Archive for August, 2006

Indians Comeback!

After being down 10-1 after the first inning, the Indians pulled off the comeback of the year to win the game 15-13 in extra innings. This is one of the craziest win probability graphs I’ve seen this season.


It even bests the August 9th comeback last season, when the Indians overcame a 5 run deficit by scoring 11 runs in the top of the 9th.


Cabrera Mr. Clutch for Marlins?

Browsing the player captions in my Sportsline fantasy league, I noticed that they had this to say about Miguel Cabrera:

News: 3B Miguel Cabrera has been Mr. Clutch for the Marlins all season, but he has turned it up in the second half. Cabrera had batted .395 with runners in scoring position since the All-Star break going into Monday, fourth in the NL.

Cabrera is in fact #1 by a long shot for the Marlins with a 3.39 WPA this season. This ranks him 14th in baseball and 7th in the National League. Since the All-Star break he’s wracked up a 1.13 WPA, good for 32nd in baseball during that same time period.

Yet since the All-Star break, part time 1st-baseman and pinch hitter, Wes Helms has a WPA of 0.93, nearly as good as Cabrera, with Josh Willingham not far behind with 0.89. Cabrera and Helms have had consistent positive contributions compared to Willingham, whose main contribution was a 2-run walk off homer against the Mets on August 1st. That single shot was worth 0.70 WPA, nearly all of Willingham’s post All-Star value.

Cabrera’s actual Clutchiness for the season is -0.35, so while he’s certainly been the most valuable Marlin, he’s hardly been contributing above and beyond what a “non-clutch” player would with the same stats. Hanley Ramirez actually leads the Marlins this season with 0.75 Clutchiness, while since the All-Star break, Willingham leads the team with a 0.81 Clutchiness.

Crede and Dye

Joe Crede is finally hitting the way Baseball America thought he would. Remember, when Crede was a 24-year-old third baseman in AAA Charlotte, he batted .312/.359/.571 with 24 home runs in just 359 at bats. His minor league career was filled with honors (twice the MVP of his league and twice the White Sox’s minor league player of the year) and an All-Star major league career seemed inevitable.

Before this year, it hadn’t quite happened. Over the last three full years of play, he’s batted .261, .239 and .252 with a high of 22 home runs. His OPS (On-Base plus Slugging) was below the league average each year, as was his batting WPA . Despite his fine glove and World Series heroics, Crede was considered a disappointment by most White Sox fans.

Something has clicked for Joe this year. He’s batting .298/.333/.545 and he’s already set a career high with 25 home runs. He’s creating 6.8 runs a game and his batting WPA is 1.53, making him three wins better than average.

As you can see on the following graph of his Runs Created per Game, he’s been consistently fine this year, suggesting he truly has moved up to another level of production.


How has he done it? He’s striking out less and hitting more flyballs, without losing any power. Crede has always managed to put the bat on the ball, but this year he’s been particularly adept at it:


When a player makes better contact, you might expect him to give up something in power, but 13% of Crede’s outfield flyballs have been home runs, the same as last year. His home run totals are up because he’s hitting more outfield flies:


The White Sox’s Jermaine Dye is also clearly having a career year. He’s creating 9.7 runs a game and his WPA is 4.25. But his production is driven by a big jump in his BABIP (from .286 last year to .353 this year) and flyball production (17% home runs last year to 25% this year). No one should expect him to maintain that pace.

Joe Crede’s story appears to be different. Of course, he could also go back to his old self the rest of this year and next year, but the stats indicate something more permanent. At the age of 28, Joe Crede is finally meeting the expectations folks had for him.

New Features: More Stats & Integration

Quite a few feature additions just went into production.

-First off, WPA is no longer expressed as a percentage. Every player still has the same value, it’s just 100 times less.

-WPA is now displayed in each player’s game log and all the dates in the game logs are linked back to the correct Win Probability graph.

-There’s a new table for both batters and pitchers in the player stat pages that include the following Win Probability & Leverage stats:

WPA: Win Probability Added.
-WPA: The total of a player’s negative contributions towards their team’s win/loss.
+WPA: The total of a player’s positive contributions towards their team’s win/loss.
pLI: Average Leverage on a plate-appearance basis.
inLI: Average Leverage when a pitcher starts an inning.
gmLI: Average Leverage when a pitcher enters the game.
exLI: Average Leverage when a pitcher exits the game (game ends not included).
Pulls: Number of times a pitcher has been pulled from a game before it ended.
G: The number of games pitched in.
phLI: Average Leverage when pinch-hitting.
PH: Pinch-hitting opportunities.
OBP Wins: Player’s wins in a context-neutral environment.
Clutchiness: Difference between WPA and a Leverage adjusted OBP Wins.

