Junior Rejuvenated?

After Ken Griffey Jr. was traded last week, I had a discussion with someone about the impact of the deal on the White Sox. My point (as stated in the post about the trade) was that Junior's bat is still pretty good, but it's no longer great and that his horrible defense (in RF no less) counteracts it, and then some. The counter-argument was that Griffey is going to a contending club, which should "energize" and "rejuvenate" him, and "give him back his fire". And so on, and so forth. There was more to it than that, but the main points were (1) Intangibles make players play better (and also the players around them) vs. (2) Intangibles can allow a player to play to their fullest potential, but Junior's potential isn't that high anymore and so even if he has been "rejuvenated", he still won't help Chicago that much (if at all).

This raises a simple question: does going from "playing out the string" to being in a pennant race in the middle of a season cause players to step up there game? I decided to find out. To do that, I looked at the trades from June, July, and August from 2001 to 2007 in which a player went from a club out of contention to one that was fighting for a play-off spot. Now, it was a little complicated to look at the records (and standings) of the teams at the time of the trade, so I estimated based on the final standings. Generally speaking, if a team was within a few games of a division title (or the wild card in many cases) I counted them as a contender. If a team ended the year 64-98, then I assumed a player wouldn't be giving it his all in the dog days of the summer. (Actually I don't really think that - I think a guy will rarely make it to the majors if he does that kind of thing to a significant degree - but for the purposes of this exercise, it's what I went with.) I only counted players with at least 75 at bats or 50 innings pitched both before and after the trade to try to minimize the small sample size issues a little. Since there are other factors affecting performance, like the change in leagues or ballparks, I used OPS+ and ERA+ as my comparison stats since they account for such things. Also, since both are on the same scale, I can combine the pitchers and the hitters together.

There were 54 cases to look at, with three players being traded multiple times (Kenny Lofton and Carl Everett twice each (with the White Sox acquiring Everett both times in consecutive years), and former O's favorite Jeff Conine three times). I looked at the actual change in OPS+ or ERA+ (going from 80 to 100 is a +20) as well as the percentage change (going from 80 to 100 is a +25%). So, the results?

The average OPS+/ERA+ change ended up being +0.8; a player, an average, would go from an OPS+ of, say, 113 to 113.8. In other words, it was pretty much exactly the same. The average percentage change was +4.8. How about when taking out the outliers (those cases more than two standard deviations from the mean like Julio Lugo going from a 124 to a 41 OPS+ (-83) or Woody Williams going from an 81 to a 189 ERA+ (+108))? Well then the average change goes all the way up +1.8, and the average percentage change actually goes down to +2. If there is a "contention factor", it is clearly not very significant. I would also like to mention that I originally stopped at trades back to 2002. At that point, the average change was negative at -3 (which means players actually got a little worse when the pressure was on) and the average percentage change was 0. I'm not going to discount 2001, but I think that goes to show that using a one point increase in OPS+ from 01-07 to support the intangibles theory isn't sound.

Not one to stop right away having seen that my argument was supported, I broke the cases down into groups by their OPS+/ERA+ before the trade. Maybe OK hitters (Junior had an OPS+ of 103) go up significantly more than others, or whatever. Here's a poorly constructed table showing the results. Notice something? That would be regression to the mean. Every group moved in the direction (and partially) by the amount that would be expected. The bad generally got better (except for Rey Sanchez, who started at an OPS+ of 76 and still managed to drop 45 points) and the good got worse (though Mark Teixeira did what he could to help Atlanta get to the postseason last year).

When the groups were split between pitchers and hitters, the former were the ones that actually improved (by 9 points, on average) while the hitters did worse (by about 3 points). They do say that when a pitcher faces a hitter for the first time, it's the pitcher that has the advantage - that may be what is going on here as in most cases the pitchers switched leagues and would thus be more likely to face batters unfamiliar with them. The batters switched leagues slightly less often.

My favorite part was that the changes actually fit a normal distribution (mostly).That doesn't necessarily mean that they're random, but I think it makes it difficult to support the notion that getting traded to a contender causes a player to play better. Sample size caveats and all that not withstanding, I think I've made my point. More could be done with this, such as looking at players going the other direction (from good to bad teams) or traded players in general. I made my argument here with numbers though,and USSMariner made a similar argument here with words - that's good enough for me.

