Cy might have been able to sustain that towards the latter part of his career, wasn't babip below .250 some of the deadball years? But he wouldn't be able to pitching in today's conditions.

But .217, really? You’re smarter than that.

Sir, you flatter me!

Has anyone looked at Cahill's pitch fx data? I tried finding it at fangraphs but didn't have any luck with what I was looking for, which is, does he really throw a sinker and, if so, how much it sinks and what are the results of the pitches he throws that do sink. .217 is probably a lot of luck, but I think the question of what kind of pitches lead to a greater % of outs when they are put in play is interesting.

Preparatory to a larger (and future) discussion about how we should talk about Cy Young candidates, Dave Cameron introduces a highly relevant subject:

Cahill's BABIP is obviously not sustainable. He projects to have a near-average BABIP and a higher ERA going forward. But, in the context of the CYA, who cares what he projects to do going forward? Jered Weaver projects to have a lower K rate and worse ERA going forward; should we value his 2010 season with a regressed K/9?

Awards are given for past performance, not future projection. The only thing that should matter is determining how much of Cahill's low BABIP is due to his defense and park.

Has anyone looked at Cahill's pitch fx data? I tried finding it at fangraphs but didn't have any luck with what I was looking for, which is, does he really throw a sinker and, if so, how much it sinks and what are the results of the pitches he throws that do sink.

I don't know where to see a list of how much his sinker sinks compared to other players. But, according to Fangraphs, Cahill's sinking fastball has been the most valuable pitch in the AL this year in terms of batted ball outcomes. According to the same measure, Gio's had the best curveball in the majors.

I don't know where to see a list of how much his sinker sinks compared to other players. But, according to Fangraphs, Cahill's sinking fastball has been the most valuable pitch in the AL this year in terms of batted ball outcomes. According to the same measure, Gio's had the best curveball in the majors.

This is what bugs me about the waving away of Cahill's performance using babip. Maybe it is just luck, or maybe there's something to be learned from Cahill's success.

On a more fanboy level, if the A's can just find a couple of bats to support Anderson, Cahill, Gio, Braden and Mazzaro next year...

But, in the context of the CYA, who cares what he projects to do going forward? Jered Weaver projects to have a lower K rate and worse ERA going forward; should we value his 2010 season with a regressed K/9?

Yes. I hate, hate, hate the use of predictive stats in award discussions.

Was it sustainable when DiMaggio hit in 56 consecutive games, or Williams hit .400. Of course not. But no one suggests regressing BABIP or HR rate for hitters.

If a guy goes 25-5 with a 2.00 ERA, and has a 4.0 K/9 and 4.0 BB/9 and a 0 HR rate, he's the CY Young regardless of what you project for next year.

FIP, xFIP, tRA, etc. are predictive stats, not value stats.

Continuing on the fanboyish level, I saw an interesting list of comparisons by a commenter on the A's Scout board yesterday:

Cahill-Hudson
Anderson-Mulder
Gio-Zito
Braden-Lilly
Mazz-Lidle

It's not perfect in terms of styles, but it's also not that far off. Of course, these are ceiling projections...

So keep Vin Mazzaro as far away from single-person aircraft as humanly possible?

Cahill-Hudson
Anderson-Mulder
Gio-Zito
Braden-Lilly
Mazz-Lidle

It seems to me that, while that is an interesting-ish list, Hudson was significantly better than Cahill, Zito was significantly better than Gonzalez, and Lilly was significantly better than Braden.

Sounds like you guys prefer the BB-ref WAR calculation in giving credit for what happened. I do to, otherwise I would have come up with a different system. Here's what I wrote on the last Cahill thread which got buried:

On BB-ref the pitching WAR leader is Felix Hernandez. I don't vote on straight WAR. There are 11 guys within 1 WAR of Felix, so that would be my candidates group. The only Mariner games I've watched this year are the ones they play against the Angels, so I've seen Felix pitch 3 times, and all three times Jered Weaver (another candidate) beat him.

But I've got to put my biases aside and endorse Felix for this award. It's not just the ERA, FIP, or WAR calculation. What really stands out is how much of a horse he's been. Leads the league in innings by 17, and about 30 ahead of Weaver, a great pitcher in his own right who has not missed a start. Among the two with lower ERA, Felix has 65 more innings than Buchholz and 50 more than Cahill. Felix doesn't have the wins, but it seems like the award voters have stopped worrying about that, punishing him for having historically bad offensive support, and paying attention to who has the best pitching numbers. Last year Felix had the wins and they gave the award to the best run preventer in Greinke.

I think it's Felix's turn to win this one.

