I think we can all agree that luck had nothing whatsoever to do with this.

I always thought Juan Pizzaro should be in the Hall

Oh noes! Here we goes!

That list tells me that clutch factor is ####### useless.

Did Steve Trachsel ever actually get around to, you know, pitching the freaking baseball in a high leverage situation? 'Cause it sure seemed like he just fidgeted around the mound for an hour or so until he either got taken out of the game or everybody just left the park and they called the whole thing off.

Anything that paints Steve Trachsel in a positive light is crap.

Put Gator in the Hall? Before Steve Trachsel, the clutchiest clutcher who ever clutched? The only way you can possibly justify that is that Trachsel's induction speech would take longer than Nicholas Nickleby.

What's the saying "A stat should show you 80% of what you know and 20% should surprise you" (sorry, I'm butchering that phrase, but you get the point).

Well I know that Steve Trachsel sucks and Suppan is about as league average-ish as you can get. Anything that paints them as the 2 clutchiest pitchers of the last 50+ years, has to be flawed. Back to the spreadsheet, you nerd.

Where's Jack Morris??

Just glancing at the list, it seems to have a lot of groundball type pitchers. So, perhaps it's not so surprising. GB pitchers will have more base runners, creating more high leverage situations. Yet their GB tendency, if they are good at using their sinker, produces fewer HRs in those situations and allows them to get more GIDPs and force outs.

I'm still not sure what the significance of this levERA is. A five percent difference in ERA in high leverage situations? Doesn't seem like a big deal.

I wonder how much this is affected by having a great bullpen finishing up behind the starter. It seems like you could have two pitchers, one with great bullpen support and one with the Royals bullpen, and any time they left in a close game you'd have a high leverage situation. The guy whose baserunners get stranded is going to get a boost to his levERA, and the Royal starter is going to see his levERA plummet like a stone.

Well I know that Steve Trachsel sucks and Suppan is about as league average-ish as you can get. Anything that paints them as the 2 clutchiest pitchers of the last 50+ years, has to be flawed. Back to the spreadsheet, you nerd.

The problem with this is that the baseline is themselves
It doesn't say that Trachsel is "clutch" compared to another pitcher, what it says is that Trachsel sucked less in hi leverage situations than he did in low leverage situations
you can look at his BBREF splits, he gave up a .733 OPS in hi leverage situations and .778 overall

Halladay by this measure is less clutch even though he only gave up a .664 OPS in hi leverage situations, because that wasn't as much as an improvement over the .671 OPS he gives up overall.

BTW Mr. Pitch to the Score (Jack Morris) had a 100 tOPS+ in hi leverage situations, medium leverage situations and low leverage situations
his game margin tOPS+s (tie game, within 1 run. 2, etc,,,) are all around 100 except "late and close" where it was 88 (once again there is no statistical evidence whatsoever, that Morris "pitched to the score"- a talent I believe some pitchers do in fact have - Like Glavine- but Morris the MSM posterchild for the idea did not- basically he was good, durable and very very lucky)

I'm still not sure what the significance of this levERA

simple: Ron Guidry has a good one, so it's a valuable stat

I wonder how much this is affected by having a great bullpen finishing up behind the starter. It seems like you could have two pitchers, one with great bullpen support and one with the Royals bullpen, and any time they left in a close game you'd have a high leverage situation. The guy whose baserunners get stranded is going to get a boost to his levERA

yes, which is why if you are going to do this you should use tOPS+ or something...

BTW- Blyleven did in fact pitch worse in hi leverage situations than he did in low leverage situations, he gave up an OPS of .687 in hi leverage situations versus .668 overall.

Javy Vazquez for his career has given up a .804 OPS in hi leverage situations versus .730 overall

What I want to know is this:
Does that mean Javy "chokes"?
Does that mean he changes his approach- in a detrimental manner- depending upon the situation- and if so can coaching help?
Does this mean that his managers have done a poor job in figuring out when to pull him?
All of the above?

I wonder how much this is affected by having a great bullpen finishing up behind the starter.