Does all that mean that Ken Griffey Jr. won't play any better in Chicago? Of course not. He was actually a bit unlucky on balls in play earlier this season, and so his offensive stats should improve some. Then again, he is going to the harder league. Frankly, I don't really care about his bat that much (it's not outstanding and not horrible) - it's his atrocious defense (especially in center) that is going to make it unlikely for him to really improve the team. Now, one could make the argument that I didn't look at defense here, but it seems silly to me that a player would play better on defense, but not when hitting or pitching. It's a little harder to find a defensive metric to work with, so I'm going to pass on looking at that myself. Maybe instead I'll do the 1,526,039th study showing that "clutch" doesn't really exist. Or go look for some more living room furniture on craigslist - you know, something that might actually benefit someone. Read more ...

98 On The Black

John Walsh at the HardballTimes wrote a very interesting article about the effectiveness of the fastball. He broke the strike-zone into sections and looked at the result of fastballs at various speeds. A strike thrown down and away from a righty pitcher to a right-handed batter was up first: "Let me say that in plain English: An 87-mph fastball in this part of the strike zone is every bit as effective as a 96-mph heater." Here's the breakdown for each zone:



It's only on a pitch inside (especially thigh high) that throwing harder makes a big difference. To try to figure out why speed doesn't have as large an effect on pitches down and away, Walsh looked at various factors, one of which was called strike percentage (they are all technically in the strike-zone).

"Around 30 percent of all taken pitches in the low-and-away zone are called balls and the faster the pitch, the higher chance it has of being called a ball. Remember, essentially all these pitches are actually in the strike zone and should have been called strikes. So, here we see a definite disadvantage to throwing hard—you have a greater chance of a bad call from the umpire. That sucks doesn't it? Ninety-eight mph heater at the knees on the outside black: BALL! calls the ump. Same exact location, but now it's thrown at 86 mph: STRIKE! Maybe this is the reason that guys like Tom Glavine, who work the outside corner with soft stuff, seem to get more than their share of close calls."

The difference in inside vs. outside is largely the byproduct of home run rate:

"This is the percentage of balls in play resulting in home runs. Keeping in mind that we're looking at small sample sizes here, we see that a large part of the dependence on speed for inside pitches comes from the ability of hitters to drive the slower pitches out of the park. Outside pitches seem more difficult to hit for homers, no matter what the speed."

I've often heard it said that a large reason that velocity is a good indicator for success is that it helps make up for mistakes that a pitcher makes. This appears to be backed up by the data - if a pitcher misses at 91 it is more likely to get him into trouble than if he misses at 96. Also, Leo Mazzone's "locate the fastball low and away" philosophy makes a lot of sense - if a pitcher can execute it consistently. So, to sum up:

" 1. fastballs outside don't depend much (if at all) on speed for their effectiveness;
2. conversely, inside fastballs are more effective the harder they are thrown (this one I already knew);
3. most of the observed effect appears to come from home runs: outside pitches are rarely hit for homers and when they are, a fast pitch is as likely to be hit out of the park as a slow one;
4. a pitch thrown hard is more susceptible to the ump's bad call than a soft toss. "

Pitch/FX data is the next major advance in sabermetrics, and I'm am keenly awaiting the insights that can be drawn from it. Read more ...

Hit By Pitch And The MoQuo

After the O's victory today, I took a look at some hit batsmen statistics. I wanted to see if there was another stat that went along with HBP, as I have heard that pitcher with good control, such as Greg Maddux, actually hit more batters than wild pitchers (defined by their walk rates). I took all of the active pitchers with about 1,000 career innings pitched or more (so that there was a good sample size of data. There where 70 pitchers) and ran a regression of their HBP per Batter Faced (BF) against their HR / BF, BB / BF, K / BF, and Wild Pitches (WP) / BF. [Interestingly, the list of pitchers was riddled with guys that aren't very good such as Jamey Wright, Ron Villone, and Glendon Rusch.]

