Ten Things I Didn’t Know Last Weekby Dave Studeman
July 07, 2005
Every week, there's something new. Here is a list of my baseballistic ruminations over the last seven days.
Teams are loving that home cooking.
I don't know why, but teams are winning at home more often than they have in a long time. Through Tuesday's games, the home team has won 56.4% of all major league games, which is a higher rate than any year I looked at, going back to 1980. The overall average over the last 25 years has been 53.8%.
Particularly the Colorado Rockies.
Here are the teams with at least a .200 point difference in their home/road winning percentage this year:
Team Home Away Diff COL .525 .179 .346 HOU .649 .310 .339 CIN .533 .200 .333 OAK .619 .342 .277 WAS .744 .488 .256 MIL .605 .357 .248 TB .452 .205 .247 ATL .667 .444 .222 NYN .595 .395 .201The Rockies often have the biggest home/road differential in a given season, but .346 will be the biggest in their history if they maintain it over the full season.
Breaking things down a bit, the Rockies' offense has only scored half as many runs on the road as at home, while the pitching has been consistently bad regardless of the setting. On the other hand, the Nationals have allowed only 3.2 runs a game at home compared to 5.0 on the road.
Following is a list of the teams listed above, along with their runs scored and allowed records at home and away:
Runs Scored Runs Allowed Team Home Away Home Away COL 6.0 3.0 5.9 5.5 HOU 4.6 3.6 3.5 4.9 CIN 5.8 4.0 5.7 6.2 OAK 5.2 3.7 4.5 4.4 WAS 4.2 4.1 3.2 5.0 MIL 4.7 4.0 4.0 4.2 TB 4.7 4.4 5.4 7.1 ATL 5.3 4.4 3.8 4.1 NYN 4.4 4.6 4.0 4.8For most teams, the reason they play well at home lies primarily with either the offense or the defense, except for the Houston Astros, who seem to just play better at home every which way.
The Astros have been winning everywhere lately.
Speaking of the Astros, they are on a roll. Here are the top ten teams in overall winning percentage the past three weeks:
OAK .778 LAA .778 HOU .778 CHA .765 ATL .737 WAS .706 STL .667 CLE .650 BOS .647 NYA .611I knew that Oakland was on a run, and the other teams on this list are in the thick of their pennant races. But Houston snuck up on me. They are now tied for second place in the National League Central, and I'm sure Astro fans are hoping for a second-half run similar to last year's.
The Astros have basically had two half-seasons in the past half season. (Does that make them quarter-seasons?) As you can see on the following graph, they've been a different team since the beginning of June.
Andy Pettitte has been awesome the last month and a half, making Pettitte, Roger Clemens and Roy Oswalt the Big Three everyone thought they could be. And Lance Berkman found his stroke after his first month back, joining Morgan Ensberg, Craig Biggio and Jason Lane to form a pretty decent lineup.
Home teams win more one-run games.
I promised that I wouldn't talk about one-run games after last week's column, but Walt Davis posted a fascinating fact on the Baseball Primer discussion board: the home team wins one-run games about 60% of the time.
Over the last five years, home teams have a .616 winning percentage in one-run games and only a .509 winning percentage in all other games. In other words, one-run games account for nearly the entire home field advantage in baseball.
I don't really know why this is so. Batting last in extra-inning games explains part of this trend, but only part of it. I have a feeling there's more to learn here. By the way, home teams are winning one-run games at a .619 clip this year, in line with past trends.
Bullpens can make a difference in close games, and we know which ones.
In that one-run games article, I also discussed the potential impact of bullpens on close games. It occurred to me that I could learn a bit more by using the bullpen stats from the wonderful Baseball Prospectus site.
Baseball Prospectus tracks some very interesting bullpen figures, including Win Expectancy and Leveraged Index for relief pitchers. Leveraged Index is a measure of the importance of a relief pitcher's appearance; that is, how important his appearance was to the outcome of a game. As you can imagine, the overall Leveraged Index of a bullpen is related to the number of close games a team plays, as this graph shows:
Basically, the more close games a team plays, the higher the Leveraged Index, and any material differences are probably the result of a manager's specific bullpen usage.
Bear with me on this next step. We can use this information to gauge how instrumental a bullpen has been to its team's record in close games. The key is that we can use the Leveraged Index to project the bullpen's Win Expectancy Added, and if a bullpen beats that projection, it has probably contributed more to those close wins. If you're not sure what these terms mean, you might want to review the Win Probability article I wrote last December.
In other words, by "normalizing" a bullpen's Leveraged Index, which is driven by opportunity rather than skill, you can uncover the true contribution of the bullpen to close games.
Maybe this graph will help. In the "X" axis I've graphed each team's record in close games, and in the "Y" axis I've graphed each team's variance in bullpen Win Expectancy from projected Win Expectancy (based on Leveraged Index). If a team is above the line, that means it has probably contributed positively to the team's record in close games. Below the line, it hasn't:
You can pretty safely say that the Nationals and Cardinals, who have excellent records in close games, have been helped by their bullpens in close games quite a bit. But the White Sox's excellent record in close games is less attributable to the bullpen, and probably more attributable to starting pitching and offense.
They've been tracking something like Win Probability at the Nats Blog.
