I'm an easy sell on this one. Wolverton does a nice job of explaining why.

Haven't read the article yet, but I've thought about it some. I can go either way on this.

Intellecutally I know RA is a better measure, since errors are ridiculously assigned, and there's no pitcher-responsibility difference between a dropped grounder and "past a diving Jeter". Moreover, RA will map more closely to team wins and losses.

That said, I have a hard time giving a statistical penalty to a pitcher when an easy double play ball to Knoblauch becomes an RBI triple. It seems like there are so many anecdotal cases that run contrary to the general statistical and intellectual awareness I have.

The best approach, obviously, is to just start removing pithcer responsibility for runs of any sort, focussing on DIPS stats to assess pitchers. If we could do that, I'd be comfortable with using RA over ERA as a semi-interesting peripheral stat. But until pitchers start being evaluated that way, I'm not sure I wouldn't prefer sticking with ERA over RA.

I use RA for handicapping.

What is needed, if RA is to take flight, is a Defence Factor, to be applied in a similar manner as Park Factor (except, of course, to all innings, not just home innings).

Over the past few years I have noticed much more "homerism" coming into scorers decisions. We have all seen, on numerous occasions, an error scored as a hit so the star home-town hitter pads his average and does not have to yell at the scorer after the game. Errors are unreliable due to the human factor. We regularly hear scorers talk of plays that should have been made. That is often code speak for errors.

I remember hearing of pitchers calling scorers after the game to attempt to get the scorers to change their decision so as to make runs unearned. I am sure some bauxites will remember who is alleged to have done that.

I agree, banish unearned runs.

To me the best solution is the 5 man umpire crew (which has been talked about in the past). Basically it would go home plate, then to the press box to track the stats [including errors] and perhaps to run QuesTek [hmm...might change the perception eh?], then to third, then to second, then to first, then home again.

That would kill off homerism and cut down on players directly complaining (don't want that ump using it against you next time you are at the plate) plus with the QuestTek addition it would cut down on umps complaining about the system not being set right. Probably would result in a lot faster calls on errors too.

Why not even be more radical and ditch "errors"? In most sports, if you score by fluke or opponent's mistake, you still get credit for it. Soccer is a notable exception.

Ditch errors, and give a batter OBP credit for putting the ball in play (while punishing the pitcher for allowing a ball in play).

Catcher's throw sails into centre on a steal attempt? Give 'em two stolen bases, same play. After all, it was the runner's risky and aggressive play that got him to third; but for his attempt, the catcher would never have hurled the ball into the outfield.

Errors are so subjective anyway. If it means that we have to evaluate defence based on advanced metrics, or -- heaven forbid -- with our eyes, then so be it. It would bring the mainstream baseball community on board in the Elusive Quest For An Accurate Defensive Statistic.

Mike, see this article for Tangotiger for the beginning of "defensive evaluation using our eyes" as you put it. Very interesting food for thought.

Changing the way earned/unearned runs are scored has been one of my causes for several years now. I was actually thinking about bringing this up yesterday after Hentgen's lousy start. He pitched a lot worse than his current nifty 1.59 ERA would indicate, and he has this archaic and idiotic rule to thank for that. It's silly to pin the blame for Munson's home run solely on Carlos Delgado, but that's what the game's scoring essentially did.

I remember having a debate on this issue in another forum back in 1999 after this Roy Halladay start. Halladay was awful, allowing five hits, four walks and four runs in four innings. But Halladay didn't allow a single earned run and his start was subsequently viewed as a success by some. It's like when a really bad hitter manages to rack up a lot of RBIs -- a person can watch every single game a horrible player is involved in, but the perception of their performance can be coloured by one misleading statistic.

It's time this silly scoring rule started getting attention. The league can't get rid of it soon enough as far as I'm concerned.

And if they're NOT getting rid of errors, start giving more of 'em to outfields for heaven's sake. An outfielder loses a ball in the sun/roof? Too bad-error. That's always been the worst call ever-the pitcher induces a lazy fly ball and gets whacked for a double-and likewise lucky for the hitter. Two guys have a communication problem and the ball falls? Give 'em both errors.

The other problem with ERA (and RA) is the failure to apportion the runs scored by runners left on base at the time of a pitching change.

