MattS wrote:
Lee: 3.73 combined (4.07 CLE, 3.05 PHI)

Bastardo: 4.69 (but that's a pretty small sample size...but I guess a benefit of SIERA is that it will stabilize quickly like QERA)

Kendrick: 4.92, 5.30, 4.26 in 07, 08, 09 (though the sample size plus pitching in relief makes the 09 number not really all that great)

thanks for the quick answers matt. very nice new stat.
hamels better than lee. cool.
i asked for both bastardo and kendrick just for the fun of comparing small samples... well, that and the fact that they look to be slotting into key roles in the first part of 2010.

that some of kendrick's relief appearance were pretty long, together with the use of a quickly stabilizing metric together with squinting cross-eyes at the pitch/fx data gives me some hope that the improvement will carry over into the 5th starter role this season. i'll keep my fingers crossed.

So Lee on the Phils was better than full-year Halladay. Here I was kind of hoping (unfairly) that Halladay would just go T-1000 and blow that away.

Barry Jive wrote:So Lee on the Phils was better than full-year Halladay. Here I was kind of hoping (unfairly) that Halladay would just go T-1000 and blow that away.

Remember, Lee might not pitch at all this year. "Minor" surgery could mean anything.

" For the same reason, pop-up rate was allowed to negatively affect SIERA because it is a symptom of the pitcher throwing the ball that generates an upward trajectory, which could lead to an increase in home runs. "

Matt, I'm only a candidate for Matt-pre-skool, but this assertion caught my eye. "symptom that generates an upward trajectory", but here's my humble question -- is it reasonable that some pitchers because of the movement, rotation and speed of their balls - end up overpowering or befuddling hitters at a greater rate, and thus popups are not necessarily predicitive of hrs based on upward trajectory at all? That for some, it does, others have a skill set that says pop-ups is my way of demonstrating hitters can't get all the way around or make good contact clean contact on a lot of my pitches?

please be gentle sen-sei...

yeah, i thought pop ups were a skill. like ground balls are a skill.

When I first saw the formula I thought I logged into a discussion about Altman's Z-Score. Can I use SIERA for HTH fantasy purposes?

The thing is that popup rate is highly correlated with flyball rate and home run rate, so it's indicative of that problem. It's a skill, but it's not something that good pitchers want to do. If you look at pitchers with a given K and BB rate, the higher their popup rate, the higher their ERA will be the next year as hitters will tend to hit the ball just a quarter inch lower down on the bat and hit it out of the park. We could have gotten cheap points in same-year ERA prediction but we would have less predictive ability for subsequent years. It's a matter of being honest about the process-- we're capturing skills rather than DIPS, I guess.

traderdave wrote:When I first saw the formula I thought I logged into a discussion about Altman's Z-Score. Can I use SIERA for HTH fantasy purposes?

I'm sure it would help fantasy, but I don't really do fantasy baseball so I don't know much about how. I'm sure you could improve upon a projection system by letting SIERA pick high and low on certain projections.

MattS wrote:The thing is that popup rate is highly correlated with flyball rate and home run rate, so it's indicative of that problem. It's a skill, but it's not something that good pitchers want to do. If you look at pitchers with a given K and BB rate, the higher their popup rate, the higher their ERA will be the next year as hitters will tend to hit the ball just a quarter inch lower down on the bat and hit it out of the park. We could have gotten cheap points in same-year ERA prediction but we would have less predictive ability for subsequent years. It's a matter of being honest about the process-- we're capturing skills rather than DIPS, I guess.

But popups againts guys like Clemens , Schilling and Pedro in their great years, or a reliever that throws in the high 90's with good movement ... where you could assert that pop-ups were a result of not being able to get on top of those pitchers power, be a positive not a negative, compared to other pitchers where the pop-up was more a matter of the batter just blowing it? And more predictive of an out of some kind, as they fall behind in the count, or get word down, or strikeout on a diff kind of pitch with a diff trajectory likely? Just seems like making it a universal negative indicator is questionable? (again, with all deference sen-sei)

Philly the Kid wrote:

MattS wrote:The thing is that popup rate is highly correlated with flyball rate and home run rate, so it's indicative of that problem. It's a skill, but it's not something that good pitchers want to do. If you look at pitchers with a given K and BB rate, the higher their popup rate, the higher their ERA will be the next year as hitters will tend to hit the ball just a quarter inch lower down on the bat and hit it out of the park. We could have gotten cheap points in same-year ERA prediction but we would have less predictive ability for subsequent years. It's a matter of being honest about the process-- we're capturing skills rather than DIPS, I guess.

