Did we underperfom or overperform expectations? - Pitching
by GiantFan on Oct 14, 2009 3:12 AM PDT
36 comments
In the previous post i set out to investigate whether we under-performed or over-performed our projections both from an individual and team perspective. In this post I'm going to undertake a similar analysis but instead looking at the pitching staff. It's a bit trickier with pitching as there are other factors to disentangle, but to do the analysis, I've looked at FIP.
The projections I've used are an average of the Marcel, CHONE and ZIPS projections again removing Marcel for players with limited major league experience. In order to keep consistent with the previous post, positive run differences correspond to an over-achievement. The pitchers are listed by number of innings.
Outperformed Their Projection
Underperformed Their Projection
- Matt Cain FIP = 3.89, projected FIP = 3.74, Run difference = -3.6
- Jonathon Sanchez: FIP = 4.17, projected FIP = 3.99, Run difference = -3.3
- Randy Johnson FIP = 4.91, projected FIP = 3.97, Run difference = -9.8
- Bob Howry FIP = 3.85, projected FIP = 3.84, Run difference = -0.1
- Justin Miller FIP = 4.91, projected FIP = 3.65, Run difference = -7.9
- Brad Penny FIP = 4.35, projected FIP = 4.31, Run difference = -0.2
- Joey Martinez FIP = 4.86, projected FIP = 4.64, Run difference = -0.7
- Ryan Sadowski FIP = 4.72, projected FIP = 4.44, Run difference = -0.9
A quick look shows that it's more balanced that the offense, with around half under-performing their projections and around half over-performing them. If you total up the difference then you can a over-performance of the projections by around 22 runs. Like in the offensive case, an adjustment is needed when we're talking about the team as a whole. A quick back of the envelope calculation suggests doing so means we outperformed our component statistics by around 9 runs (taking into account defensive numbers from UZR).
That works out to an over-performance of around 30 runs in total. if we were to combine that with the offensive numbers then we performed very similarly (in terms of RA/RS) to what the projections suggested in pre-season. Converting that into an expected win record we ended up winning 2-3 games more than projected.
This FanPost is reader-generated, and it does not necessarily reflect the views of McCovey Chronicles. If the author uses filler to achieve the minimum word requirement, a moderator may edit the FanPost for his or her own amusement.
1 recs |
36 comments
Email
Print
Comments
Quibble
Just as with the offense, is it really wise to divide this along simple plus/minus lines? Any pitcher is probably just one or two odd bounces (which might not show up in the stats used in the calculations) from arbitrarily gaining or losing one or more runs especially if they saw heavy playing time. It would be useful to know what the league-wide standard deviation is.
Oh to hell with it… Timmeh IS Gohd!
Fred Lewis can stand under my umbrella.
31 May 2007, 21:38 EST - the last time Matteh's career W-L wasn't below .500
We are at war with Los Angeles. We have always been at war with Los Angeles.
Lowering the Quality of Internet Discourse Since 1985™
by S.F. Giangst on Oct 14, 2009 6:47 AM PDT reply actions 0 recs
You’re right, small differences don’t imply a huge amount and i’d imagine the league wide standard deviation would be quite sizeable on a single player sample size. It’s more relevent when looked at a team wide scale or looking at trends. It’s also just interesting to see how players did compared to their expectations.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 14, 2009 7:29 AM PDT up reply actions 0 recs
Any chance you can do this again using tRA instead of FIP…?