The results:

HR / BF - There was no relationship. Some guys hit a lot of batters but gave up few HR (Julian Tavarez); some hit few but give up a lot of HR (Brett Tomko); some have few of either (Tom's Gordon & Glavine). No one gave up a lot of both. I assume that is because a guy who hits a lot of batters and gives up the long ball wouldn't stick around long enough to get 1,000 IP.

BB / BF - A weak positive relationship (R-squared = 0.18). There was a small group of pitchers that hit a lot of guys, but didn't walk many. It included Bronson Arroyo and Pedro Martinez. On the flip side is Tom Gordon who hit few but walked many. In general though, the more guys a pitcher walked, the more guys he hit.

K / BF - There's actually a very slight negative relationship. Kerry Wood strikes out a ton of guys, but also hits a lot of batters. Trevor Hoffman had a very good K rate, but he almost never hit anybody. Jamey Wright hit batters more frequently than any other pitcher, but he also had the lowest K rate.

WP / BF - No relationship, and the data points are all over the chart. I was surprised by this, as I thought that a HBP would often by a WP but higher and inside.

The conlusion is that walk rate works, but not all that well. Since I already had the data up, I decided to look at how often the pitchers would put batters on base (BB + HBP / BF).

Everyone guess who had the lowest number there?

Did you guess Greg Maddux? It makes sense - he made a career out of great control. He's not #1 on the list though. (He's actually #2).

How about Tom Glavine? He pitched to the outside corner - that should limit HBP. Plus he had great control. Not even close; it turns out he had a slightly below average walk rate for this group and ended up #34.

The answer is (surprise!) Jon Lieber. Yeah, that Jon Lieber. Over his career, he has actually walked guys less frequently than Greg Maddux (or anyone else on here). #3 was Ben Sheets - I've talked about how awesome Sheets is before. In th #4 spot is the Moose, Mike Mussina. Trevor Hoffman is going to the Hall of Fame, but he's a spot behind Odalis Perez. Mariano Rivera is a couple spots back of Brian Mohler, and Glendon Rusch is ahead of Pedro Martinez.

I decided to do some more stuff here. I calculated each pitcher's MoQuo (Made-up name - I just called it that because it rhymes). It is equal to 100 * [1.5 * (K / BF) - 5 * (HR / BF) - 2 * (H + HBP / BF) - 0.2 * (WP * BF)]. There's very little reason for the weights, but I think it works OK. It's pretty similar to component ERA.

#1 was Pedro Martinez at 16.6. Aaron Harang is #19, and is the last pitcher above 0. Jamey Wright is last at -20.2, which is more than 5 worse than the next closest pitcher. Only five pitchers were above 10, and two were relievers - Mo Rivera and Hoffman. Glendon Rusch was ahead of Tom Glavine. It doesn't mean anything; it's just something to look at..

[For the record, Daniel Cabrera is at -11.9 - between Kent Mercker and Jason Jennings. Former O's Ace Erik Bedard is at 3.9 - between Mike Mussina and Josh Beckett. Keep in mind, these are career marks.]

The full list:
16.6 - Pedro Martinez [Man, Pedro is awesome.]
14.3 - Trevor Hoffman
13.0 - Mariano Rivera
11.7 - Randy Johnson
10.6 - Johan Santana
7.6 - Roy Oswalt
7.1 - John Smoltz
6.8 - Jake Peavy
5.3 - Ben Sheets
5.2 - Greg Maddux
4.1 - Mike Mussina
3.5 - Josh Beckett
2.5 - Brandon Webb
1.6 - Roy Halliday
1.3 - Javier Vazquez
1.2 - Kerry Wood
0.6 - Kevin Millwood
0.4 - Andy Pettitte
0.3 - CC Sabathia
0.1 - Aaron Harang
-0.2 - John Lieber
-0.3 - Tom Gordon
-0.5 - Aj Burnett
-1.5 - Tim Huson
-1.6 - Brad Penny
-1.6 - Jeremy Bonderman
-1.9 - Derek Lowe
-1.9 - Todd Jones
-2.6 - Matt Morris
-3.0 - Carlos Zambrano
-3.3 - Dontrelle Willis
-4.5 - Odalis Perez
-5.1 - Glendon Rusch
-5.1 - Tom Glavine
-5.4 - Mark Buehrle
-5.5 - Randy Wolf
-5.9 - Brett Myers
-6.3 - Esteban Loaiza
-6.7 - Barry Zito
-6.9 - Bronson Arroyo
-7.2 - Ted Lilly
-7.3 - Jake Westbrook
-8.0 - Livan Hernandez
-8.2 - Doug Davis
-8.2 - Vincente Padilla
-8.3 - Dave Weathers
-8.6 - Jamie Moyer
-8.6 - Mark Redman
-8.8 - Brett Tomko
-9.0 - Kenny Rogers
-9.5 - Adam Eaton
-9.9 - Brian Mohler
-10.2 - Ryan Dempster
-10.2 - Paul Byrd
-10.3 - Jarrod Washburn
-10.4 - Gil Meche
-10.4 - Kyle Lohse
-10.4 - Julian Tavarez
-10.7 - Miguel Batista
-11.3 - Kip Wells
-11.8 - Kent Mercker
-12.3 - Jason Jennings
-12.4 - Jeff Suppan
-13.3 - Jon Garland
-13.6 - Ron Villone
-13.8 - Darren Oliver
-14.7 - Ryan Franklin
-15.0 - Jason Marquis
-15.1 - Josh Fogg
-20.2 - Jamey Wright Read more ...

