Catching up to the AL Central: Starting Pitching
by kcdc1 on Aug 6, 2011 10:25 AM EDT
Email
PrintI've been thinking about putting together a series about how the Royals can catch up to the other teams in the AL Central and make the playoffs in 2012, so the first thing I wanted to do was see how close the Royals are. In doing so, I discovered something that really surprised me:
The best run differential in the AL Central is Detroit's -4 mark. The Royals are third in the division at -49, a mere 45 runs behind the Tigers.
Naturally, I wanted to get a sense for how the Royals could close that gap, and the first place I looked was the starting rotation. I figured I'd start by looking at the luck-neutral metrics for our current starters, and then I'd take out the crappiest starter and plug in someone good like CJ Wilson or Edwin Jackson and say, "Look! This is how close we could be if we just made a big FA signing!"
Well, it turns out I never made it that far because I made a second discovery that really surprised me:
80% of the gap in run differential between the Royals and the Tigers disappears when you neutralize luck in the Royals starting rotation.
For my luck-neutral pitching metric, I chose SIERA. For those who aren't familiar with SIERA, it gives you a sense of what a pitcher's ERA should be if they had neutral luck with BABIP and HR/FB. In this sense, it's very similar to xFIP, but it gives pitchers who miss a lot of bats or induce a lot of groundballs or flyballs credit for being hard to square up, and slightly lowers their expected BABIP. It's actually slightly harder on most of the Royals starters than is xFIP.
Then I calculated how many more or fewer earned runs each pitcher would have given up were their ERA equal to their SIERA. Here's what I found:
| IP | ERA | SIERA | delta(runs) | |
|---|---|---|---|---|
| Francis | 135.7 | 4.38 | 4.17 | -3.18 |
| Paulino | 62.3 | 3.75 | 3.17 | -4.02 |
| Hochevar | 145.3 | 4.95 | 4.3 | -10.50 |
| Davies | 61.3 | 6.75 | 4.21 | -17.31 |
| Chen | 82.3 | 4.37 | 4.69 | +2.93 |
| Duffy | 67.7 | 5.05 | 4.13 | -6.92 |
| Total | 554.7 | 4.88 | 4.11 | -39.0 |
In other words, if these Royals starters had had neutral luck this year, they would have given up about 39 fewer runs combined, pulling them almost even with the division-leading Tigers in run differential.
I ignored the starts from O'Sullivan, Mazzaro and Adcock, figuring that every team gets crappy spot starts from AAAA pitchers during the season. I decided to take those starts as built-in 'bad luck' that will be back next year, so I didn't neutralize it. But if you wanted to say that O'Sullivan, Mazzaro and Adcock's numbers should be luck-neutralized, you'd get another 6 or so runs back, which would make the Royals dead-even with the Tigers.
This method has some obvious short-comings. I didn't neutralize luck for any of the other teams in the division, I didn't look at luck in the Royals bullpen, I didn't regress anything to the mean, and half of the starters I listed won't even be back next year.
Still, this one simple step of neutralizing luck in the starting rotation puts the Royals near the top of the AL Central in run differential without any significant moves or FA signings. I think that's a good starting point when considering the Royals chances to contend in 2012.
49 comments
|
2 recs |
Do you like this story?
Comments
Don't know
This computer sucks. It took way to long just to get the few numbers I used. But they’re just as likely to have had good luck as bad luck. According to SIERA and xFIP, the Royals starters just happened to have an unusual amount of bad luck this year.
tableizer
http://tableizer.journalistopia.com/
If you’ve got them in a spreadsheet, this will convert it to HTML.
Glad I came, just wish I hadn't stayed so long.
Rock Chalk Talk
Jeff, KSinDC, Matt, Scott and/or Will
Do any of you have interest in doing (or helping me do) a “Reasonable Expectations for the 2012 Royals” column or series of columns that takes a look at what the Royals could expect in terms of run differential (and therefore record) next year if they made no roster changes? I’d want to neutralize luck for pitchers, regress each player’s 2011 numbers to the mean, and apply a simple adjustment for aging/development.
I suspect it would produce interesting results, but I could use some help and some neutrality or pessimism to balance my optimism.
I’d love to, but I really don’t have the time to devote to it. And Jeff, Matt, Gopherballs and probably others would be of more help with the methodologies and math in this kind of projection.
You may know me as NYRoyal.
by Scott McKinney on Aug 4, 2011 2:58 PM EDT up reply actions
I have interest, but I leave for Alaska for two weeks tomorrow
So I’ll be on internet blackout until Aug 18. If you’re still working on it then, shoot me an e-mail.