Based on the above, is Nationals' closer Chad Cordero the National League MVP? Or is Peter Gammons correct when he claims Jose Guillen is their MVP? Well, the good news is that the Nats Blog has been tracking each Nationals' game this year with a pseudo-Win Probability system. It's called Win Value, and it provides a fascinating, win-based look at a pennant-contending team.
The guys at Nat Blog posted Win Value leaders for the first half of the year, and Chad Cordero is virtually tied with Nick Johnson for the team MVP title. Guillen is a distant third.
Congratulations to the Nats Blog for a great job. This information should provide great fodder for future columns.
Sometimes you've got to look between the numbers.
The USS Mariner ran a great study of why it's important to look at context as well as raw stats. They specifically looked at Willie Bloomquist's stolen bases in 2004 and 2005 and found that the guy really has stolen bases when it mattered. Stolen bases are a great example of looking at more than just raw stats because, unlike home runs, you can choose to steal when it matters most.
Tom Ruane posted a long article about Value Added Evaluations on the Retrosheet site that assessed the value of each offensive event based on the context in which it occurred. I found the section on stolen bases most interesting, because he compared the general value of stolen bases and caught stealings against their actual "run value" for all basestealers of the past forty years.
Here's a list of the players whose stolen base records added more runs than their raw totals indicate:
Name SB CS VASBR LWSBR DIFF Paul Molitor 504 131 61.1 41.6 19.5 Pete Rose 198 149 -3.2 -22.5 19.3 Rod Carew 353 187 15.3 -3.0 18.3 Carney Lansford 224 104 12.1 -1.1 13.2 Willie Wilson 668 134 89.6 76.5 13.1 Ron LeFlore 455 142 47.0 34.2 12.9 Lou Whitaker 143 75 7.9 -4.2 12.1 Tony Phillips 177 114 -5.7 -17.0 11.3 Brian Downing 50 44 1.9 -8.7 10.6 Dave Winfield 223 96 11.7 1.4 10.2Paul Molitor was a great, heady player, and this kind of analysis illustrates that. It's interesting that former leadoff hitter non-base stealer Brian Downing is also on this list, as well as Pete Rose, whose understanding of when a steal mattered (or ability to avoid a crucial out on the basepaths) helped to offset his overall poor record as a basestealer.
Here are the largest negative differences:
Name SB CS VASBR LWSBR DIFF Chuck Knoblauch 407 117 5.2 21.0 -15.7 Jose Cruz 317 136 -11.1 2.4 -13.5 Julio Cruz 343 78 26.4 36.6 -10.2 Tony Bernazard 113 55 -11.8 -1.7 -10.1 Von Hayes 253 97 -1.4 8.0 -9.4 Mike Cameron 216 59 2.4 11.4 -9.0 Bert Campaneris 649 199 42.9 51.4 -8.5 Brady Anderson 315 100 4.9 13.0 -8.2 Felix Jose 102 57 -12.9 -5.1 -7.7 Pokey Reese 144 26 7.4 14.5 -7.1 Kenny Lofton 545 145 26.6 33.7 -7.1Chuck Knoblauch was a fine basestealer, but his record of successfully stealing when it most counted was only slightly better than breakeven.
Clutch hitting: maybe, maybe not.
Tom Ruane also posted a study at Retrosheet in which he took a different look at clutch hitting to see if he could find concrete evidence for its existence. This is a hot topic because the "Fog crowd" has focused on clutch hitting studies as a test of Bill James's contention that year-to-year correlation studies aren't adequate to uncover trends that can get lost in the fog of large contextual variances. Or something like that.
So Tom decided to look at career stats instead of year-to-year correlations, and still couldn't find evidence of clutch hitting. Which is interesting, because Andy Dolphin also conducted a study of career clutch hitting stats and found definite evidence of its existence.
I know I'll be accused of being an anti-sabermetrite, but I am willing to go on record as saying that some players are better clutch performers than others. I also predict that I will not be the one to create the breakthrough study that finds them.
AL run scoring is returning to its "normal level."
Earlier this year, we noticed that, for the first time in a long time, the American League wasn't scoring more runs per game than the National League, and we wondered if this would persist.
You'll be glad to know that things are back to normal. The AL is scoring 4.8 runs a game over the full year, compared to 4.5 in the National League. In three of the last four years, the AL average has been about 4.8.
Catcher's Glove Hands Take a Beating
Okay, you probably knew that. But you might be interested that a new study in the Journal of Bone and Joint Surgery found that catcher's glove hands are most likely to have permanent damage. According to this New York Times blurb (registration required), the fault lies with the kind of mitts catchers have been using since the 1960's.
Can you believe that If you build it, he will come is only 39th on the list?
References and Resources
Baseball Prospectus's definition of Leveraged Index is not the same as TangoTiger's, who coined the term. You can read more about Tango's work here.
There is one other study on clutch hitting I should have mentioned. Elan Fuld, a student at the University of Pennsylvania, received some publicity when a Penn press release mentioned his study of clutch hitting. Fuld also found evidence of clutch hitting, and you can download his study from his personal site.
Dave was called a "national treasure" by Rob Neyer. Seriously. Comments about this article can be sent to him through the miracle of e-mail.
<< Return to Article