where there was a “mental error” that wasn’t counted? A ball dropping in the outfield with three people standing around with their thumbs up their butts, for example. Since the wrongs could go both ways, I see them as cancelling, leaving the slightly more useful ERA (though still not very).
“Both seem to have equally predictive value, so the discussion may be overblown.”
I’ll agree on that.
“But if we have two essentially equal methods, but one is based on simple counting and the other is based on a fallible and often biased human being piecing together in his mind what he thinks should have happened,”
But you completely ignore the application with that analysis. It may be simple counting, but it’s counting the performance 9 players, not one. So when you turn around and apply it to measure the performance of a single player…do you see how it could be misleading? It’s not that ERA is perfect, but it’s an attempt to correct this very problem. To take a team statistic and make it an individual one. If you can come up with a better way to do it (and I’ve no doubt some of you math wizzes could find a way) then be my guest.
As you say, this is overblown, especially at the major league level. But wasn’t a big part of Beane’s “revolution” the application of statistical analysis to the collegiate level performance as a predictor of future expectation? Let me throw out a hypothetical then.
Take two approximately equal pitchers, in the same division. One has a good defense, one has a really bad one (and there are plenty in college). Isn’t it problable that the one with the worse defense is going to have a much higher RA? Well, ERA fixes that, and they reflect their abilities with close ERA’s. Comparing RA across those pitchers would mean nothing. This is the point I’ve tried to make all along, though I’ll admit I articulated it poorly.
Take two approximately equal pitchers, in the same division. One has a good defense, one has a really bad one (and there are plenty in college). Isn’t it problable that the one with the worse defense is going to have a much higher RA? Well, ERA fixes that, and they reflect their abilities with close ERA’s.
No. Absolutely not, at least on the major league or high minor league level (which is where this discussion started).
The largest factor between good defense and bad is in the fielders range of the players, not the errors.
The NL’s Fielding Percentage last year was .983. 12 teams were within +/- 3 points of that and the biggest outlier was -6 points. Six one-thousandths of a point!
The NL team DER (essentially the percentage of balls hit into the field of play that were turned into outs) was .695, but the spread was much larger with the difference between best and worst in excess of 34 points, more than three times the spread in FP. Stated more precisely, the standard deviation in DER is much higher than the SD of FP.
Or another way of looking at it is a team that makes a lot of errors will make less than 0.3 extra errors per game while a team that doesn’t turn many balls in play into outs will allow more than one extra hit per game. The run scoring impact is more than three times as great for the range issue than for the errors issue, especially considering that a significant portion of errors (throwing Es) allow runners to advance, not runners to reach.
So what ERA does is attempt to account for one type (the smaller type) of bad defense and then lead one to believe that it significantly adjusts for defense. That is grossly misleading.
If you can come up with a better way to do it (and I’ve no doubt some of you math wizzes could find a way) then be my guest.
DIPS, mentioned before does a much better job — at least at the major league level — of determining what the pitcher is responsible for. FIPS (presented at the Hardball Times web site) is a similar formula. Bill James’ ERC (component earned run average) does the same type of thing.
The NL’s Fielding Percentage last year was .983. 12 teams were within +/- 3 points of that and the biggest outlier was -6 points. Six one-thousandths of a point!
I thought I was done here, but this is just incredibily misleading. The spread between the most and least team unearned runs in the NL last year was 67, or 0.4 runs a game. It might not seem like much but that’s a big deal. In the AL it was 51 runs.
I just don’t see that it’s better than runs.
Unearned runs are pretty useful if a pitcher has significantly more or less than the league (or team) average. If he’s got more unearned runs it indicates that he’s not as good as his ERA suggests (and not as bad as his RA would suggest). And vice versa.
Unearned runs can indicate that a pitcher was hurt more by his defense than the average pitcher, and that’s something I want to know if I’m evaluating players. While the information coming from unearned runs is not perfect, and does not account for the larger effect of range differences, that’s still know reason to simply throw valuable information away.
“No. Absolutely not, at least on the major league or high minor league level (which is where this discussion started).”
Maybe I wasn’t clear, but I was talking about college pitchers. Most of you post was aruging something irrelevent to what I stated. I already stated the differences are not significant enough at the major league level. I specifically mentioned how it could affect the evalution of college prospects. Sorry if I was not clear enough on this point.
“While the information coming from unearned runs is not perfect, and does not account for the larger effect of range differences, that’s still know reason to simply throw valuable information away.”
Exactly the simple point I’ve been trying to make all along.
Here is how that run was scored in the 1-0 game against the Expos that you mentioned Mac.
BRAVES 7TH: DeRosa walked; Ortiz struck out; Furcal doubled to
center [DeRosa to third]; HORGAN REPLACED DOWNS (PITCHING);
Marrero was walked intentionally; J. Drew reached on an error by
Johnson [DeRosa scored (unearned), Furcal to third, Marrero to
second]; C. Jones struck out; AYALA REPLACED EVERETT (PITCHING);
SLEDGE REPLACED HORGAN (PLAYING LF); A. Jones struck out; 1 R, 1
H, 1 E, 3 LOB. Braves 1, Expos 0.