I was always asking the same thing, so start here:
http://www.fangraphs.com/blogs/index.php/predicting-a-teams-wins-using-underlying-player-talent/
I was about ready to start working with it my self. Let me know if you have any questions wydiyd ~ hotmail
- .-. ..- … – / – …. . / .—. .-. - .. . … …
by Jeff Zimmerman on Aug 4, 2011 4:30 PM EDT up reply actions
I've got a lot of spreadsheets going
For now, I’m just planning on using wRAA for offense and SIERA minus league(SIERA) for pitching. I’ll probably ignore defense and baserunning for simplicity.
How should I handle regression and aging/development?
Would it be difficult for you to use an average of SIERA, tERA and xFIP? Each metric has benefits, but also weaknesses. I think you even out the weird outliers that each system has (every such metric has some) if you average them.
You may know me as NYRoyal.
by Scott McKinney on Aug 4, 2011 4:48 PM EDT up reply actions
At this point, I think that'd be a little more work than it's worth to me
Those metrics are almost always very close to each other, so I don’t think it’d make a big difference. And it’s really for illustrative purposes anyway. I want to get a sense of whether the Royals true talent is closer to 70 wins, 80 wins or 90 wins, and I don’t think decimal points here and there will make much of a difference. Broad strokes.
I try do this sort of thing myself every season
what do you want help with? I did the “full version” two years ago, and although i learned a lot, ti took FOREVER.
I started doing partial version for 2011, then said “screw it.”
I did it in 2010 because I did an AL Central preview for the 2010 Royals Authority Annual, and I wanted to do it for myself at least once. It’s rather simple compared to what the real saber-nerds do, but I think it does okay.
I first got projections (or took projections and derived wOBA, etc.) from them for wOBA, then fielding, then pitching (I used FIP scaled to RA for simplicity). Then I did depth charts and playing time… for every team in baseball (curse you, interleague play! That took a long time, even on a rather informal/eyeball basis). I then generated a simple runs allowed and runs scored. RA is easy enough, of course. RS is a little more complicated, but it basically calculating a team wOBA then converting it to wRC (“absolute” wOBA runs created). The number of innings pitched is stable (there’s a crude way of allocating starter/reliever innings), and I use a very crude formula for increasing PAs for teams with higher wOBAs (crude in that it assumes a universally linear relationship between team wOBA and team OBP… but it’s probably better than nothing).
From that, generate a generic Pythag win% for each team (I did make a league adjustment). Then take a spreadsheet of the schedule (it’s not that hard to find if you look around, but it won’t be ready until after this season) and substitute each teams “neutral win%” (not really an accurate name, but you get the idea) and run a simplified log5/odds ratio for each one (expected neutral team win% – expected neutral opponent win% – .500) to get the win %, multiply that times 162, and you get the projected number of wins for each team.
One way of saving time would be to do what Rally did prior to 2010 and do a “all things work out” version for each team: just forget projecting playing time for the bench, and assume all the starters/top 5 pitchers stay healthy and stuff. A much shorter and easier version would just to do the short version for each team in division and compare “neutral” win%.
Not saying it’s brilliant, but that’s I’ve done it. That’s the short version, if you want more maybe we can do it by email. Heeck, you might be able to sucker me in to doing it again.
Making watching baseball as fun as doing your taxes.
My Twitter feed.
by Matt Klaassen on Aug 4, 2011 6:08 PM EDT up reply actions
Haha
Your version would be infinitely more in-depth than the version I would do. My plan is just to take a very simple projected wRAA for the 9 starting position players, and a simple projected ERA minus leagueERA for each current member of the pitching staff.
I’m just looking for a rough sense of where the Royals stand right now to see if they’re a few FA signings away from contention. I’m looking for whether they’re on the order of 70, 75, 80, 85 or 90 wins in true talent, so I wasn’t going to bother bench playing time, projecting other teams or the schedules. Hell, I’m even planning on ignoring defense. Broad strokes.
My plan for the hitters is to take their RoS ZiPS projection, add about 10 points of wOBA for the guys that are 24 and under, and subtract about 10 points of wOBA for the guys that are 28 and older. Will that at least get me in the right ballpark?
I don’t have a good idea for how to project development from the pitchers because pitchers don’t seem to have as predictable of an aging curve. Still, I’d like to give Duffy a little boost because I think his numbers have suffered slightly from rookie jitters.
Hmmm...
I think you at least have to do the other teams in the division to compare, since a “raw” set of ERA (or whatever)/wOBA numbers in a vacuum doesn’t tell you all that much. I guess it might tell you how the team is compared to the current numbers, although given regression to the mean, you’d get probably about what you’d expect: better pitching (because it’s been so bad), and probably the offense about the same or worse (Gordon, Melky, Frenchy all regressing down, Butler about the same, Hosmer and Moose up).
