Errors

[538280]

I just recently bought Craig Wright's book The Diamond Appraised, and I finished the chapter about how errors are the most useless stat in baseball. Wright made some good points. He mentions how errors being committed are not completely dependant on the fielders. He notes how drawing a "base on error" is actually something that certain batters do consistently while others almost never will. He notes how Frank White consistenly has gotten lots of base on errors over the years. a fast runner who hits lots of ground balls to the left side of the infield like White should do that. Wright argues that White should be credited for this.

He also notes how a fly ball hitter who strikes out a lot rarely gets on base because of an error. Wright argues that those players should be credited or penalized for that, and thus the game would benefit from errorless scoring.

He criticizes ERA because of the definition of an earned run. He says a pitcher should be penalized for all runs he gives up. He points out that ground ball pitchers with few strikeouts always give up more unearned runs than fly ball pitchers who strike out a lot of batters. He says the fly ball/strike out pitchers should be credited for this.

This is the part of the analysis that I found to be particularly interesting. There has been lots of talk over the past few years about "Defensive Independant", or DIPS, pitching analysis. Wright pointed out that ERA is actually already supposed to be defensive independant, yet isn't because it acutally is a "skill" to reduce unearned runs. This increased my interest in DIPS analysis.

He also says that fielding percentage is barely worth measuring in today's game, though it was important in the deadball era. He says that players make such a high percentage of plays versus errors anyway that it makes very little difference.

I think here Wright again has a good point, but he can be taken too literally. Obviously it is a skill to be sure handed and make few mistakes in the field. Particularly for a third basmen.

What do you guys think about all of this? I frankly found it to be fascinating and something I have never really thought about. Particularly the discrediting of ERA.

[SABR Matt]

I think errors for fielders should be tracked, but "Reached on Error" should also be tracked for hitters.

a) I don't believe it is a "skill" for a hitter to reach on an error in most cases. For most hitters, reaching on errors is directly related to what happens when you make contact and how frequently you do this...and to your groundball/flyball rate. There are rare exceptions (really fast players force more errors if they put it on the ground a lot, see Suzuki, Ichiro), but generally, the creation of errors against is just related to batter tendencies.

That said, those tendencies are phenominally important....for one VERY big reason. When we use Linear Weights to project how many runs a player creates for his team, we generalize the average result of each ball in play "out" (including the errors) and each strikeout (errors being one of two primary reasons Ks are a little worse statistically than BIP Outs)...when we do this for whole teams, we end up with an average error of 15-35 RC away from what the team ACTUALLY scored. I believe a big chunk of that error is in the averaging process. My analysis suggests that high-K/Flyball offenses have their RC slightly OVER-estimated and groundball offenses have their RC slightly UNDER-estimated.

As such...b) Although I don't think hitters have the power to create errors with their bats other than by getting more grounders...I think the value of hitting it on the ground must be considered ESPECIALLY for analysis of pre-1920 baseball. This is IMHO why RC estimators get so horribly inaccurate when you go back to before Babe Ruth and the live ball.

The same thing can be applied to analysis of pitching results.

Unfortunately, until we somehow manage to fill in PBP data back to the earliest times in baseball history, we'll never know EXACTLY how many errors offenses benefitted from.

[digglahhh]

Chris,

I think that this is an extremely interesting topic, especially in light of the discussion we had in the Chipper Jones thread about the subjectivity of official scoring. It doesn't take too much time working with official scorers to realize that you are dealing with an art and not a science. This is one of the reasons why I set the bar higher than most when I'm reading somebody's conclusions driven from PBP data.

Here is an exerpt from the post I was talking about

Take for example, a very simple situation:

One out, runner on first and third. Batter hits a pop of between the second baseman and rightfielder. The runners hold because they don't know if it is going to be caught. The 2B is going back and gets his glove on it, but it drops. Now if the runner from 1B can't make it to second base before the 2B can recover and force him at second him at second, this is a force out and the batter is credited with an RBI when the runner on 3rd scores. If that the second baseman doesn't recover and the runner from first reaches 2B safely, we have an error, and no RBI when the runner on third comes home. In fact, say the next two guys strike out, now that run is unearned too.