It shouldn't be affected at all if the stat is based on the clutch stat at B-R.com. The clutch stat is based on the WPA and WPA/Li figures for a pitcher, which cumulate the incremental win probabilities for a pitcher in a game. If a pitcher leaves a game with a runner on second the pitcher is charged with the probability of that run scoring; whether the run is ultimately allowed or not because of the relief pitcher is irrelevant.

Just glancing at the list, it seems to have a lot of groundball type pitchers. So, perhaps it's not so surprising. GB pitchers will have more base runners, creating more high leverage situations. Yet their GB tendency, if they are good at using their sinker, produces fewer HRs in those situations and allows them to get more GIDPs and force outs.

Again, if I understand the stat correctly, this should be irrelevant. If the LevERA+ stat is based on WPA (as it appears to be), a pitcher who puts runners on but escapes the inning without allowing a run sees no real impact on his WPA for the game - the WPA just zeroes out for the inning, going up for each runner allowed but going back down by approximately the same amount if the run doesn't score. The stat at B-R.com that measures a pitchers performance given the base-out context in an inning is RE24.

simple: Ron Guidry has a good one, so it's a valuable stat

Gator gave up a .662 OPS in hi leverage situations versus .672 overall
He was a good pitcher (at his brief peak he was truly a great dominating pitcher), and he pitched well in pennant races, and to the extent you can statistically show that someone had clutch performances - he did

but his peak was too short and his career was too short- and nothing Tommy in CT is gonna come up with is going to change that

#11, thanks for taking my snarky sarcastic post and logically explaining it.

I just wanted to act like *insert MSM sportswriter* for a moment. It felt good.

This thread reeks of anti-Yankee bias.

The problem with this is that the baseline is themselves
It doesn't say that Trachsel is "clutch" compared to another pitcher, what it says is that Trachsel sucked less in hi leverage situations than he did in low leverage situations

Exactly. The cause is not necessarily better performance in high-leverage situations; it could be worse performance with low leverage. If Trachsel half-asses his job when the leverage is low, then it will appear he performs better in the clutch, when really all he's doing is not-half-assing it.

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

To whoever wrote that article, please never do regression analysis again until you know how to do regression analysis. Please stop interpreting differences in bivariate correlations as if they tell you anything about how important one variable is relative to another in predicting your dependent variable.

I will agree that Ron Guidry belongs in the HoF as much as Vida Blue and Al Leiter do, so you've achieved your goal.

I'm surprised any modern starter shows up on the list actually. They almost never pitch in high-leverage situations.

Each of these pitchers have an ERA+ that can fairly be said to be deceptively low. Another way to put it is that they were better pitchers than their ERA+ figures might indicate.

I didn't RTFA, but if somebody has a high levERA+, and they pitch better in high leverage situations than they otherwise do, doesn't this positively impact their ERA+ because they don't allow runs in those situations? Or are we supposed to ignore the fact that they pitch worse in low leverage situations and get themselves into high leverage situations in the first place. Isn't this like saying that a tightrope walker has a lower risk of death than a normal person does because he dies less often when walking on a tightrope than a normal person would?

If the LevERA+ stat is based on WPA (as it appears to be)

Maybe it is (the article doesn't say how it's calculated and I'm not going to go looking). But then it's being compared to ERA+ rather than, say, component ERA+ or some such.

And, 202 pitchers, 15 of them above 4%. What's the mean (0?) and standard deviation (I'll guess about 2.5%)? Is it roughly normally distributed? What number of pitchers would we expect to be above 4% by random chance alone?

Does this mean that his managers have done a poor job in figuring out when to pull him?

Bobby Cox is the master of getting his pitchers out of the game before they blow up. I don't know how he does it, but it's possible that he just knows when a pitcher is losing it. It seems like most managers only pull a pitcher based on results. A pitcher can be really struggling, but the batters line out, or the guy "gets out of trouble" or whatever. The manager will still put the pitcher back out the next inning. Cox doesn't seem to do this.

It's no surprise that Javy had his best year in Atlanta. Girardi seems to be more typical in this manner, although I have not analyzed Javy specifically for each start. I do know that Hughes seems to only get pulled after giving up runs.