I’m not quite sure what you need help with… sorry. I’m on vacation and I think the Claritin needed to ward off the effects of my mom’s dog (of course WE were never allowed to have one when we were kids) has taken the edge off an already not-too-sharp mind.
Making watching baseball as fun as doing your taxes.
My Twitter feed.
by Matt Klaassen on Aug 4, 2011 6:49 PM EDT up reply actions
Just looking for runs vs league average as a guess at record
It’d be better to compare against the competition, but just runs scored above average and runs allowed above average should give a sense of whether they’re in the picture.
And the big thing I’m wondering is whether the +10 points of wOBA for position players under 25 and -10 for position players above 28 is a reasonable thing to do for aging curves.
And how would you handle adjusting for age/development on the pitching side?
Off the top of my head
+/- 10 flat wOBA seems a little extreme. A better idea might be something like +2 pts for each year younger than 27, -1 for each year older. That’s a ballpark, off-the-top-of-my-head.
It’s more important to regress. It’s a pain, but without it you’ll get kind of a wacky forecast, though. What were you going to add the points to? This season’s stats? Last three years’ average?
Honestly, for a generic pitching thing where you aren’t looking at components (and leaving aside which stat to use and, of course, the regression issue) and are just using one stat, I’m tempted to say don’t make any adjustments for age at all.
But if you really want to get crazy, maybe next simplest thing would be to not make any adjustments for young guys, and add.04 runs of ERA for each year over 27. If you want to be optimistic, maybe subtract .02 runs of ERA for each year under 27. But that’s kinda wacky…
Making watching baseball as fun as doing your taxes.
My Twitter feed.
by Matt Klaassen on Aug 4, 2011 7:59 PM EDT up reply actions
oh, yeah, sorry
one thing is that this late in the season ZiPS ROS is also projecting ROS playing time, which makes using them for measures of true-talent somewhat problematic. See this.
Making watching baseball as fun as doing your taxes.
My Twitter feed.
by Matt Klaassen on Aug 5, 2011 3:03 PM EDT up reply actions
Eh, with 50 games out, that shouldn't be a big problem
I’m okay with some rounding error
I fear with all of the “roughness” of your projections, regressions, etc. that perhaps the final numbers you come up with won’t tell us a lot. For instance, if your rough numbers come up with a true talent level of 80, do you think that will mean that this team has 80-win TTL for next season? Given all of the rough estimates you’re using, do you think the number you end up with will reliable as even telling us the general vicinity of the TTL of this team in 2012? I’m not saying don’t do it. I’m just questioning the conclusions we should draw for whatever numbers you come up with. If you come up with 80, then should we conclude that the Royals are pretty close to contending so they should “go for it”? Or would the number be for novelty purposes because you could be off by several wins?
You may know me as NYRoyal.
by Scott McKinney on Aug 5, 2011 11:45 PM EDT up reply actions
Wow
If you take the SIERA for the pitchers currently on the active roster, it comes out almost exactly at league average. And if you use linear weights (wRAA) to calculate batting runs above average from the current active roster, you get +31.5 runs above average. The team UZR is +5.4.
Taken together, the advanced metrics for the current roster suggest a team that’s about 40 runs above average, or about 60 runs above average over the course of a season. That comes out to about an 86 win team.
That’s without regression or aging/development.
Ran the offensive wRAA with the ZiPS RoS projections for each player
The offense dropped from +31.5 runs to +5.7 runs. That’s closer to what I expected.
Still need to regress the pitching, but it’s looking like I’ll come up with about a .500 team at present.
But for purposes of 2012, that counts regression of offensive players, but not aging/development, which should be solidly on the Royals side. For example, Hosmer has a .340 wOBA this year, and his ZiPS RoS projection is a .329 wOBA. Given normal development curves, I think Hosmer’s 2012 projected wOBA will be significantly higher than .329.
Whatever you do
don’t forget to factor in another year of comraderie under the belt.
We have met the enemy, and he is us.
Interesting
I’ve followed the Royals defensive rankings thru the year via DER, and we still rank way near the bottom like we always do. (Some of that is due to the bigger park, but only some. Any way you look at advanced metrics, the defense has still been well below par overall).
Now, watching every game, it appears to me, and most observers agree, that the team’s overall defense has improved. My first thought was perhaps the starting pitchers are so bad, they are allowing an inordinately high LD % or something. That’s simply not the case though – they are allowing only 17% line drives.
So, perhaps this validates your conclusion that the team’s starters have simply been unlucky this year, and are closer to average than their 13th ranking suggests?