Read that scenario again, Chirs. In this case, the RBI, error and unearned run are determined by whether or not the runner on 1B can reach 2nd before the fielder can recover from his misplay- that may or may not be scored an error. The same exact play can lead to two vastly different statistical outcomes. I can give you countless scenarious like this.

Here is another one that happens all the time. Two outs, home team pitching, runner on first. Runner reaches second on some kind of error, an errant attempt to pick him off by the catcher or something. Next batter wallops a double into the corner. Next batter pops up. Official scorer rules this an unearned run, without the error the runner is on first, the double only advances him to third. The OS has the latitude to make this decision, and he does.. a lot. Now, you and I both know how frequently runners score from 1st on doubles with two outs- but that doesn't matter. The next morning in the box score what do we have- objective reality, but not really.



Now it is quite obvious that in certain scenarios the determining factor of whether or not something is an error, RBI, earned or unearned run hinges upon a player who is indirectly related to the statistical ramifications of the play. That is only within the physical game; umpires, official scorers and the PR crews that pressure OSes to rule the outcome of the play in the manner that is most statistically beneficial to the team they represent, add even more levels to this dynamic. I'm serious about these PR guys, they're like the Congressional lobbyists, and OSes bow to the pressure more often than you would like to think they do.

I would also advocate the "team error." This is actually something that I have been meaning to start a thread about for a while now. When two or three fielders converge on a pop-up, miscommunicate and the ball drops between them- that is an error! You can't attribute it to one player, so it is not ruled as such, but it is certainly errant. Manny diving and cutting off that throw for no reason- that's an error. Sometimes errors are made, and the fielders are bailed out b/c the runners don't successfully advance, so the is no error charged. A team committing no errors is certainly not automaitcally indicative of them playing expemplary defense. There are certain mistakes you can make and not be penialized statistically for, and others that you can't.

What about a "hit" that was really the product of a ball being hit a few inches out of the range of a poor fielder. A ball that is two feet out of the range of Michael Young, is probably a ball that should be caught by the average defensive SS. Now DIPS recognizes these trends for pitchers and their teams, but it gets deeper, and infiltrates much of the most common stats. What if that play happens with two outs, and the next guy hits a homer, those two runs are earned, but are they really? Or are they given by Young? You can't tell the difference between a legit hit and one that was a 1few inches out of the reach of a bad defender simply by looking at PBP data.

Look, if that scenario happened to a guy like Santana once more than it happened to Millwood, he would have won the ERA title. Errors and ERA are closely related, and these two rulings have tons of situational idiosyncrasies that make their outcomes suspect, especially when trying to analyze data from 30 years ago just looking at PBP record.

These are the types of things I am concerned about when I talk about how all complex metrics have some flaws before the first calculation is made. The data itself has biases, the system is more concerned with being standardized, repeatable and "correct" than being descriptive and accurate.

[SABR Matt]

Wait a second...

In that scenario digglahhh...you assumed that if the runner reached second base, it would be scored an error...if it was even slightly difficult for the second baseman to get to the ball, most of the time, that's scored as a SINGLE...not an error, and the RBI counts as does the ER.

[digglahhh]

Not necessarily, Matt.

Actually I believe this exact (or at least very similar) play happened last year in the World Series with a Craig Biggio pop-up.

[SABR Matt]

Not necessarily, Matt.

Actually I believe this exact (or at least very similar) play happened last year in the World Series with a Craig Biggio pop-up.

More often than not...a flyball that is "between the RF and 2B" is ruled a hit unless there's a very obvious miscue on the part of the second baseman in handling the ball (it hits the heel of his glove and bounces away)...if he gets a glove on it, but just the fingertips or the webbing...it's clearly a hit.

[Chisox]

I would also advocate the "team error."

So would I. How about when an outfielder throws home and the collision forces the ball to the backstop allowing the hitter/runner to advance when a simple throw in wouldn't have? Since the outfielder threw the ball, and an unearned base was taken on a miscue, the outfielder gets the error. But is anyone going to say the outfielder made a mistake in trying to get the out by throwing it right on target? You see these consantly, but is it really fair for one player getting penalized for doing something right?

