Archive for May, 2005

Introducing PrOPS

The Hardball Times has published my latest research project, Introducing PrOPS. Using some of THT’s excellent hit-type data I develop a metric that predicts player hitting performance (OPS). The goal of this project was to separate out hitting luck from actual performance. You can think of Predicted OPS or PrOPS for hitters as similar to DIPS and Component ERAs for pitchers. If you’re interested, read the article for more information. This is just my first attempt to developing this metric, so please feel free to send me suggestions for improvement.

If you’re still confused as to what I’m trying to find with PrOPS, this quote from Moneyball captures the spirit .

[W]hat is a double? It really isn’t enough to say that a double is when a runner hits the ball and gets to second base without a fielder’s error…. There are lucky doubles and unlucky outs. To strip out the luck what you need, really, is something like a Platonic idea of a double.

So, I guess you could say I’m trying to measure a Platonic OPS. Maybe I should have called it PlOPS?… No.

Some interesting things I’ve noticed from looking at the PrOPS numbers:

  • Jason Giambi hasn’t been as bad as his stats: PrOPS=.891 versus OPS=.736
  • The A’s are due: Not one single player has an OPS >= PrOPS. They are all hitting better than their stats.
  • Tony Pena maybe should have stuck around for some better times: Only 3 players have an OPS>PrOPS. The next manager may get some undeserved credit when the bounces even out.

Studes on Balls in Play

Studes posts some interesting findings on what impacts hits on balls in play over at The Hardball Times this morning. I’ll summarize his findings if you promise to read the article.

  • He improves FIP ERA by adjusting for the percent of fly balls that become home runs. (Braves fans, don’t look at Mike Hampton’s numbers.)
  • He develops predictive models of hitter BABIP and HR rates based on a few simple metrics (line drives, fly balls, strikeouts, etc.).
  • And just in time for Star Wars: The Revenge of the Sith, he rolls out an interesting new stat to measure hitting power known as “Force.” The stat has a lot of potential (LD% + HR/OF fly), but will be most useful for prospect discussions among geeks: “The force is strong with this young prospect.”

More Stats on the Web

Good news for those of you looking for more baseball stats. For the general baseball enthusiast out there The Hardball Times has started posting its 2005 player stats. This page is a must for sabermetricians looking for their favorite pet stats. THT provides a great service. Thanks to Studes you no longer have to break-out the spreadsheets and torture html tables to view the stats you want to see.

And for the Braves fans out there, John Wright has started tracking advanced pitching stats for the Braves relievers. WPA, WPP, etc. ; they are all there. Unfortunately, John has confirmed that sinking feeling we all have when Dan Kolb enters the game. He also posts game comments on each game. Stop by and say hi, and don’t forget to check out his blog, Shots from the Bleachers.

More Arguing Balls and Strikes

I’ve investigated the impact of managers on balls and strikes Questec and non-Questec parks a little further. In my earlier post, I looked at the effect only on the pitching side. Now, I want to report what I found for hitters. I used the same methodology (switching the relevant control variables from pitchers to hitters) to examine the managerial impact on the strikeout-to-walk ratio for hitters. If a manager has influence over ball and strike calls in a way that helps his hitters, then I would expect the K/BB to be higher in Questec parks than non-Questec parks (the opposite of what it was for pitchers). The results below are regression coefficients estimating the impact of managers in Questec parks. I re-report the managerial impact on K/BB for pitchers for comparison purposes.

Manager                Hitting        Pitching
Hal McRae        	1.468        -0.487
Bruce Bochy        	0.741        -0.409
Frank Robinson        	0.472        0.271
Ken Macha        	0.457        -0.967
Felipe Alou        	0.456        -0.085
Luis Pujols        	0.403        0.428
Jerry Narron        	0.319        -0.283
Bobby Cox        	0.239        -0.275
Jeff Torborg        	0.140        -0.084
Jimy Williams        	0.073        -0.584
Lloyd McClendon        	0.015        -0.406
Jack McKeon        	0.000        -0.255
Tony Pena        	-0.039        0.215
Jim Tracy        	-0.083        -0.869
Clint Hurdle        	-0.099        -0.494
Alan Trammell        	-0.139        0.289
Mike Scioscia        	-0.162        -0.453
Jerry Royster        	-0.200        0.000
Bob Boone        	-0.202        -1.017
Tony LaRussa        	-0.204        -0.966
Ron Gardenhire        	-0.311        0.749
Joe Torre        	-0.329        -0.639
Carlos Tosca        	-0.362        -0.024
Mike Hargrove        	-0.376        0.348
Lou Piniella        	-0.379        -0.230
Bob Melvin        	-0.493        -0.114
Jerry Manuel        	-0.497        -0.156
Grady Little        	-0.536        0.197
Buck Showalter        	-0.558        0.483
Eric Wedge        	-0.623        0.369
Dusty Baker        	-0.641        0.024
Art Howe        	-0.646        0.160
Larry Bowa        	-0.693        -0.743
Bobby Valentine        	-0.862        -0.387
Bob Brenly        	-0.996        0.771
Ned Yost        	-1.920        0.113