by Missing Barry on Oct 14, 2009 6:57 AM PDT reply actions 0 recs
Well i chose FIPover tRA as it’s conveniently displayed on Fangraphs. The problem with doing it again using tRA would be how to calculate tRA projections. I would imagine that the projecton systems don’t publish tRA projections and so you’d have to calculate them yourself. However, to calculate them you’d need projections for LD rates, etc and again i’m not sure that the projection systems publish those.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 14, 2009 7:34 AM PDT up reply actions 0 recs
Ah, good point. I dunno why but I have a harder time with FIP being context neutral and it’s assumptions that other variables are out of a pitchers control than I do with using a context neutral hitting stat like wOBA…
by Missing Barry on Oct 14, 2009 9:12 AM PDT up reply actions 0 recs
Yeah, that’s understandable, i think wOBA is better reflection of past performance than FIP and so i find it easier to trust offensive calculations of that type.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 14, 2009 10:04 AM PDT up reply actions 0 recs
How did Merkin Valdez overperform and Justin Miller underperform? Also, did Johnson really play long enough to be worth that many negative runs?
The very bad man traded my son...So now I'd like you all to meet my new son, Ryan "Aaron" Garko...Dammit it's just not the same!
by boonitez on Oct 14, 2009 9:47 AM PDT reply actions 0 recs
RJ pitched 96 innings on the year. While it was much less than what we would have liked, it’s nothing to sneeze at. Miller had a shiny ERA but pretty god awful peripherials.
by xanthan on Oct 14, 2009 10:10 AM PDT up reply actions 0 recs
Also, RJ’s FIP sucks because he gave up an unlucky amount of home runs. His HR/FB% of 19.2% is a good 7-8 points higher than league average. His xFIP (normalized HR rate) of 3.79 is much nicer.
by xanthan on Oct 14, 2009 10:15 AM PDT up reply actions 0 recs
As mentioned, two reasons really. One is that Miller outperformed his peripherals whilst Valdez underperformed his so Valdez actually had the lower FIP. The second is that Miller was projected to have a much lower FIP than Valdez in part due to his previous two seasons of low FIP pitching.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 14, 2009 2:24 PM PDT up reply actions 0 recs
I don’t understand the point of analyzing the difference in runs if you’re going to do it according to FIP and not actual runs allowed.
GROUGTHINK ALERT
The first Chester Arthur fanboy ever.
by groug on Oct 14, 2009 12:17 PM PDT reply actions 0 recs
Why, FIP is a (albeit flawed) measure of how a pitcher did in terms of runs. The problem is if you do a similar analysis using ERA then you’ll find that most of our pitchers outperformed their expectations. The problem is that won’t be because the pitchers actually outperformed but rather we had great defense. FIP is used to try and remove that defensive component.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 14, 2009 2:27 PM PDT up reply actions 0 recs
Kind of OT, but how did we end up with great defense? I mean, did it seem like a great defense at the time? Not to me. The left side was pretty bad. Rowand was mediocre at best. Molina was mediocre at best. Second base was a mixed bad. The corner outfielders were great, and Ishikawa was great when he was on the field. Far be it from me to argue with defensive efficiency ratings and all, but does that really add up to the best defense in baseball?
by Evan on Oct 14, 2009 4:18 PM PDT up reply actions 0 recs
It wasn’t the best defense in baseball, but it was Top 3-5. The infield was not great by any means – but we have a fly ball pitching staff. The outfield D was great, and compared to other teams our outfielders got a lot more chances relative to our infielders. The whole AT&T suppressing HR’s helps our outfield D get more chances and perform better, too…
by Missing Barry on Oct 14, 2009 4:45 PM PDT up reply actions 0 recs
Prospectus says it was the best, applying park factors or some other adjustments.
But that’s a good point about the nature of the pitching staff. The shaky infield just didn’t get the chance to do a lot of damage.
by Evan on Oct 14, 2009 5:03 PM PDT up reply actions 0 recs
Our infield defense wasn’t shaky though, our 1B defense was fantastic. Even if you ignore that (which I’m guessing was what you meant) our 2B, SS and 3B defense were all around average. And that’s not really surprising either, we only really have two below average defensive players, Sandoval (who isn’t bad enough to have a big effect) and Velez (who didn’t play enough infield to have an effect).