Among the 202 pitchers who've pitched 2000 or more innings since 1952, these are the only 15 with a positive clutch adjustment factor higher than 4.0%.

So, does everyone else have a negative clutch adjustment factor higher than 4.0%? Or a positive clutch factor worse than -4.0%?

EDIT: damn strikethrough malfunction

So, let's see here. This guy makes up a list of 15 pitchers who score really well in a dicey stat and then singles out the TENTH guy on the list for Hall consideration. One of the nine guys ahead of him is in the HoF (Palmer). None of the others are, or are even very close, except for Billy Pierce. This tenth guy ranks behind Juan Pizzarro and ahead of Tom Seaver, which tells you that this stat does not match pitcher quality at all. The author can claim that the list isn't supposed to measure pitcher quality, just a small feature that, at best, adds a footnote to the pitcher's overall quality. But if he claims that, he's just shot the importance of his stat right in the foot. And even if his stat was any good, it would make a HoF case for Pierce or Johnny Podres, not Rod Guidry.

The stat itself is such trash that when I read the article, even I could see that he had no idea how to do regression analysis, and I'm not real good at regressions. Walt Davis was exceptionally polite in comments 20 and 22, considering what he must not just think, but know about the value of this stat.

The author makes the point more than once that this stat correlates better with winning percentage than ERA+. Well, the stat is designed to locate exactly those pitchers who were unusually lucky in close games. Of course the stat correlates well with winning percentage. It is designed to locate exactly those pitchers who have been luckier at winning than their ERA+ would suggest. I believe this is called "selection bias", although there may be a special term for this particular failure.

I would also like to note that Billy Pierce has many many more IP and decisions than Ron Guidry. In spite of his much longer decline phase, Pierce has exactly the same career 119 ERA+ that Guidry has. Pierce ranks higher on this list than Guidry does, meaning that the difference in winning percentage between the two is not a result of clutch pitching by Guidry, but instead is the result of pitching for the Yankees instead of the Chisox. In short, if this list does anything, it enhances Pierce's HoF candidacy, not Guidry's.

Not to be overly blunt, but this is the kind of nonsense that makes people hate the Yankees, or at least Yankee fans. A bad list where the author singles out the tenth guy for the HoF because he's a Yankee, when he could make a much better case for a guy who played somewhere not New York. I wonder if the author even knows who Billy Pierce was, since he wasn't a Yankee. - Brock Hanke

I read the title and the list and decided not to click the link, I don't think I'm missing anything.

Just for laughs, I looked at another list that Steve Trachsel is near top of, career HR/9. There's quite a bit of overlap between that list and this one.
Steve Trachsel: 1 and 2 (Woody Williams snuck past him at the very end of his career for the #1 spot)
Pat Hentgen: 2 and 9
Jeff Suppan: 3 and 4 (and moving up fast)
Pedro Astacio: 7 and 6
Bartolo Colon: 14 and 14

Where can I find a definition for LevERA+? I've searched Google and looked on B-R and Fangraphs to no avail. Any help?

The level of discourse whenever "clutch" comes up, especially on threads about Guidry or Blyleven, is embarrassingly poor. Many people are deriding the stat without knowing what it is, either due to the name, or the results, or the source.

The stat itself is such trash that when I read the article, even I could see that he had no idea how to do regression analysis, and I'm not real good at regressions.

What is the stat? We should at least know how it's calculated before calling it "trash".

Walt Davis was exceptionally polite in comments 20 and 22, considering what he must not just think, but know about the value of this stat.

Walt Davis admitted that he doesn't know how the stat is calculated and wasn't going to try to find out. That's not polite; that's lazy. There's a total double standard on this site.

Walt Davis admitted that he doesn't know how the stat is calculated and wasn't going to try to find out. That's not polite; that's lazy. There's a total double standard on this site.

Why should Walt? It's not his job to chase down every garbage stat that finds it's way onto the internet.

I read the title and the list and decided not to click the link, I don't think I'm missing anything.

It links to exactly where you think it links. It's like deja vu all over again and such.

What is the stat? We should at least know how it's calculated before calling it "trash".