Bold: statistically significant at the 5% level

Hal McRae seems to lose the most from his hitters when playing in a Questec park, which suggests he may be the best arguing balls and strikes for his hitters. However, the estimate is not statistically significant, so it’s unclear it the effect is real. Only Ned Yost has a statistically significant impact, in a way he might prefer not to, as his hitters seem to do much better in Questec parks. Bob Brenly is having trouble again. Maybe he’s just used to managing in a Questec park in Arizona, and therefore gets taken by other managers on both sides of the ball. Or it may be that the team assembled under him was just well-accustomed to Questec calls and played poorly when out of that environment. The only problem with this explanation is that other managers with Questec home parks don’t seem to show the same tendency. But, we must remember that the results is not statistically significant, so it’s probably not worth thinking about too hard. While the “best” estimates of Brenly show him not doing well, they are not statistically different from no effect given the standard error. Overall, it looks like most managers seems to have little impact over balls and strikes called for his hitters.

There’s one other interesting test I did, which was to compare the estimated impact for both hitting and pitching. If a manager raises the K/BB for his pitchers does he then cost his hitters and vice versa? Or, if he’s good at working the umps is he able to help pitchers and hitters at the same time. A simple correlation between the impacts generates a correlation coefficient of -0.28, which is statistically significant at about the 10% level. This gives some support to the idea that managers may have some ability to influence umpires on both sides of the ball. That is, a manager that is able to raise the K/BB for his pitchers is associated with lowering it for his hitters. In the end, in most casese I don’t think managers are influencing balls and strikes when they argue, but it’s possible that Questec just isn’t changing any impact they have. Given all of the complaining by the umpires, I tend to believe the former is true.

Arguing Balls and Strikes: The Most Influential Managers

In an earlier post I discussed the possibility of managers influencing the calling of balls and strikes by umpires. This was, of course, motivated by Tony LaRussa’s recent comments about Bobby Cox’s excessive complaining to the umpires over the strike zone. Using Questec as a controlled environment, where complaining would be less effective, I looked at how Cox’s and LaRussa’s pitchers performed, in terms of the strikeout-to-walk ratio, in Questec and non-Questec parks. Simple t-tests indicated that while Cox had no influence, LaRussa’s pitchers performed significantly worse in Questec parks, where lobbying the umpire should be less effective. While this test was fun, there could be many outside factors that explain the difference. What about the park, batter quality and pitching quality? These could all contribute to explaining the difference. So, I decided to take it up a notch to control for these factors and examine the lobbying impact of all the managers in the sample.

The data I used is the same as before. I looked at managers in away games (eliminating any impact from home field advantage) in 2002 and 2003. Instead of just concentrating on Cox and LaRussa, I used a sample of all managers who managed at least 60 away games over this time period. I used a multiple regression to estimate the impact of various factors on a visiting team’s strikeout-to-walk ratio in each game. The factors I controlled for were:

  • the average K/BB of the pitchers on the visiting team on the road in that year
  • the average K/BB of the batters on the home team at home in that year
  • the park in which the game was played (using indicator dummy variables)
  • the league of the home team (to control for the designated hitter effect)
  • the manager of the visiting team (using indicator dummy variables)
  • whether or not the game was played in a Questec park
  • the manager of the visiting team in Questec parks ( an interaction term to pull out the individual effect of each manager in Questec parks

The results were quite interesting. And once again, the evidence indicates that LaRussa should not be the one complaining about the behavior of other managers. Here is a list of managers and their estimated increase/decline on their teams’ strikeout-to-walk ratio in Questec parks.

Manager        		Marg. Impact
Bob Boone    		-1.017
Ken Macha    		-0.967
Tony LaRussa    	-0.966
Jim Tracy    		-0.869
Larry Bowa    		-0.743
Joe Torre    		-0.639
Jimy Williams   	-0.584
Clint Hurdle    	-0.494
Hal McRae    		-0.487
Mike Scioscia   	-0.453
Bruce Bochy    		-0.409
Lloyd McClendon   	-0.406
Bobby Valentine    	-0.387
Jerry Narron    	-0.283
Bobby Cox    		-0.275
Jack McKeon    		-0.255
Lou Piniella    	-0.230
Jerry Manuel    	-0.156
Bob Melvin    		-0.114
Felipe Alou    		-0.085
Jeff Torborg    	-0.084
Carlos Tosca    	-0.024
Jerry Royster    	0.000
Dusty Baker    		0.024
Ned Yost   		0.113
Art Howe    		0.160
Grady Little  		0.197
Tony Pena    		0.215
Frank Robinson    	0.271
Alan Trammell    	0.289
Mike Hargrove    	0.348
Eric Wedge    		0.369
Luis Pujols    		0.428
Buck Showalter    	0.483
Ron Gardenhire    	0.749
Bob Brenly    		0.771
Bold: statistically significant at the 5% level
Italics: statistically significant at the 10% level