It seems to me that, while that is an interesting-ish list, Hudson was significantly better than Cahill, Zito was significantly better than Gonzalez, and Lilly was significantly better than Braden.

Right. As I wrote: "Of course, these are ceiling projections..."

Sounds like you guys prefer the BB-ref WAR calculation in giving credit for what happened. I do to, otherwise I would have come up with a different system.

Can you remind me how you go about adjusting for fielding?

So keep Vin Mazzaro as far away from single-person aircraft as humanly possible?

Single-pilot. I think the plane he was in seats four.

The more important thing is to watch the damn weather and have a contingency plan ready before you need to actually implement the contingency plan.

Can you remind me how you go about adjusting for fielding?

Right, we adjust for ballpark and we adjust for run support to try and get a neutral context between pitchers. If his BABIP is driven by excellent defense behind him, shouldn't we adjust for that?

If his BABIP is driven by excellent defense behind him, shouldn't we adjust for that?

How do we know how much of his BABIP is driven by excellent defense and how much of it is driven by balls that were easily fielded?

Is Trevor Cahill genuinely benefitting from great defense, where the A's defenders are making plays behind him that other defenders wouldn't make? By all means, that should be adjusted for on Cahill's record.

Is Cahill just getting lucky in that the balls in play he's yielding are going straight at defenders (more than the average pitcher)? Lucky, but only partially the product of his defense? How do we adjust for that.

Or is Cahill producing more weakly hit balls, chances that any defense would routinely turn into outs? I think that should be credited to Cahill entirely, whether it's true skill or the lucky result of hitters just making bad swings against him.

#13 That's what Sean's WAR does and it makes total sense to me to at least make a quick and dirty effort to account for the defensive difference between pitchers. But for award voting I only adjust for how good the team was generally, not how effective they were for the specific pitcher.

And I'm voting based on actual runs (after adjusting for park and defense. And if somebody can show he faced a truly unusually tough or easy set of opponents I'm all for adjusting for that too). I'm not particularly concerned about the number of runs you'd expect from a given stat line. Clutch pitching may be mostly luck (or not -- I'm not aware of any good studies on the subject), but it's luck that has value.

I don't agree that FIP is purely a predictive stat. It does say something about the value the pitcher provided to his team. How much value is debatable, but I don't think it should be ignored. If two pitchers have identical 2.50 ERAs, but one has vastly superior peripheral stats, I think it's fair to argue that the one with the better FIP is providing more value to his team. It's not an ironclad rule, and FIP shouldn't end the discussion, but it's relevant IMO.

This is very different from Dimaggios 56 game hitting streak or Maris's 61 home runs. In both of those cases, the value the hitters provided to their teams was obvious, even if the performance wasn't sustainable. Assigning value to pitchers with mediocre peripheral stats is a lot more difficult.

Cahill's BABIP is obviously not sustainable. He projects to have a near-average BABIP and a higher ERA going forward. But, in the context of the CYA, who cares what he projects to do going forward?

If something is not repeatable, we're not sure it was actually attributable to that person. That's why we give credit that way.

If something is not repeatable, we're not sure it was actually attributable to that person. That's why we give credit that way.

No one thinks Jose Bautista's .346 ISO is repeatable. Should he not get credit for all those homers?

No one thinks Jose Bautista's .346 ISO is repeatable. Should he not get credit for all those homers?

Offense and pitching are very different. Those HRs provided obvious value to the Blue Jays, and no other member of the team participated in the HRs in any way.

No one thinks Jose Bautista's .346 ISO is repeatable. Should he not get credit for all those homers?

I think that's a good question. In the awards context, we don't tend to give full credit for unsustainable numbers like that. Guys who have fluke years are usually passed over in favor of stars when it comes time for MVP and All-Star voting.

#19 makes a good point about offense vs. pitching. There's also defense stats to throw into the mix, which I think get treated waaaay differently than offense. For some reason, many people want to fully regress single-season defensive numbers, but then combine them with non-regressed offensive numbers. Sure, the offensive numbers need LESS regression, but they're not perfect either.

I don't agree that FIP is purely a predictive stat. It does say something about the value the pitcher provided to his team. How much value is debatable, but I don't think it should be ignored. If two pitchers have identical 2.50 ERAs, but one has vastly superior peripheral stats, I think it's fair to argue that the one with the better FIP is providing more value to his team. It's not an ironclad rule, and FIP shouldn't end the discussion, but it's relevant IMO.

I don't agree.

If you want to adjust for defense, and can figure out a meaningful way, fine. Park adjustments make sense too, of course, but we already have that in ERA+ or RA+.