Psych, Psych, Super-Psych

I hadn't read this when it came out originally, but the HardballTimes had an article investigating the "hidden" effects of having a base-stealer on first as it relates to disrupting the pitcher. Here's the wrap-up:

"The goal of this study was to confirm or refute the notion that good base stealers disrupt the opposing pitcher/defense simply by his presence on first base. I reasoned that any disruption would show up in the performance of the batters who came to the plate with a Stealer on first base. I studied the performance of 219 batters (almost 3700 plate appearances) when one of the top 10 base stealers was on first base (with second base open) and compared that to what would be expected from those particular hitters. I then used a custom hit-trajectory matrix to convert the Stealer-on-first base performance into a defense-independent context that can be compared to the generic case. Finally, a small correction for pitcher quality was found to be necessary.

The results obtained indicate a small effect of disruption, amounting to about 0.17 in RC27, for the Stealer on first base situation. An independent cross-check was made considering what happens with any runner on first and again a very small effect of disruption was found (0.07 runs per game for that check). Finally, I found that pitchers hit a few more batters than expected when a Stealer was on first, although the effect was not statistically significant. It could be real, but we can't say for certain. Pitchers do commit more balks with a Stealer on first base.

Assuming these small effects are real, how much are these base stealer intangibles worth over a season? A typical Stealer is on first base for 180-220 plate appearances per season. That corresponds to about 128 outs (for this group), which adds up to 4.75 games. Assuming the improvement due to disruption is 0.17 runs per game, this gives a measly 0.8 runs over the whole season. Let's throw in an extra balk and a half-HBP (I didn't include the HBPs in the OBP calculation for simplicity) and we get an additional 0.5 runs (more or less), for a grand total of about 1.3 extra runs a year.
So, the next time you tune into the White Sox game and Hawk Harrelson is telling you that "Scotty" Podsednik, by virtue of his ability to disrupt the pitcher, is worth more than what his statistics show, well you now know he's telling you the truth. Podsednik is worth a little over one more run per season."

Since the 3,700 PA sample size was a bit small, it was expanded to the top 20 base-stealer.

"The result I found is a little more stringent: the 3.2 runs per season number goes to about two runs per season. So, instead of saying that Podsednik adds a little over one run per season due to his "disruptive" powers, we must content ourselves with saying that he very likely adds no more than two runs per season. "

This is a result that I've known to be true for a while (the "disruption factor" is very small overall) but it is interesting to see it quantified. Two runs is not insignificant (10 runs are about one win) but it is clearly hyperbole when the commentator goes on and on about how Brian Roberts is psyching out the pitcher. Now you'll know the truth. And isn't that really what we're after?

[I really wish I could do this kind of in-depth analysis. I guess I'll have to settle for understanding and appreciating it.] Read more ...

Said Earl Weaver - He's "lucky he's in f****** baseball."

