Loading ...Time.com has an interesting Q&A with Bill James. Here’s the part I found most interesting:
You often show that conventional baseball statistics aren’t as important as they appear. In the book, you write “every year that passes, the ERA (Earned Run Average) becomes a little more irrelevant.” Why is that?
The reason the ERA is becoming a little more irrelevant every year is that pitchers don’t pitch whole innings anymore. Relief pitchers anyway. If you go back to 1915, 1920, really, all pitchers pitched full innings 99% of the time. And you could measure a pitcher’s effectiveness by how many runs he allowed in those whole innings. But modern pitchers, in particular modern relievers, pitch portions of an inning. And in a situation where each pitcher pitches a portion of an inning, who you charge the run to becomes critical. And the rule on whom we charge the run to is so careless and sloppy that it doesn’t work. It often leads to pitchers having ERAs that do not reflect how they really pitch, either because the reliever allowed a bunch of runs to score that were charged to somebody else, or because the starting pitcher who left guys on base got hurt by it.
Plus the fact that defense isn’t even considered.
All aboard the DIPS train.
I wonder if Dusty Baker qualifies for any of James’ awards?
Jim Kern was saying the very same things in 1982. Glad Bill James finally got with the program.
I can understand the sentiment that ERA is overrated as a pitching metric (it is), but why should inherited runners who score be counted against a relief pitcher? What if a someone comes in with the bases loaded and strikes out the first two batters he faces and then allows a boop single to score a couple runs? Why should the first pitcher be given leeway in how many baserunners he gives up, but the guy who comes in after him be expected to be perfect? In my mind it is perfectly reasonable to charge the runs to the guy who let them get on base, especially if they got into scoring position.
I think the point is that, unless a pitcher pitches the whole inning himself, there’s really no satisfying way to assign runs.
I agree with Dan. We’re not making moral judgments here, just find the best way to evaluate pitchers.
Didn’t Bob Howsam’s no-beards rule cause Kern to ask for a trade?
I just looked at his game log for ‘82. In his first three outings as a Red, he entered the game with the bases loaded. Guess he had Johnny Mac’s confidence.
I remember those games well. Kern was a disaster. No matter how bad things get these days, they can never top ‘82. It was just awful, and Kern set the stage right away. If memory serves me correctly, in the first week or so, Kern had something like 78 inherited runners to score and about 27 wild pitches. One game he had 2 or 3 wild pitches in the same inning that caused yet another loss and people were going nuts because the new closer (Kern) couldn’t throw a strike and the new catcher (Trevino) couldn’t catch.
Yeah, Kern raised hell about the beard thing. He was really, really unhappy and harping about it every day. he said he looked like Big Bird without his whiskers. Funny thing, he really did.
He finally talked his way out of town, helped by his lousy pitching. But before he left, he gave Tim Sullivan of the Enquirer a really terrific and insightful piece on new stat standards for relievers and why the old stats were out of date.
Wish I had saved that piece. Some really thoughtful perceptions by Kern, who I heard has a book about his time with the Rangers.
What’s considered the best barometer for a relief pitcher? Inherited runners/inherited runners allowed to score? Is that a readily available stat somewhere?
Yes, that number is around but not all these free stats sites list it. I believe the IR-IS stat was the premise of Kerns’ theory. He had a formula and it made lots of sense. I gotta find that article.
I think they’re all flawed in some way, much like defensive stats. You have to look at everything. For example, straight IR-IS is good, but is flawed b/c it treats “2 outs, runner on 1st” the same as “0 outs, runner on 3rd.”
The Win Probability Added stuff over at FanGraphs.com might be the best, especially if you’re only asking about what has happened, and aren’t concerned with the predictive value of the performance.
Prospectus has “Reliever Expected Runs,” which I don’t even begin to understand.