Notes. No News, No Quotes. Just Notes.
Expecting Wins
The Mets beat the Cubs in dramatic fashion Saturday afternoon, with a couple of home run blasts by two rookies. First, Victor Diaz hit a three-run home run in the bottom of the ninth, with two outs and two strikes, to tie the game. Then Craig Brazell won the game with a solo shot in the 11th. Of course, this was a welcome sight for Mets’ fans, who have nothing but the future to look forward to — as in two or three years from now. For Cubs’ fans, this was agony.
How unlikely was this turnaround? Well, from 1979 through 1990, there were 137 games in which the home team was at bat, behind three runs, with two outs and runners on first and third. And the home team only won seven times. In 95% of the cases, the visiting team won. Billy Goat curse, indeed.
Here’s another example: the Dodgers and Giants played a nail-biter Friday night. The Dodgers were winning 3-2 entering the bottom of the ninth in San Francisco. At that stage their chances of a win were pretty good (80%).
After retiring the first two batters, their chances increased to 95%. Incredibly and unfathomably, with Barry Bonds on deck and J.T. Snow in the hole, Gagne proceeded to throw four straight balls to Pedro Feliz, decreasing the Dodgers’ chances of a win to 90%. Of course, Mr. Bonds was walked and, less obviously, Snow eked out a free pass too. Bases loaded, two out, and the Dodgers chances of winning were now 72%.
Luckily for Dodger fans, Torrealba lined out to left and the win was preserved, and the Dodgers’ chances of winning turned to 100%. This sort of thinking is called “Win Expectancy,” and I’ve found it a fascinating way to watch a ballgame. Once you get a feel the tables and the odds of each base/out situation, you start to think about the game differently.
You can take the concept further and assign win expectancy values to players, based on what they do during the play. For instance, Victor Diaz can claim 45% of a win, because he increased the Mets’ chances of winning from 5% to 50%. Craig Brazell gets about 37% of a win for his game-winning shot. Jay Bennett has been doing this sort of analysis for awhile. His system is called Player Game Percentage (PGP), and you might find his analysis of the 2003 World Series interesting.
Pitching Fits
By the way, the Cubs’ pitching has been red-hot in September. Their September ERA is 3.09 through Saturday’s games, and their :FIP: is an even better 2.85. That is the best month of pitching in the majors this year.
Although Jason Schmidt’s late summer decline has been well documented, it’s the Giants’ pitching that has kept them in the playoff race. Here is their FIP by month:
April May June July August Sept
4.76 4.34 4.53 4.78 3.81 3.75
Over the last twenty games, Schmidt has actually been their worst starter — Tomko, Reuter and Lowry have been carrying the load. Tomko, in particular, has been lights out.
Here’s another team that has seen its pitching improve throughout most of the year.
April May June July August Sept
4.21 4.35 4.12 3.87 3.77 3.55
That would be the Minnesota Twins, with even Kyle Lohse starting to contribute. As I’ve already mentioned, Johan Santana deserves serious consideration for the MVP; it will be interesting to see if the voters give it.
Is the Yankees’ Pythagorean Variance historic? Huh?
This has been pointed out before, but the Yankees do not have the best run differential (Runs Scored minus Runs Allowed) in the league. The Red Sox have them roundly trounced in Run Differential. Unfortunately for Sox fans, real victories are all that matter.
But one little statistical quirk is coming out of this. The Yankees are beating their predicted victory total (based on Run Differential, through a calculation called the Pythagorean formula) by ten games. That’s actually quite a lot, and it’s been the source of some interest throughout the year.
Over at Baseball Think Factory, someone (I forget who — sorry) asked if the Yankees have the largest Pythagorean Variance in history among division winners. The answer is no, but it’s close. Here’s a list of the top Pythagorean Variances among all teams who qualified for postseason play since 1900:
Team Year Wins Pyth Wins Diff Reds 1970 102 91 11 Reds 1961 93 83 10 Giants 1997 90 80 10 Athletics 1931 107 98 9 Athletics 1930 102 94 8 Twins 2002 94 86 8
Now, the Yankees could still finish eleven, even twelve games, over their Pythagorean projection, but it doesn’t seem likely. They’d probably have to lose another 22-0 game to do so.
The real key to the Yankees Pythagorean variance is their 46-23 record in games won by two runs or less. They’ve won more close games than any major league team this year. Now, many fans believe that bullpens are the key to positive Pythagorean variances, and the Yankees obviously have an excellent pen.
The historic oddity on this list is the two-year reign of the Philadelphia Athletics. The 1930-1931 A’s featured Lefty Grove out of the bullpen (in addition to starting, of course). On the other hand, over half their games were complete games, including 49 by Grove himself.
Career Home Runs and Win Shares — never seen together at the same time…
Barry Bonds is now in third place in both career Win Shares and career home runs. You know what’s eerie? The all-time scales for the two stats are eerily similar. Here is a list of the top twenty career Win Share leaders, along with the top twenty career home run leaders.
Win Share Leaders Home Run Leaders Babe Ruth 756 Aaron 755 Ty Cobb 722 Ruth 714 Barry Bonds 662 Bonds 702 Honus Wagner 655 Mays 660 Hank Aaron 643 Robinson 586 Willie Mays 642 McGwire 583 Cy Young 634 Killebrew 573 Tris Speaker 630 Jackson 563 Stan Musial 604 Schmidt 548 Eddie Collins 574 Sosa 539 Mickey Mantle 565 Mantle 536 Walter Johnson 560 Foxx 534 Ted Williams 555 Palmeiro 528 Pete Rose 547 McCovey 521 Rickey Henderson 535 Williams 521 Mel Ott 528 Banks 512 Frank Robinson 519 Mathews 512 Joe Morgan 512 Ott 511 Roger Hornsby 502 Murray 504 Nap Lajoie 496 Gehrig 493
See what I mean? The names are different, but the milestones (500, 600 and 700) are the same. The only difference is that Win Shares follow a smoother, more normal distribution.
That wacky Bill James. Think he did it on purpose????
References & Resources
The Win Expectancy data comes from Phil Birnbaum’s site (and many thanks to Jon Daly and Tangotiger for the guidance). The records of historical pennant winners courtesy of Sean Lahman’s database. All of our great in-season data comes from our friends at Baseball Info Solutions.