I think part of it is we’ve been spoiled by good defense (partly because Sabean isn’t a fan of big offense/ bad defense players) and so our perception of defense is slightly skewed (which leads to incorrect conclusions such Lewis is a terrible defensive LF). That similar effect leads to pitchers being slightly under-rated and hitters being slightly over-rated. It’s obviously a natural effect in sports, especially in bi-polar teams. I’m obviously talking in general terms rather than you, me or anyone as an individual.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 15, 2009 2:27 AM PDT up reply actions 0 recs
I’m saying I don’t think the defense is really that good. Our starters collectively strike me as an unremarkable unit, probably a tick above average. My read is that the team’s exceptional performance on defense this year was caused by (a) lots of innings for a defense-oriented bench, (b) a great fit between pitchers, park, and defensive strengths, as mentioned above, and © luck.
The second point is the interesting one. I suspect Sabean is aware of this symbiosis, in his own unscientific way. So I’m betting we’ll see Cameron or Figgins or Anderson in the outfield next year instead of Dye or Abreu or Bay.
by Evan on Oct 15, 2009 9:44 AM PDT up reply actions 0 recs
Fair enough, it’s certainly possible we aren’t that good. Although looking at the numbers for the starters they seem pretty in-line with what i’d expect.
Winn has been great for a number of years, Ishikawa has always had a great defensive reputation, Renteria has always been around league average, Rowand has always been league average or above, Sandoval looks below average but not horrendous, our 2B options have all been or looked around average. The only one that looks questionable is the other OF spot as i’m not sure whether Schierholtz, Velez, etc are that good defensively.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 16, 2009 2:52 AM PDT up reply actions 0 recs
I exected the defense to be pretty bad.
More precisely, I expected and atrocious infield defense to be bad enough to sink the above average outfield defense. As it turned out, the infield defense was about average, and the outfield defense was better than expected (thanks in large part to a parade of excellent-fielding reserves).
VAE PVTO DEVS FIO
by Bhaakon on Oct 14, 2009 4:56 PM PDT up reply actions 0 recs
rowand is actually a pretty good defender
by sfoakbay on Oct 14, 2009 6:18 PM PDT up reply actions 0 recs
Well that depends on what you mean by pretty good. He’s probably pretty average for a CF, but then again CF’ers represent one of the better defensive groups in baseball, so compared to all other players I think it’s fair to call him pretty good, but comparing him just to his position that wouldn’t be true…
by Missing Barry on Oct 14, 2009 6:50 PM PDT up reply actions 0 recs
I pointed this out in the OF thread, but if you’re an average CF, that leaves you as probably one of the 20 best outfielders in baseball.
Please hit better, Randy Winn.
by oldjacket on Oct 15, 2009 7:49 AM PDT up reply actions 0 recs
I wouldn’t mind toying with the idea of moving Rowand to a corner next season. He would probably rate pretty well on defense.
by xanthan on Oct 15, 2009 7:58 AM PDT up reply actions 0 recs
So if LF = Jermaine Dye, would you rather have Rowand or Lewis in RF?
by Missing Barry on Oct 15, 2009 9:34 AM PDT up reply actions 0 recs
I don’t know, it’s most likely a toss-up but I would probably go with Rowand.
by xanthan on Oct 15, 2009 9:45 AM PDT up reply actions 0 recs
I don’t even think I’d hesitate to go with Lewis for a second. I might consider a platoon, but that’s the closest Rowand would get to regular PT for me…
by Missing Barry on Oct 15, 2009 10:53 AM PDT up reply actions 0 recs
Cameron would be an interesting signing if he came on a short term contract. He’s probably as good offensively as some of the other options (Dye, Nady) and would be a big defensive upgrade (over them, not who he’d be replacing)
Proud parent of Waldis Joaquin!
by GiantFan on Oct 15, 2009 8:14 AM PDT up reply actions 0 recs
I’d even say it’s fair to call Cameron a sizable defensive upgrade over Rowand. Cameron is above average for a CF – there are only a few defensive OF’ers better than him in all of baseball.