[SABR Matt]

Also the team error would allow scorers to issue an error in cases where no player has touched the baseball but where the ball clearly should have been caught ("I got it! I got it! You got it! *thud*) but for miscommunication or collision.

[digglahhh]

More often than not...a flyball that is "between the RF and 2B" is ruled a hit unless there's a very obvious miscue on the part of the second baseman in handling the ball (it hits the heel of his glove and bounces away)...if he gets a glove on it, but just the fingertips or the webbing...it's clearly a hit.

Okay, so its just a pop up to the 2B, SS or a shallow pop directly to the right fielder, whatever. The spirit of the point remains. The scoring is supposed to explain/describe the events that lead to the outcomes, but sometimes the outcomes dictate how the events are described.

The process of official scoring forces all events to fall within a somewhat limited rubric of rulings, this leads to those in decision making capacities to often have to pick not the "best answer available." The quirks involved in the recording of data and classification thereof is one the many reasons why there is really no substitute for intently watching the games.

[538280]

Chris,

I think that this is an extremely interesting topic, especially in light of the discussion we had in the Chipper Jones thread about the subjectivity of official scoring. It doesn't take too much time working with official scorers to realize that you are dealing with an art and not a science. This is one of the reasons why I set the bar higher than most when I'm reading somebody's conclusions driven from PBP data.

Oh, I think I forgot to report on the other part of Wright's study. He also talked about subjectivivy of the official scorers. He mentioned how Cleveland scorers from the period 1976-1987 (when he had PBP) had charged a real lot of errors while other scorers had charged very few.

He also talks about how it appeared Boston scorers were always very favorable in giving Wade Boggs hits. He noticed how very few errors were scored when Wade Boggs was batting as opposed to other players.

I would also advocate the "team error." This is actually something that I have been meaning to start a thread about for a while now. When two or three fielders converge on a pop-up, miscommunicate and the ball drops between them- that is an error! You can't attribute it to one player, so it is not ruled as such, but it is certainly errant. Manny diving and cutting off that throw for no reason- that's an error. Sometimes errors are made, and the fielders are bailed out b/c the runners don't successfully advance, so the is no error charged. A team committing no errors is certainly not automaitcally indicative of them playing expemplary defense. There are certain mistakes you can make and not be penialized statistically for, and others that you can't.

I agree. I remember when I was scoring a game a while ago (don't really remember what game) and two fielders miscommunicated on who would catch the ball and it landed between them. Under the way we score games, there is absolutely no way the batter should be creditied with a hit on that play, yet there is no way to determine who to charge the error to because it was between two players. The player would get a hit he didn't deserve. I like this idea of a "team error".

[west coast orange and black]

the idea of "team errors" is an interesting one to me.
as with some of you guys, at times i have no idea to whom to charge an error. and then at other times it kinda pains me to be forced to pencil in a hit when it obviously wasn't.

sometimes the official rules book takes the pencil outta your hands, so to speak, even though scoring is actually very subjective at times.

[digglahhh]

On Retrosheet there's a huge study (by Ruane, I think) and the goal is to find out whether some players reach on errors more than others and why. In this study the author goes through error rates at particular ballparks and seems at a loss to explain the random jumps in error rates that appear for certain teams in certain years and then come back to earth. I don't really have data on who the official scorers were for which teams and when, I could probably find it at work if I did enough digging, but that may explain some of the differentials when the other variables seemed to remain constant.


I'd go one further than the team error, and sometimes I feel (I flip flop on this) that players should be able to be given errors even if runners don't advance. I mean a screw up is a screw up, the fielder shouldn't necessarily get bailed out because the runner didn't take advantage of it. To me that's kind of like finding out a D student plagairized his paper and not penializing him because the work was poor anyway, you know... Of course those runs wouldn't be unearned. I don't know maybe I'm just a bitter taskmaster, but I waver on this, does anybody else have an opinion?