Walt Davis was exceptionally polite in comments 20 and 22, considering what he must not just think, but know about the value of this stat.

Walt Davis admitted that he doesn't know how the stat is calculated and wasn't going to try to find out. That's not polite; that's lazy. There's a total double standard on this site.

Usually I'd agree. but in this case the burden of proof is on the author. Extraordinary claims require extraordinary evidence. Trachsel and Suppan have to be among the worst of the 202 pitchers to amass 2000+ IP since 1952, and they rank as the best? I mean, I don't necessarily doubt that they do rank 1-2 in this stat. But since they do, to claim that this stat has value in telling us anything meaningful, the burden is on the author to be a hell of a lot more transparent about his methodology. Failing that, I feel free to dismiss it out of hand.

Where can I find a definition for LevERA+? I've searched Google and looked on B-R and Fangraphs to no avail. Any help?

Everything I can find using a fairly general search eventually links back to the Gator Blog.

There's a link available on this page, which is supposed to provide a definition, but it currently links to nothing.

Where can I find a definition for LevERA+?

Here (in case it's not obvious; this stat was invented by the author of this Gator blog thingy).

EDIT: My link is the same as Ryan's second link except that his link has a stray "not" in it.

Lies, damn lies and statistics.

As pointed out several times above, all statistical analysis has to at least pass some kind of smell test and any analysis that puts Trachsel at the top of any baseball achievement list is intuitively flawed.

Though I must say when I saw the headline about Gator, I though oh no, another Tommy in CT special...

Trachsel and Suppan have to be among the worst of the 202 pitchers to amass 2000+ IP since 1952

The worst pitchers (by ERA+) with at least 2000 IP since 1952.

Okay, I tried to look at this a little bit. First, I can't replicate his +7.5% number for Steve Trachsel. I get +6.7%. But second, I think he's double-counting "clutchiness" here.

What he's doing is adjusting ERA+ by looking at Fangraph's "clutch" statistic, which is the difference between WPA and WPA/LI. But just looking at Trachsel's WPA and WPA/LI numbers, I think the extent to which his WPA is better is already being reflected in the actual number of runs he allowed.

Trachsel in his career had an ERA+ of 99 in 2,501 innings, which, if I do the math means Trachsel allowed something like 12 runs more than an average pitcher in his career. This is roughly consistent with Trachsel's career WPA on Fangraphs of 1.10 (wins) (so, something like 10-12 runs above average). But it's clearly nowhere near consistent with Trachsel's career WPA/LI on Fangraphs of -8.75 wins, or something like 80 runs below average. So, it looks like Trachsel was "clutchy" in his career because he gave up fewer runs than we would have expected, but that's already being reflected in his ERA and ERA+.

Now, having said that, there also looks like, at least in Trachsel's case, there's something systematic about his "clutch" scores. He has a positive clutch score in 14 of his 16 seasons. Now, I suppose one could argue that, well, of course the "clutchiest" player of the last 60 years was "clutch" every year. But Steve Trachsel has a "clutch" score of +0.5 wins in 1999 when he had a W-L record of 8-18 (admittedly with an ERA+ of 81, so he deserved his crappy record that year).

I mean, how is it that the clutchiest pitcher of the last 60 years, with an almost perfectly-average ERA (career ERA+ of 99) managed to have a career winning percentage of .474?

So, let's see here. This guy makes up a list of 15 pitchers who score really well in a dicey stat and then singles out the TENTH guy on the list for Hall consideration.

For what it's worth, I think you're misinterpreting this because you're not familiar with the site. He's not singling out Guidry for the Hall because he finishes 10th in this stat. The site is "Put Gator in the Hall," dedicated to a single noble cause. Gator's spectacular finish in this stat, in addition to his many other Hall-worthy accomplishments, is just another reason why he should be Cooperstown bound.*




* Some folks not posting from inside a CAT Scan machine seem to have a bit of a problem with the basic premise.

Why should Walt? It's not his job to chase down every garbage stat that finds it's way onto the internet.

No, it's not his job. But why should he? Because he's commenting about this metric here (and not just Walt -- that's true of a number of posters in this thread). People are calling this stat crap without even knowing specifically what it measures, how it's calculated, etc. Everybody just keeps cracking jokes, patting each other on the back, etc. I'm starting to wonder if this is parody.