by Missing Barry on Oct 15, 2009 9:35 AM PDT up reply actions 0 recs
I agree with that, but I doubt they’d shift Rowand anyway.
by Evan on Oct 15, 2009 9:45 AM PDT up reply actions 0 recs
You’re right, he probably is a defensive upgrade over Rowand, i was more comparing him to the combination with Winn (as that’s who he’d be replacing) I guess it’d still be an upgrade as Rowand would rate much better at the corners and obviously it’d be a big upgrade offensively.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 16, 2009 1:35 AM PDT up reply actions 0 recs
Well, we had three positions where we had the best defensive in baseball (1B, LF) or very close to it (RF) The other four positions (excluding catcher defense) we had around league average defense. So it’s not surprising that we had one of the best defenses.
For comparison if a team had three players that were around the best in baseball for their positions offensively and the rest were around league average offensively, you’d expect them to be one of the best offensive teams wouldn’t you?
Proud parent of Waldis Joaquin!
by GiantFan on Oct 15, 2009 2:10 AM PDT up reply actions 0 recs
But isn’t the runs saved category here just a measure of the runs that would have been saved with an average defense? And isn’t this disregarding luck? I mean, even if he had the same great defense, I have a really hard time seeing Matt Cain outperforming his FIP by almost a full point again next year. So while it isn’t a thing that is likely to happen again, there’s just no way you’ll convince me that Cain didn’t overperform expectations this year with regards to the runs he actually gave up.
GROUGTHINK ALERT
The first Chester Arthur fanboy ever.
by groug on Oct 15, 2009 10:17 AM PDT up reply actions 0 recs
That’s a fair point. I don’t know of any real solution to that problem, though. ERA is inadequate because of the role defense plays in it, and basically every other stat is context-neutral. I have to say I think about this sometimes, too, I do believe Matt Cain gave more value than his FIP indicates to the Giants this season, just as I think guys like Penny/Smoltz gave less value to the Red Sox than their FIP indicates, but I’m not sure how to quantify that.
The best I can do is say over a players career, the context aspect of it will even out. I also think there are exceptions where FIP or even tRA may not properly value certain pitchers – the problem is to figure out who those pitchers are you need a big enough sample to where that point isn’t really relevant anymore….
As for your first point about defense, it’s ok that it’s runs saved with an average defense. This point isn’t a problem – for example if a team allows 5 runs but the pitcher has a 4 FIP (assuming neutral luck), that extra run just gets credited to the defense, which makes sense from a theoretical standpoint because the defense caused it to happen by not making the outs an average defense would have made.
by Missing Barry on Oct 15, 2009 11:00 AM PDT up reply actions 0 recs
The runs saved category is indeed supposed to measure how many runs would have been saved with an average defense and with average “luck” (not necessarily the best term but I’ll use it here) In other words it is supposed be a measure of his “true perfomance level”.
In the case of Cain, his true performance level was around what was expected (slightly less). In terms of the runs he actually gave up his performance was much better than expected (around 20 runs better) The simplistic way of viewing that is that it was because he had great defense behind him and was “lucky”.
The problem is, as Missing Barry explained, is that it’s an overly simplified view and that’s why statistics like tRA are created, as we haven’t really found the ideal pitching statistic. Over larger samples like a career, or a team as a whole it’s pretty effective though. The reason for that is that any “luck” will normalize and defensive contributions are much more accurate over the larger samples.
For reference, as a team we conceded around 60 runs less than what FIP tells us we should be. That works out to being a much smaller difference than Cain had. UZR claims that around 50 of those runs were due to an above average defense. Which means that around 10 of those were attributed to “luck”.
Proud parent of Waldis Joaquin!
by GiantFan on Oct 16, 2009 2:44 AM PDT up reply actions 0 recs
Comments For This Post Are Closed
Secondary Sidebar
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2009 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