But popups againts guys like Clemens , Schilling and Pedro in their great years, or a reliever that throws in the high 90's with good movement ... where you could assert that pop-ups were a result of not being able to get on top of those pitchers power, be a positive not a negative, compared to other pitchers where the pop-up was more a matter of the batter just blowing it? And more predictive of an out of some kind, as they fall behind in the count, or get word down, or strikeout on a diff kind of pitch with a diff trajectory likely? Just seems like making it a universal negative indicator is questionable? (again, with all deference sen-sei)

You're right but it's okay anyway. Those guys-- Clemens, Schilling, Pedro-- have infield fly rates that are roughly average actually. Power pitchers do tend to have slightly lower BABIPs but because the coefficient on strikeouts in SIERA says "how much lower are pitchers' ERAs when their strikeout rate is higher by X, holding walks and GB/FB constant," it actually credits the pitcher with some of this skill. The reason that I think superstar pitchers don't actually have great popup rates necessarily is that they also are getting hitters to just barely nip the top of the ball, leading to dribbling ground balls, as often as they are getting hitters to just barely nip the bottom of the ball, leading pop ups.

yo, tell me about JC Romero.

(he seems like a guy who would get killed in other metrics)

Now I can't read the title of this thread without picturing Matt in a cheap tux with a microphone at a seedy tit-bar...

jamiethekiller wrote:yo, tell me about JC Romero.

(he seems like a guy who would get killed in other metrics)

romero (as a phillies only):

07- 4.66
08- 4.07
09- 5.31

Romero enters the game with one or two outs already recorded a lot so his walks and ground ball singles don't lead to as many runs as other pitchers with comparable skills. This probably represents what his ERAs might have been with neutral luck, defense, and park and ALSO had he started every inning with none on and none out.

MattS wrote:

jamiethekiller wrote:yo, tell me about JC Romero.

(he seems like a guy who would get killed in other metrics)

romero (as a phillies only):

07- 4.66
08- 4.07
09- 5.31

Romero enters the game with one or two outs already recorded a lot so his walks and ground ball singles don't lead to as many runs as other pitchers with comparable skills. This probably represents what his ERAs might have been with neutral luck, defense, and park and ALSO had he started every inning with none on and none out.

I have thought that Romero is kinda unique in that he gives up a slew of walks, more than pretty much any pitcher who has a job in MLB. But he has given up very few hit; very few HRs and gets a pretty good amount of ground balls and Ks. He's pretty deadly vs. lefties and I'd guess he faces a lot of those guys.

So doesn't the fact that he surrenders a tremendous amount of walks make him pretty unique and also mess up any metric designed to measure guys who aren't so unique?

What I mean is, a guy who is as wild as Romero is almost never effective. But Romero has been effective most of his career. And that makes him pretty unique I'd guess. And unique guys are usually exceptions most rules and formulas and stuff. (Like the way Moyer "broke" PECOTA).

Does this make any sense at all to you?

I would think Romero is a good example of the accelerating importance of GBs, with all the GBs he has induced the past few years to get out of (usually self-induced) jams. That, and the importance of very good infield defense.

I also think it's pretty clear he's been highly fortunate throughout his Phils career, his RISP splits are absurd each of the last few years, and a lot of that is microscopic BABIPs. He seems to have a tendency as a Phil to hit the spot/throw the strike when he absolutely has to, but his career RISP splits also suggest he doesn't have special powers to bear down in higher leverage spots. To those points

Career splits with RISP: .254/.381/.373, .297 BABIP
07 splits RISP: .150/.342/.183, .180 BABIP
08 splits RISP: .173/.337/.235, .246 BABIP
09 splits RISP: .125/.276/.125, .150 BABIP

Romero has gotten a lot of big outs for this team but I will continue to be scared when he takes the hill in important spots

Maybe he's just been really lucky.

I'm looking through his game logs. There are exactly three times in the last 3 years he's given up 3 BBs in one appearance/inning. All three times he has 0 ER and 0 strikeouts.

smitty wrote:

MattS wrote:

jamiethekiller wrote:yo, tell me about JC Romero.

(he seems like a guy who would get killed in other metrics)

romero (as a phillies only):

07- 4.66
08- 4.07
09- 5.31

Romero enters the game with one or two outs already recorded a lot so his walks and ground ball singles don't lead to as many runs as other pitchers with comparable skills. This probably represents what his ERAs might have been with neutral luck, defense, and park and ALSO had he started every inning with none on and none out.

I have thought that Romero is kinda unique in that he gives up a slew of walks, more than pretty much any pitcher who has a job in MLB. But he has given up very few hit; very few HRs and gets a pretty good amount of ground balls and Ks. He's pretty deadly vs. lefties and I'd guess he faces a lot of those guys.