For more information about these stats, take a look at the following links:

-Tangotiger’s Critical Situation Series: Part 1, Part 2, Part 3.
Clutchiness: The Blog.
The One About Win Probability.

If you notice any problems, please let us know and we’ll fix them as quickly as possible.

New Feature: Find The Right Game

In the past, navigating you’re way through the Win Probability graphs was probably a bit of a pain since you had to go game by game until you got to the correct date. This was finally fixed today with the addition of a calendar where you can easily go directly to the game you’re looking for.


Additionally, the calendar is color coded so you can see which games were won (green) and lost (red) on each particular day they played. This may bring back some painful memories of some of the worst months in recent history like the Tigers’ 3-win April in 2003, or the Orioles 24-loss September back in 2002.


All the best months are in there too, such as the Athletics’ 17 straight wins to finish August in 2002 and the Red Sox’s 21-win August in 2004. Try it out and let us know what you think!

Update: It should be working for mac users now, but the formatting is going to be a bit off in safari until I can figure out what’s causing the issue.

More Win Probability!

You may have noticed that the Win Probability section has changed slightly. Thanks to our stats provider, Baseball Info Solutions (BIS), we now have 2002-2006 game logs, which we have used to compile the 2002-2006 Win Probability graphs and stats. Note that we were not using any of BIS’s data previously to compile Win Probability stats and you may notice that a few of the values have changed.

The data BIS provides is much more accurate than our previous data source and as a result there should be few-to-no errors. If you do happen to find an error, chances are it will be corrected the next day.

Hopefully there will be quite a few updates in the next week or two which should include:

– Play-by-Play for each game with Win Probability stats.

– Additional leverage statistics for relievers and pinch hitters.

– Win Probability +/- breakouts.

– Season Leaderboards. (FINALLY!)

– All Win Probability and Leverage stats in the regular player stats pages and game logs.

In the meantime, the 2002-2005 Win Probability data looks quite interesting, especially if you’re still debating the 2005 AL MVP race.

Carlos Lee and Kevin Mench

A week ago, the Rangers and Brewers swapped leftfielders and a few other players. Texas acquired Carlos Lee and minor leaguer Nelson Cruz for Kevin Mench, Laynce Nix, Francisco Cordero and a minor leaguer. At the time, I thought Cordero (an A reliever most of his career) was the key to the deal, and I called Mench a “poor man’s Lee.”

Most Internet posters seem to think that the Rangers got the best of this deal. For instance, ESPN’s Keith Law opined…

Unless the Brewers have a second move in mind involving Mench, Cordero, or Turnbow, it’s hard to see how this is a good return on arguably the most attractive position player on the trade market.

But the erstwhile MGL, in this thread feels that once you include fielding and baserunning, Mench is actually a better player than Lee — and he’s cheaper to boot. So I thought it would be fun to compare the two. Let’s start with a basic Runs Created graph, showing each player’s Runs Created over their career:


You’ve got to say that the Brewers picked a fine time to trade Lee, who is having the best year of his career and will be a free agent at the end of the season. Mench isn’t having as good a year, but his production was very similar to Lee’s prior to 2006.

Breaking down their stats a little, Mench and Lee have exhibited the same level of on-base skill throughout the years…


..but the difference between the two this year has been their power.

Almost 18% of Lee’s outfield flies have been home runs, compared to a previous average of about 13%. Given his track record, I’d say it’s highly unlikely he will maintain that rate for the rest of the year.

In comparison, Mench has kept his home run/outfield fly rate at about 11% (aided by his old home park), but his 2006 slugging decline is more related to a higher groundball rate (42% vs. a previous career average of 36%). That could be a disturbing trend, because changes in batted ball rates can signal abrupt changes in a batter’s true performance. At least, that’s my hypothesis. Maybe I’ll test that someday…

If you had compared Lee and Mench at the end of last year, you might have said that Lee has a slight edge in power but not much else. Does this year–particularly Mench’s increase in groundballs–change that assessment? I will leave that to you.

As for fielding, Mench ranked 15th among leftfielders last year and Lee ranked 23rd, according to John Dewan’s Fielding Bible. Lee truly looks like a bad baserunner, however. The Hardball Times Annual gave him -2.8 baserunning runs and Mench received a positive 2.2.

Overall, there appears to be about a 10-run edge for Mench in fielding and baserunning (equal to one win) and prior to this year, you might have rated Lee and Mench relatively even in batting prowess. Add in the fact that Mench is younger and won’t be a free agent for two years, and you might actually believe that Mench really isn’t that “poor” a relation to Lee. In the meantime, watch his groundball rate.