Update: I corrected the estimates from my previous posting. Brenly, Hargrove, and Gardenhire no longer have statistically significant impacts. Jim Tracy’s estimate is now statistically significant.

As you can see, most managers experience no statistical difference between Questec and non-Questec parks (rounded to thee decimal places). Only 3 managers have a statistically significant impact on K/BB. And right at the top of the list is Tony LaRussa. LaRussa’s ability to sway umpires is not surprising considering that he is a lawyer, which is about as a well-kept secret as John Kerry’s military service. I wouldn’t be surprised if LaRussa is also good at manipulating MLB’s umpire complaint department.

And then look at Bob Brenley (although his impact is large it’s not statistically significant. Why is he so “bad?” Well, there may be some as yet to be explained factors, but I’ve certainly never been convinced by anything he’s said during a TV broadcast. In fact, I’ve changed opinions based on the fact that he agreed with my thinking. I wonder if his arguing skills are so weak that he ends up hurting his pitchers. Maybe other managers take advantage of him when Questec is not around, and he can’t counter with anything. Or, it could be that protects his hitters to the extent that it hurts his pitchers, which was not such a bad strategy given the good pitching he had during these years.

Overall, most managers don’t seem to have any real impact in arguing balls and strikes, which is consistent with rent seeking outcomes. Everyone wastes energy arguing but nothing is gained. At least maybe Questec can get the excessive bickering over balls and strikes out of the game.

Given that Boone is no longer managing, I think it’s safe to say that Tony LaRussa is the best rent seeking manager in the game today. Although, rent seeking is generally considered to be bad, it certainly is good news for the Cards. So, I think it’s a quality to be admired, but certainly Tony LaRussa has no moral high ground to accuse any other manager for influencing the game through umpires.

Does Bobby Cox Win by Rent Seeking?

That’s the implication Tony LaRussa gave in some recent comments about the Braves.

Rent seeking is “the socially costly pursuit of wealth transfers,” in the words of my former professor Bob Tollison. Most of the time we discuss rent seeking in law-making; for example, lobbyists spend effort to encourage legislators to adopt policies that give interest groups legislated benefits. Rent seeking involves expending energy to change the rules of the game to your benefit, instead of using resources to seek rewards within the rules of the game. In baseball, it means mangers, players, and fans attempting to influence the calling of the rules of the game. Now, the social loss of rent seeking in the game is a bit complicated, but, from my standpoint, watching managers both plead theie cases to an umpire just so we end up where we started in the rulebook is a waste. We could argue over that, though.

Although LaRussa toned down his assertions a bit from Saturday, on Sunday he had a little more to say about Cox’s influence on the game.

“All I said was that when you go against somebody who is complaining all the time, you worry about how it affects the umpires,” said La Russa, who doesn’t think he should be fined for his critique of Dale Scott’s strike zone in the seventh inning when he and pitcher Al Reyes were ejected.

“When I was a young manager, Earl Weaver at Baltimore played it that way. They challenged ever strike. If your wife is getting on you every minute, you get fed up. Pretty soon, you may tune her out and do whatever it takes to get through it. If I was an umpire, I’d put [the Braves] in their place.

“I didn’t say anything to challenge anybody’s integrity. I have great respect for Bobby. He’s out to beat you. He does all that stuff sincerely. I didn’t say anything personal about Bobby Cox.”

Well, the last stuff kind of seems to contradict the first stuff, but that’s not what I’m really interested in. LaRussa makes a bold claim about the Cox-managed Braves that I think is legitimate: Cox’s complaints to the umpire influence the game. I strongly believe that players, managers, and fans harass referees/umpires in a way that can effect the outcome of the game. I don’t think it’s a huge problem, nor necessarily something that any sports league needs to address. Heckling and complaining have long been a part of the game and will persist. I will admit that I think it can go too far and some restrictions would be beneficial.

But in reference to LaRussa, he seems to be saying something further than this; namely, that Cox has more influence than other managers. He’s a successful rent seeker, stealing games by manipulating umpires rather than outfoxing other managers on his way to glory. As someone who has found the success of Braves pitchers quite exceptional without any particular explanation as to why, I wondered if this might be the reason. So, I put on my economic thinking cap and went to work. How could I identify managers influencing umpires? Questec.