If strikeouts are indeed fascist - then find me some starters that believe in fascism
It made me think about the defense too
The DER is pretty bad, but they make up for it somewhat with outfield assists. You’d think that, on balance, the Royals defense would be fairly neutral for its pitcher’s numbers. I don’t see a good explanation for their ERA being so much worse than their xFIP, SIERA, etc.
So, the new Meme:
1) acquire mediocre starting rotation
2) neutralize their luck
3) PARADE!
If strikeouts are indeed fascist - then find me some starters that believe in fascism
If we turned the banner lighting down when our opponents where hitting
and made it really bright when we were hitting…. would it help?
They've always said
You win 1/3 of your games.
You lose 1/3 of your games.
It’s what you do with the other 1/3 of games that matters.
But this shows that Davies isn't the worst pitcher ever?
Then clearly your research is wrong.
Kansas City Royals: your 2006 and 2007 NL Central champions!
He is the worst pitcher ever, I am probably the worst ever
He is the worst given given the number of innings he has pitched.
- .-. ..- … – / – …. . / .—. .-. - .. . … …
by Jeff Zimmerman on Aug 6, 2011 11:25 AM EDT up reply actions
His WAR is .6 this year
So he’s been twice as valuable as Joakim Soria, which proves has flawed WAR can be.
I do like the research but there is something inherently wrong with the premise, “every single Royal starter has had bad luck!” The truth is more often than not they have more control than it is assumed for their “bad luck,” and it has more to do with their skill, talent level, pitching performance than not.
I do have high hopes for a Hochevar, Duffy, Paulino, Crow starting rotation next year as long as a legitimate pitcher is added to that mix of 4
by GobbleforCyoung on Aug 6, 2011 11:33 AM EDT up reply actions
All that proves is that Relievers have less of an impact on games than people think they do, because they pitch less innings
Also, Soria had a really shitty start to the season if you can remember
SBN's most random and mysterious lurker guy who posts too much
Follow me on Twitter if you want: Lum_SM
nice job. seems to me such close looks will eventually
help explain causes in terms of why a team with .550 talent (Royals) has a .400 record. there’s another factor(besides luck) seems to me that needs to be accounted for—coaching. a manager who consistently removes starters after they fall on their azz instead of before e.g. or manager/pitching coach that perpetually fails to discern who’se “on” in the bull pen for a particular night—manager’s intuitive,perceptive abilities. Or, managers that manage pitching staffs on other basis besides performance/injury prevention. Interestingly we saw e.g.s to these in last night’s game where Yost did remove Paulino before major damage, and did pitch Crow in a crucial situation where Crow had nothing. What effect on 2012 rotation does the manager/pitching coache’s methodology affect outcome? as an aside, nice to see giovetella as an obvious highly conditioned athlete from the photos. compare his body composition to Soria/Crow’s—same set of photos. does this make a difference in optimal performance, particularly over time?
Good lord
Are you for real?
You may know me as NYRoyal.
by Scott McKinney on Aug 6, 2011 11:47 AM EDT up reply actions 2 recs
He's In The
Best shape of his life.
I used to be an A's fan until they left town and got good.
by philofthenorth on Aug 6, 2011 12:13 PM EDT up reply actions
Are they car engines or something?
SBN's most random and mysterious lurker guy who posts too much
Follow me on Twitter if you want: Lum_SM
Dont forget xGRIT stat for Grittiness
by KCTiger on Aug 6, 2011 12:10 PM EDT via mobile up reply actions
cc sabathia and prince fielder are fat....
mitch maier is a physical specimen
Fire Everyone
by billybeingbilly on Aug 6, 2011 12:39 PM EDT up reply actions
So it sounds like a manager's most important quality
is knowing what will happen before it happens. Or is it the ability to go online and criticize managers after the fact for not predicting the future?
It would be impressive
if a manager could single-handedly change a team with a 89-73 pythagorean to a 65-97 team.
A 5 man rotation is too small a sample size to assume their collective luck will evens themselves out
They could just be 5 unlucky losers.
I am probably the only Royals fan in Hong Kong?
by Yamfun Cheng Kamfun on Aug 6, 2011 1:11 PM EDT reply actions
Interesting.
It would be fun to see this done for each Central team’s starters AND relievers, AND something similar for batters (comparison to ZiPS? I dunno).
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Royals Review
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
In Association With
Big League Stew - MLB - Yahoo! Sports
- Eric Hosmer redefines the Baltimore chop with infield double (Video)
- College player makes impossible over the fence catch to save tournament game (Video)
- Security!: Camden Yards guard collides with Royals’ Mike Moustakas
- The Juice: Nelson Cruz matches career high eight RBIs, Pujols stays hot in Seattle
- Benches clear at Fenway as Red Sox-Rays rivalry heats up
