Best Pitching Units by tRA, 2008.5

by Matthew on Jul 14, 2008 6:02 PM PDT in The More You Know Comment 38 comments

We're roughly half way through 2008 and we just took a look at 2007, so let's see how teams are faring thus far this season shall we?

First off, some league wide numbers to provide that oak-casked, hoppy context. Word of warning though, the run and out constants have not been updated in over a month, so these are the slightest bit off, as you can tell by the cumulative figure for O-xO. The system is currently expected too few expected outs which means that once adjusted, the overall tRAs will go down a skosh.

Entry	O-xO	xO	xR	tRA
NL - SP	-55	23231	4155	4.83
NL - RP	-43	12869	2105	4.42
AL - SP	141	20868	3607	4.67
AL - RP	17	10328	1629	4.26
Table 1: 2008 tRA Figures

O-xO = Outs - eXpected Outs: A positive number means more outs were record than the underlying numbers would suggest (good luck). A negative number means the opposite.

Offense and scoring is down league wide so far this year, which isn't much of a surprise since it's usually down at the start of the season and peaks in the hot July and August months which we haven't gotten to yet. Nonetheless, the average MLB hitter had a .758 OPS in 2007 and so far through 2008 it stands at .740, a hefty drop. While both leagues are down offensively, it's the AL that has seen the bigger drop as you can see here in the respective tRAs.

This effect is only magnified in reality by the opposite polarity in the O-xO column for the AL against the NL. By the lower tRAs we see that AL pitching should be doing better at run prevention than their NL brethren and with the AL seeing many more outs than expected on top of that, we'd expect to see run prevention much improved in the AL. Indeed, AL teams are surrendering 4.44 runs per game while their NL counterparts are up at 4.63.

BEST ROTATIONS IN 2008

Arizona Diamondbacks, 3.73
Los Angeles Dodgers, 4.09
Toronto Blue Jays, 4.09
Chicago White Sox, 4.13
Oakland Athletics, 4.19

This seems like a grab bag of teams based on actual MLB standings, but take a look at their expected wins according to BaseRuns: 86, 81, 87, 97 and 92. Those are hefty numbers and should be good enough to land at least two in the postseason were they to play to those records.

WORST ROTATIONS IN 2008

Texas Rangers, 5.70
Pittsburgh Pirates, 5.66
Baltimore Orioles, 5.56
Florida Marlins, 5.23
Cincinnati Reds, 5.20

Pittsburgh and Cincinnati are the only two truly terrible teams on this list as Texas and Florida make up for their loathsome starters with tremendous offense and the Orioles with good defense. Worst to first is nearly two runs per game which is a fantastic amount.

BEST BULLPENS IN 2008

Chicago White Sox, 3.25
Atlanta Braves, 3.57
Philadelphia Phillies, 3.61
Los Angeles Dodgers, 3.63
Boston Red Sox, 3.80

The White Sox and Dodgers appear on both best of lists which is pretty telling why Chicago is where they are and just how bad the Dodger's offense has been to keep them from running away with the pathetic NL West.

WORST BULLPENS IN 2008

Pittsburgh Pirates, 5.36
Texas Rangers, 5.14
Milwaukee Brewers, 5.12
St. Louis Cardinals, 5.02
Cleveland Indians, 4.99

The Pirates and Rangers again appear 1-2 on the worst list though they swap spots this time around. That's remarkably horrific when you consider it. The Brewers are feeling the loss of Francisco Cordero for sure while the Indians at least finally made the move to cut Joe Borowski out of their life. Worst to first here is comfortably over two runs a game.

Seattle ranks near the bottom of the pack with a SP tRA of 5.13 and right about average in baseball (though below average in the AL) with a bullpen tRA of 4.35.

4 recs | Comment 38 comments

Story-email Email Printer Print

Comments

Well, now we know what Neil Huntington needs to fix first

by Gomez on Jul 14, 2008 6:22 PM PDT reply actions 0 recs

Rangers pitching.

Is there any particular reason why this is always a problem?

by ThundaPC on Jul 14, 2008 7:27 PM PDT reply actions 0 recs

Arlington

that park is murder. Also, the fact that they used to have a crap FO. I think at this point they should give up on the whole pitching thing and just try to do it with a steamroller offense.

by Bearskin Rugburn on Jul 14, 2008 7:44 PM PDT up reply actions 0 recs

You know what would be cool?