Who remembered that as recently as 2003 Deivi Cruz (and his .269 OBP and 14 HR (career high!)) was the Orioles starting shortstop? I guess I blocked that one from my memory. Brook Fordyce was the #1 catcher and Tony Batista started at third that year and had 26 HR and 99 RBI to go along with his .270 OBP. Anyone still think RBI are important?

I found this out while doing a little bit of research. I read an article about how Terry Crowley shouldn't be the O's hitting instructor any more. His philosophy of an aggressive approach, and the poor results that followed, are usually cited as the reasoning behind a change. I completely agree that the offense hasn't been very good recently, and a better hitting instructor (one who preaches patience, working the count to get a good pitch to hit, and not swinging at every first pitch) should be found as soon as possible. Doing the same thing over and over again (keeping the Crow around) and expecting different results (an improved offense) is kid of crazy. The whole "aggressive approach" thing may not hold that much water, though. I went through the O's teams over past few years to check out the walk and strike-out rates of players that came over to the O's from other teams (and in some case, left again) or were here before and after the Crow. If Crowley is having a big impact on making the hitters more aggressive when they get here, it seems that such a thing would show up in their BB/PA and K/PA. I also took age into account, as players tend to walk more as the get older. There was no complicated math done - I just eye-balled the numbers to draw some conclusions. [I only picked players who were with the team for at least two years, which actually eliminated a lot of people.]

Ramon Hernandez: BB/PA improved a good deal but K/PA got a bit worse.

Kevin Millar: His BB/PA has been trending upward for years but it really went up after joining the O's. The K/PA is in the same neighborhood. He is getting older though, which is when this is expected.

Miguel Tejada: The BB/PA is about the same, but the K/PA (which had been improving) has started to regress some.

Corey Patterson: Corey's BB/PA is about what it was before, but he's been able to cut his K/PA a lot (if you can actually believe that).

Jeff Conine: The BB/PA worsened a little, but the K/PA got much better.

Javy Lopez: The BB/PA regressed a bit, but the K/PA was about the same.

BJ Surhoff (Crow became coach in his 4rd year in Baltimore): His BB/PA got worse (and then better after going to Atlanta and when he came back). The K/PA also got much worse at first, but imporved after going to Atlanta and coming back.

Tony Batista: As bad as he was for the O's, he actually improved both his BB/PA and his K/PA after joining the team.

Brook Fordyce: There was pretty much no change here.

Mike Bordick (Crow became coach in his 3rd year): His BB/PA dropped and his K/PA went up.

Delino DeShields: BB/PA went up with only a small increase in K/PA.

David Segui: Segui's BB/PA and K/PA both went up a good deal.

Albert Belle: His BB/PA was about the same, but he cut down his K/PA.

Cal Ripken Jr (Crow became coach in his 19th year): Cal's BB/PA and K/PA were both falling before 1999, but the BB/PA fell of a cliff and the K/PA went up. This was the end of the line for him, though.

Brady Anderson (Crow became coach in his 12th year): Like Millar, his BB/PA was already trending upward steadilly (about .010 BB/PA a year increase) which was not altered by Crowley. His K/PA only went up a little. [Speaking of Brady, Voros McCracken did some research to find the "flukiest" home run seasons ever (well, since 1920) - he compared a players career home run rate with their rate for their big season. Brady Anderson's 1996 50 HR explosion ranked #1. I remember being at school and hearing someone one the radio say that he had hit number 50. It was a very exciting time for the O's, as they made the playoffs that year (it was elementary school). At least they're not the Pirates, who haven't even had a winning season in 15 years (and that doesn't look like it'll end any time soon).]

Overall: This is certainly a small sample size, and it suffers by not identifying how much Crowley actually worked with each hitter. I would say that regular aging progressions are more responsible for the increased BB/PA and K/PA that is seen on average, and that while the Crow isn't particularly good at what he does, he isn't doing too much to harm the offense. It may just be a matter of getting better hitters on the team to improve the offense (duh). I would like a new hitting coach who may actually be a benefit, but I don't think it's that big of a deal. It is possible that Crowley's attitude contributed to guys like Jack Cust not being comfortable here and not getting much of a chance. If that's the case, then I think that that is the better reason to ship him out, as opposed to the bad offenses he has coached. Read more ...