That said, Walt wasn't really talking about LevERA+ or the clutch index; he was talking about the regressions in the article (and the conclusions drawn), and he was right that they were a total load of crap. I was responding more to Brock's post which called Walt posts "polite", but it wasn't really fair of me to direct my response towards Walt.

But Steve Trachsel has a "clutch" score of +0.5 wins in 1999 when he had a W-L record of 8-18 (admittedly with an ERA+ of 81, so he deserved his crappy record that year).

Even in 1999, Trachsel pitched better in high-leverage situations (321/450 allowed) than otherwise (roughly 330/455 allowed).

But just looking at Trachsel's WPA and WPA/LI numbers, I think the extent to which his WPA is better is already being reflected in the actual number of runs he allowed.

I think this makes sense. WPA/LI essentially reduces each play to a neutral leverage; runs allowed doesn't. And it seems that in Trachsel's case, runs allowed properly reflects his value. His clutchiness wasn't based on score/inning, but rather on base-out situations.

If we tended to use OPS+ allowed or something similar as our key metrics for pitchers, then LevERA+ would be a more useful addition (assuming that Kiko's inferring things correctly); but since we look at runs already, we're probably already capturing a lot of that clutch performance.

Thanks, Kiko, for the substantive criticism of the metric/post.

You haters can continue your little circle jerk. Me, I'm going to break out my old VHS tape of Louisiana Lightning's unbelievable 18-K game against the Angels. The ninth inning still gives me chills, just hearing the crowd chant "LevERA+, LevERA+, LevERA+."

SOSH says, "For what it's worth, I think you're misinterpreting this because you're not familiar with the site. He's not singling out Guidry for the Hall because he finishes 10th in this stat. The site is "Put Gator in the Hall," dedicated to a single noble cause."

Thanks! I had no idea where this stuff was coming from. No idea that this was almost certainly doomed to be circular reasoning or selection bias of some sort. I just thought someone had gotten way off the reservation about Guidry. I didn't know that he had been off the reservation for some time. The obvious comparison to Pierce was what really set me off.

I'm glad Harold added that he was directing the comment towards me, not Walt. Walt really really knows his math. If he says, "To whoever wrote that article, please never do regression analysis again until you know how to do regression analysis," that's going to absolutely mean that he found a regression analysis in the computation of the stat and he concluded that the regression analysis was horrific. That is, he saw enough to evaluate the stat's regression analysis, no matter how little he had to read to figure it out. At his level, he may have figured it all out by sentence 2. I don't know; he's way over my head. I've read Walt assassinating a bad stat. He let this guy off easy. Trust me, he was polite. If I knew enough math to be as certain as Walt is about things like that, I doubt I could be as polite as Walt generally is, much less this. When I do run into something in a specialty of mine, I often am not as polite as Walt is about baseball math. I get frustrated when I catch amateurs posing as professionals. Walt is much more forgiving.

Even in 1999, Trachsel pitched better in high-leverage situations (321/450 allowed) than otherwise (roughly 330/455 allowed).

But this is a fundamental problem with the stat. By comparing it to the player himself, rather than the league, it basically sets up a situation for a guy who sucks to score well.

I understand your point that people seem to be jumping to conclusions here but it's not like this is some first time post that's been linked or someone with a track record of insightful commentary. This guy has an agenda and consistently pushes it with questionable statistical "evidence." It's like Scott Boras' annual "player X is as good as much better player Y" binders that go out. At some point the burden is not on the reader to challenge the assertion but on the producer of the stat to demonstrate that it is meaningful.

Look at it a different way, if Rich Lederer comes up with something that demonstrates Bert Blyleven to be the 5th best pitcher ever you should start out skeptical given Lederer's obvious interest and support of Blyleven's HoF case. That doesn't mean he cannot be right, but he should be challenged. This guy has created a stat with a very basic flaw (comparison to self rather than league) that even I as a non-math guy can immediately see.

It's a stat with an obvious math flaw and a non-sensical conclusion.