Questec is system of cameras and computers that track the pitches in a game. In 2002, MLB installed Questec cameras in 10 parks to track the ball and strike calls of umpires. The umpires were none too pleased, and even players (most notably Tom Glavine and Curt Schilling) complained. While the impact of Questec on the game is a little unclear, one thing it does is keep umpires honest. Think of this situation without Questec. Cox is screaming from the dugout, “You stink ump. Can’t you see the strike-zone.” While I believe umpires are resistant to such calls, I can imagine that on close calls umps may, even subconsciously, favor the complainers. (Fellow professors know what I’m talking about.) Cox might be able to gain a strikeout or lose a walk for his pitchers just by being an ass. But with Questec, the umpire answers not just to managers, but to MLB. Umpires, while they may be swayed when Questec is operational, I think there is less movement for bending.

How can I use Questec to examine Bobby Cox’s rent seeking? One nice feature of MLB’s experiment with Questec is it’s inclusion in only some parks. This article by Nate Silver and and Keith Woolner (with an aside by Rob Neyer) identifies the ten parks with Questec, which was installed beginning in the 2002 season. As best I can tell, the usage of Questec has not changed since their first installment. I then used Retrosheet game logs to look at all of the visiting games (to avoid the home fan/park influence) in which Cox was the manager inside and outside Questec parks for 2002 and 2003. I compared his pitchers’ average strikeout-to-walk ratio in these parks to non-Questec parks. Here is what I found.

Questec K/BB (46 games) = 2.86
Non-Questec K/BB (115 games)= 2.81

A t-test reveals that the difference in not statistically significant, and if anything his pitchers do better in Questec parks. Now, of course, this does not prove Cox has no influence, only that from looking at the raw numbers no influence is observable. Certainly, pitcher and team quality may be important. Sorting through all of that is a big project that I may tackle one day, but today is not that day.

But, I didn’t want to end there. I wondered what LaRussa’s influence might be, since he made such a big deal about it. Well, it turns out that Tony may not be as innocent as he claims.

Questec K/BB (53 games) = 2.00
Non-Questec K/BB (109 games)= 2.75

It turns out, that this estimate is statistically significant at just about the 1% level. LaRussa does do better without Questec in the house to keep the umpires honest. Well, this certainly doesn’t imply that LaRussa is necessarily the cause, but it is fun to point out. While there is no evidence of Bobby Cox winning games through the means LaRussa suggests, the same cannot be said about the man making the accusation.

Note: I would like to add that for games with zero walks, I assigned one walk to the team. When I exclude these observations, the results do not change.

Cram it, Tony

I’m just going to come out and say it: I don’ t like Tony LaRussa. I think he’s a conceited jerk who has a wee bit too much confidence in his own intelligence and morality. Here are his comments on last night’s Cards/Braves game.

We had Mondesi struck out twice [in that inning] and we had Chipper out on the same pitch. …[The umpire] Scott lost the aggressiveness of his strike zone.

I watched the whole game, and I admit I thought the balls and strikes were not called well all night. Although, I think the calls were universally bad for both teams. I recall, in particular, that Pete Orr took a crucial called strike-two that was nowhere near the zone. Pete’s pretty much a free swinger and the replay showed the ball to be low and inside. LaRussa also forgot to mention that Chipper should have been on first in any event since Chipper was hit by the pitch in his injured foot. The umpire somehow thought Chipper’s jumping around in severe pain must have been an act.

But, you know, that kind of comment doesn’t bother me. All managers complain about umps, so that’s fine. But LaRussa’s next statement requires a personal apology to Bobby Cox.

The only thing they [Braves] should not be respected for is the way they [complain] and moan and beat the umpires down. That’s beneath the class of their organization.

What? Give me a break, Tony. Did you you see Chipper complain when your pitcher hit him in the foot? No. He stood in there an earned a walk. I understand Tony was upset about having two runs taken off the board when the umps determined that a ball had struck the baserunner Pujols. But, the replays seemed to confirm that the ball did hit his player. Shouldn’t the call be reversed? Nope, it’s those whining Braves who are at fault.

This reminds me of an article that appeared in Sports Illustrated about a month ago. In an excerpt from the recent book written about him (I won’t link to it deliberately), LaRussa philosophizes about his decision to plunk Luis Gonzalez in the ribs in a situation that really hurt his own team. I was embarrassed that LaRussa thought he’d reached a decision based on anything but spite (“It’s not personal, Sonny. It’s strictly business.”). I felt like vomiting at the end. I hope Gonzalez spits in his face the next time they meet. And I won’t be surprised if LaRussa orders a cowardly plunking again today.