For me, the most interesting description of a pitcher is simply a listing of opposing PA outcomes. Halladay's line for last year might read something like this: 20% strikeouts, 3% walks, 3% home runs, 40% ground balls, 18% fly balls, 10% line drives, 6% pop-ups. If you combine these numbers with opposition slug and OBP, you've got a really nice portrait of a pitcher and how he interacts with his defence.

And if they're NOT getting rid of errors, start giving more of 'em to outfields for heaven's sake. An outfielder loses a ball in the sun/roof? Too bad-error. That's always been the worst call ever-the pitcher induces a lazy fly ball and gets whacked for a double-and likewise lucky for the hitter. Two guys have a communication problem and the ball falls? Give 'em both errors.

Coach's "E-10" would take care of this.

The best explanation I've heard as to the paucity of errors given by official scorers came from none other than Buck Martinez. According to Buck, if you score a play an error both the batter and the fielder will be mad at you. However, if you score the play a basehit then only the pitcher will be angry at you.

The best approach, obviously, is to just start removing pithcer responsibility for runs of any sort, focussing on DIPS stats to assess pitchers.

I use K/9, BB/9, and HR/9 for pitchers like a lot of people use ba/obp/slg for hitters.

I can't imagine that errors and unearned runs will ever be eliminated.

Mind you, for some reason the Japanese have a more pro-pitcher Earned run Rule as the status of a run is determined right at the moment it's scored. Say, the leadoff batter triples and scores on a passed ball on the very next pitch - unearned run no matter what.

And as for a Pete Walker update, his BayStars are suddenly tied for first with a 4-2 record coming off a miraculuos sweep of defending champ Hanshin. The last two wins were shutouts : 17-0 (!) and 4-0.
Walker's turn on the mound is somewhere this weekend against Hiroshima.

In principle more information is better than less. Perhaps RA and ERA can be used simultaneously. Their difference can be compared on a pitcher by pitcher basis, to see what pitchers are perhaps more susceptible to defensive lapses of various sorts. The obvious extreme case mentioned in the article is knuckleballers (passed balls and so forth), but in practice I would think the differences would show up between ground ball pitchers and fly ball pitchers, with the latter being more dependent on good infield defense.

In any case, when evaluating a given pitcher, a high RA - ERA differential would be a red flag to investigate further. Maybe it's because of bona fide poor defense, i.e. it truly isn't the pitcher's fault. Maybe it's because the pitcher works slowly and puts the defense to sleep. Or maybe it's because the pitcher loses his head when a defensive mistake lets a runner on, and this leads to poor pitching and a bunch of nominally unearned runs being scored. In any case, for any given pitcher, it seems like something worth looking into.

Oh well, Walker indeed made his second official start in Japan here on Friday and stunk big time.

 
Yokohama  BF  NOP    IP  H  R ER  HR BB SO   ERA
Walker    14   59   2.1  7  6  6   3  0  3  7.56(season)

Hiroshima leeds 6-2 thru 7 at the time of the writing.

Back to topic, I agree with gid - the more info the better, especially the one that has a century long tradition to it. Make up new stats that better fit the player analysis as you wish, but maintain a sort of backward compatibility.

Finally, a remark to John Northey's idea, that I've seen too many times (rsbb?) and just plain disagree.

My opinion is probably biased by my domestic experiences here in Russia, but I think the umps just don't care about this part of the game. Quite a few of them consider themselves the most important ones on the field (Cowboy West comes to mind), make ignorant calls afield and will sure be even worse in the booth.
Besides, the umpires are the ones I think who can long have grudges against a particular player. ML umps make too much money to give a damn about some scoresheets.

The question is, what is a better measure of a pitchers ability - ERA or RA? I think neither is significantly better than the other. ERA takes some of the defense out of it - after all you can't get charged with an error if the play is unmakeable. In my estimation, about 70 percent of error calls are indisputably correct, 20 percent are borderline and 10 percent are crazy-wrong.

Conversely, using RA to evaluate pitchers is patently unfair - since what is evaluated is overall pitching-defence, of which the pitcher is responsible for about 70%. With ERA, that percentage of responsibility goes up to about 75%; with DIPS or FIPS it might go up to 90%.

Use your favourite defence-independent measure, making sure to include stolen bases and caught stealing against, while mixing in a little regressed RA.

There's zero chance of that kind of stat finding its way onto television screens, so I say, why ditch ERA if you aren't replacing it with something better?