So doesn't the fact that he surrenders a tremendous amount of walks make him pretty unique and also mess up any metric designed to measure guys who aren't so unique?

What I mean is, a guy who is as wild as Romero is almost never effective. But Romero has been effective most of his career. And that makes him pretty unique I'd guess. And unique guys are usually exceptions most rules and formulas and stuff. (Like the way Moyer "broke" PECOTA).

Does this make any sense at all to you?

interesting thought, and got me started on something. we've had this discussion before and finally with SIERAs available, i decided to test it. i looked for all pitchers between 2003-08 who had 40+ IP in one year and the year after (hence no 2009) who had BB/PA > .10, K/PA>.20, and GB/Batted Ball >.50.

I got 26 pitchers. They are as follows (name/year/era/siera/era-next-year)


Jose Contreras 2003/3.30/3.54/5.50
Francisco Cordero 2003/2.94/3.42/2.13
Kyle Farnsworth 2003/3.30/3.08/4.72
Danny Kolb 2003/1.96/3.36/2.98
Mike MacDougal 2003/4.08/3.90/5.56
Tom Martin 2003/3.35/3.70/3.97
B.J. Ryan 2003/3.40/3.24/2.28
Jay Witasick 2003/4.73/4.11/3.21
Matt Miller 2004/3.09/3.44/1.82
Guillermo Mota 2004/2.14/4.07/4.70
J.C. Romero 2004/3.51/3.87/3.47
Daniel Cabrera 2005/4.52/4.01/4.74
Jorge De La Rosa 2005/4.68/4.92/6.49
Ryan Dempster 2005/3.13/3.67/4.80
Jason Isringhausen 2005/2.14/3.96/3.55
Will Ohman 2005/2.91/3.86/4.13
Akinori Otsuka 2005/3.59/4.02/2.11
Bobby Jenks 2005/4.00/3.01/2.77
Fernando Rodney 2006/3.52/3.87/4.26
Scott Downs 2007/2.17/3.19/1.78
Pedro Feliciano 2008/4.05/3.94/3.03
J.P. Howell 2008/2.22/3.41/2.84
Clayton Kershaw 2008/4.26/3.97/2.79
Brian Wilson 2008/4.62/3.51/2.74

average: 3.40 ERA, 3.70 SIERA, 3.61 next-year ERA

conclusions

1) This type of pitcher is often a reliever
2) These pitchers were often lefties (9 of 26)
3) They often beat their SIERA slightly
4) That often regresses the following year to a higher ERA close to SIERA
5) That is still slightly below their SIERA but really not much
6) J.C. Romero is still pretty darn lucky

I think most of this effect is entering innings late and being paired with lefties who can't hit him that well in the first place.

Yeah. I think he's been used very intelligently by the Phils which helps. Can you figure out his SIERAs for the years 2002-2009?

For example since 2002 his ERA compared to xFIP has been:

1.89/3.56

5.00/4.88

3.51/4.01

3.47/4.90

6.70/4.91

1.92/5.05

2.75/4.37

2.70/5.50

He's had a couple of bad years. But he's had 6 years with very good ERAs. most of the time his xFIP is really high. This suggests it's kinda normal for him to far outperform his projected numbers.

It's not really like he's letting a lot of other guy's runs score either as his LOB% has been very high most years. I think there's a thing called a strand rate but I'm not sure where to find that.

Anyway, if Romero has been lucky he's been lucky a LOT and I'll take it.

I'm not sure if he'll keep it up though. He had that 50 game suspension for whatever reason and he finally has to deal with arm problems. He might not come back from them. But I do contend he's been an effective pitcher most of his career since converting to the pen.

joe table wrote:
Romero has gotten a lot of big outs for this team but I will continue to be scared when he takes the hill in important spots

I like JC and I like to try to defend his record but this here thing you say is really, really true.

It's like:

"THROW STRIKES YOU BIG DUMMY"

"HE STINKS -- HE's GONNA BLOW IT."

"ARGHHHHHHHHHHH!!!!!!!!"


"WAY TO GO JC-- I TOLD YOU HE'd DO IT!!!" "HE's GREAT!!!!!"

MattS wrote:The thing is that popup rate is highly correlated with flyball rate and home run rate, so it's indicative of that problem. It's a skill, but it's not something that good pitchers want to do. If you look at pitchers with a given K and BB rate, the higher their popup rate, the higher their ERA will be the next year as hitters will tend to hit the ball just a quarter inch lower down on the bat and hit it out of the park.

an aphorism leaps from this observation:

"Grounders produce fewer HRs than do fly balls, regardless of skillz"