FIP takes away too many things that are really part of pitching. Take two identical pitching line (A & B) and convert 2 K's per game to pop-ups for pitcher B. FIP tells you A is better, and that's just wrong. Likewise, pitching better with men on base, inducing DPs and controlling the running game are real pitching skills, even if they are prone to randomness and hard to predict.

From a value point of view, how you get your outs is irrelevent. The only relevent factor is if your defense is converting a disproportionate number of outs.

Offense and pitching are very different. Those HRs provided obvious value to the Blue Jays, and no other member of the team participated in the HRs in any way.

And a 12H, 6 BB, 2 K shut-out provides obvious value to a team. Unless the defense is making extraordinary plays behind him, you still have to credit that value to the pitcher.

Pop-up, walk, GIDP is just as good as fly-ball, single, K, K in terms of value, and saves a bunch of pitches.

It feels like there's a lot of reasonable explanations for why Cahill wouldn't necessarily be as valuable as it seems, but no actual numbers yet devised to support those explanations.

I guess I'm willing to believe Cahill hasn't been as valuable this year because he's gotten especially lucky with his defense behind him, or lucky on the balls he's allowed into play, but I need something more concrete before I turn my eyes from that shiny 168 ERA+ he's sporting.

I mean I've watched him pitch a lot, and he's tough to hit. A lot of the balls he allows in play are weakly hit grounders and crappy pop ups and lazy fly outs. But obviously he's not this good. I can grant you that. But I'm still of a "take the numbers as they are" mindset until someone can show me specifically how (or roughly how) lucky he's been, beyond just BABIP. Because right now I agree with the sentiment that if some guy goes 30-2 with a 1.20 ERA because he fluked into the greatest BABIP of all time, well, I won't draft him in fantasy the next season, but I'll give him the Cy Young.

So are you saying that ERA (in context) is the single most important stat for determining the value of a starting pitcher?

Those HRs provided obvious value to the Blue Jays, and no other member of the team participated in the HRs in any way.

I'd disagree. Its easier/harder to hit depending on the situation (runners on base, score, opposing pitcher etc), and the situation has to do with the teammates.

So are you saying that ERA (in context) is the single most important stat for determining the value of a starting pitcher?

Well, to be honest, I don't know. It doesn't seem bad. But I do agree that FIP and such seem more predictive than value-based. I know FIP is calculated into WAR, which makes me hesitant to put a ton of stock into it for something like Cy Young. I'm open to being told I'm wrong, though.


edit: I guess the way I interpret it is that, for instance, factoring FIP into WAR is kind of in a way saying that being lucky provides no value. To which I disagree, it just doesn't provide reliably repeatable value. But the value is there.

Offense and pitching are very different. Those HRs provided obvious value to the Blue Jays, and no other member of the team participated in the HRs in any way.

I don't see how this is an argument for using FIP. Preventing hits and runs provides obvious value to a team. The question is how much of that value should be attributed to the fielders. FIP assumes that a double off the wall has the same value as an infield fly. We know that's not true. In terms of value, inducing an infield fly is 99% as valuable as a strikeout. Inducing easily fielded BIP provides real value to the pitcher's team, regardless of whether the pitcher was lucky to have induced it or not. Cahill has allowed very few linedrives this year. Again, that has real value.

FIP also assumes that BB/BB/BB/HR/Out/Out/Out has the same value as HR/BB/BB/BB/Out/Out/Out, but we also know that's not true. Those two sequences might not tell us anything different the pitcher's skill, but the latter is far more valuable to the team.

I think that's a good question. In the awards context, we don't tend to give full credit for unsustainable numbers like that. Guys who have fluke years are usually passed over in favor of stars when it comes time for MVP and All-Star voting.

Taken to the logical extreme, every performance beyond a player's forward-looking projection is "flukey" and "not repeatable." Should we regress all over-performance back to projections? Regress it halfway?

There's also defense stats to throw into the mix, which I think get treated waaaay differently than offense. For some reason, many people want to fully regress single-season defensive numbers, but then combine them with non-regressed offensive numbers. Sure, the offensive numbers need LESS regression, but they're not perfect either.

I think people do that with defense because they're worried about measurement error. It's not that Player's UZR in 2010 is better/worse than his true talent, it's that Player's UZR in 2010 isn't necessarily indicative of his actual performance in 2010.

I guess I'm willing to believe Cahill hasn't been as valuable this year because he's gotten especially lucky with his defense behind him, or lucky on the balls he's allowed into play, but I need something more concrete before I turn my eyes from that shiny 168 ERA+ he's sporting.

Well, that ballpark supresses batting average more than almost any other ballpark. So there's that. And then I think most defensive stats will show their team defense has been good - in fact, their DRS and UZR are both among the best in the majors. So there is without other factors, beyond luck or his own skill, reflected in his BABIP.