I’d love to see home/road splits for movement on the Texas pitchers breaking stuff. C’mon Matthew/Jeff/others who may have attended pitchfx conferences – we need park effects for curve balls, sliders and sinkers!
We’ve always heard about the effect that Coors has had on off-speed stuff, but I wonder if other parks have similar (though perhaps smaller) effects. May need to wait until the system’s been up and running for a while to make sure any differences aren’t associated with the equipment/positioning/whatever, but this would be really, really awesome.

by marc w on Jul 15, 2008 10:57 AM PDT up reply actions 0 recs

That's already been done, by Josh Kalk I believe, among others.

by Matthew on Jul 15, 2008 11:04 AM PDT up reply actions 0 recs

Got a link, you big tease?

by marc w on Jul 15, 2008 11:16 AM PDT up reply actions 0 recs

You're not talking about the park-specific corrections

Josh uses? There was a paper on that at the conference you went to, but that’s not really the same thing.

I’m not seeing a home/road split on Josh’s pitch fx cards, but it could be somewhere else. Was this a THT article?

by marc w on Jul 15, 2008 11:50 AM PDT up reply actions 0 recs

I'm confused then by what you're asking for.

by Matthew on Jul 15, 2008 11:58 AM PDT up reply actions 0 recs

This is how I picture you right now

The Jose Lopez Watch - 113 H - 12 BB - 67 G Left

by seattlebruin on Jul 15, 2008 12:13 PM PDT up reply actions 0 recs

What I mean is:

If I’m understanding this right, Josh Kalk’s work attempts to adjust for errors/discrepancies in the recordings at various parks by adding a correction to the initial parameters of a pitch – basically regressing many of these factors towards a league average.

That’s really great, as it helps smooth out any inconsistencies in the data. But it’s focused mainly on getting comparable data for things like release point. Great stuff, and sure, knowing the release point more accurately will help you calculate break more accurately, but it’s still not quite the same thing.

What I’M looking for is the park-by-park breaks of, say, Erik Bedard’s curve ball, but again corrected for park/camera errors. That’s the tough part – right now, the errors in the system are large (as Ike Hall showed between Petco and Dodger stadium), so I’m not sure that you can glean any meaningful info on changes in break. But as the system is refined (and I’ve heard it’s a lot better in 2008 as compared to 2007), these sorts of measurements should be easy.

Just imagine josh’s player cards, but with an additional filter for park, or even home/road.

by marc w on Jul 15, 2008 1:32 PM PDT up reply actions 0 recs

I believe that is part of Kalk's correction data

is not only fixing release point stuff but also adjusting for the less break you see in Coors, as one example.

by Matthew on Jul 15, 2008 1:36 PM PDT up reply actions 0 recs

Yeah, he corrects for acceleration

which necessarily would pick up the ‘coors effect’ in this case.

But again, my point is that Josh is trying to solve a different problem. He’s trying to get uniform data by essentially translating everything into a neutral park. He’s also not directly adjusting break – just initial location, velo and accel.

All of this is really awesome, but let’s say we’re looking at his adjustments – a list of park factors for, oh, initial velocity. How can these data help us figure out if Sean Green’s sinker sinks more in Park A vs. Park B?

by marc w on Jul 15, 2008 1:43 PM PDT up reply actions 0 recs

Baseball Reference currently has their multi-year park factor at 100, yet...

ESPN’s at 117, 98, 108, and 107 going from 2008 back to 2005.

Weird. What’s up with B-Ref?

by Sky Kalkman on Jul 16, 2008 8:23 AM PDT up reply actions 0 recs

ESPN park factors are flawed.

and BR’s are wrong here. 104 is the overall PF we have for TBiA.

by Matthew on Jul 16, 2008 3:25 PM PDT up reply actions 0 recs

care to elaborate on why ESPN's factors are flawed?

Don’t they simply do compare runs scored at home by both teams to runs scored on the road by both teams? I understand they don’t do anything across seasons, but it seems like what they do is the beginning to a more accurate analysis, no?

by Sky Kalkman on Jul 17, 2008 3:51 PM PDT up reply actions 0 recs

Aside from trading so much of it away

Here’s a rough guess: could it be that the extremely pitcher-friendly park in Oklahoma inflates their opinion of pitchers on the AAA team?

by Gomez on Jul 14, 2008 7:45 PM PDT up reply actions 0 recs

I don't know that Oklahoma is a big-time pitchers park

I know MiLBsplits has it as nearly impossible to hit HRs, but at First Inning, it’s actually slightly ABOVE average for HRs (and everything else).

I don’t know what’s right, but the First Inning numbers are a bit more recent, I think, though they may also be more volatile (the Nelson Cruz effect). We’ll see.

I’m also not sure that Oklahoma is inflating anyone’s reputation; the Rangers pitching ‘spects have been hyped out of the Cal league and have generally not fared well in AAA. Edinson Volquez was fine, but still had a FIP over 4 in AAA in 2006. John Danks was slightly worse, and Thomas Diamond looks like he may not even make it to AAA. Same story with Eric Hurley, who’s pitched OK in the AL, but terribly in the PCL (FIP near 5.5).
The Rangers problem still seems to be difficulty in recognizing quality pitching.

by marc w on Jul 15, 2008 10:44 AM PDT up reply actions 0 recs

Something just occured to me

Cellular Field is a great place to be a hitter, and this metric is supposed to neutralize that (actually, how regressed are the homeruns given up? cause I’m pretty sure it’s the most homer happy place in the Majors). However, walks and Ks are also affected by local environments, sometimes significantly. So where I’m going with this is that the White Sox pitching might actually be better than it looks even by tRA.

by Bearskin Rugburn on Jul 14, 2008 7:42 PM PDT reply actions 0 recs

All the park effects are factored in

by Graham on Jul 14, 2008 7:44 PM PDT up reply actions 0 recs

Here're the factors we use (modified from THT)

by Graham on Jul 14, 2008 7:47 PM PDT up reply actions 0 recs

Is it bad that I think its funny

RFK field is where you’re most likely to get plunked?

by Bearskin Rugburn on Jul 14, 2008 7:51 PM PDT up reply actions 0 recs

Heh, Tampa, CHC and Texas all have higher HBP rates than Washington

I was looking at the extended spreadsheet and Oakland’s foul-out factor was 1.701

The Jose Lopez Watch - 113 H - 12 BB - 67 G Left

by seattlebruin on Jul 15, 2008 9:03 AM PDT up reply actions 0 recs

Oh shit. You're right

But Texas is appropriate too, although they don’t play in Dallas.

by Bearskin Rugburn on Jul 15, 2008 9:13 AM PDT up reply actions 1 recs

See JI? This is how you do subtle...

Also, I think that the question on the THT page that ZuLong linked to is intriguing, and I might have an answer for it.

When I pitched in HS, there was always one field where I noticed the mound dirt was softer, and the slope seemed different, like I could push off better. I always got the ball lower in that park, and had almost no FBs. Nobody else understood what I was talking about with the mound, but FBs were always so scarce in that park that we started playing our OF in all the time to try to nab the slow runners at first and get LDs on the fly.

I’ve always thought that the ground crew was a part of it, and this might be the same case.

by Faux on Jul 15, 2008 9:39 AM PDT up reply actions 0 recs

Huh?

by JI on Jul 15, 2008 9:45 AM PDT up reply actions 0 recs

Knowing what you will say is too easy sometimes.

by NOLAmarinergirl on Jul 15, 2008 10:18 AM PDT up reply actions 0 recs

but the picture of Janet Weiss

by JI on Jul 15, 2008 10:26 AM PDT up reply actions 0 recs

The only thing I could connect to that was Rock Band.

Yeah, I failed at that.

by NOLAmarinergirl on Jul 15, 2008 10:34 AM PDT up reply actions 0 recs

"On the internet, no one can hear your sarcasm."

BR attempted to link the RFK assassination to getting plunked, then attempted to link the whole thing to Dallas, unaware that RFK was shot in LA, and JFK was shot in Dallas.

I was more making fun of that, not you so much.

by Faux on Jul 15, 2008 11:13 AM PDT up reply actions 0 recs

Sorry, dumb question time

are the park factors calculated based on visiting team data? Because using home team results (even if comparing home/away splits) could reflect a home field adjustment, ie the Angels and the home run factor.

by Bearskin Rugburn on Jul 14, 2008 8:14 PM PDT up reply actions 0 recs