The White Sox took everyone by surprise last year in winning the World Series. I have already explained that they may have been luck, in particular in doing so well with runners on base. In the long-run hitters do about as well with runners on base as they do overall (see "Clutch Hitting Leaders, 1987-2001"). The same is probably true for pitchers since the number of runs they allow is very well predicted by the OBP (on-base percentage) and SLG (slugging percentage) they allow (see "Do Pitchers Give Up their Expected Number of Runs Based on OPS?"). So a team would be fortunate to do better with runners on base both when hitting and pitching. This was true for the 2005 White Sox. I look at how many more games they won as a result.
In 2005, the White Sox batters had an overall OBP of .322 and SLG of .425. But with runners on base (ROB), their OBP was .336 and their SLG was .443. With no runners on base (NONE), their OBP was .312 and their SLG was .412. The normal increase in OBP is about .024 when moving from NONE to ROB situations (see "General Clutch Data"). So the Sox were pretty normal there. But the normal increase in SLG is only about .011 while the Sox increased .031 or about 3 times normal.
On the pitching side, Sox pitchers allowed an overall OBP of .310 and an SLG of .387. In NONE situations they had and OBP of .312 and an SLG of .412. But in ROB situations, their OBP and SLG were .308 and .376, respectively. So although OBP and SLG normally go up in ROB situations, they both actually went down for Sox pitchers, with a huge drop in SLG.
When OBP and SLG are added together, its called OPS. I found the OPS differential (hitting OPS - pitching OPS) for each team in 2005.The table below shows how they ranked:
Notice that the White Sox were not much better than their opponents in NONE situations. But in ROB situations, the story is much different. The White Sox are number 1.
So how much did this huge improvement in performance help the White Sox? The following equation generated using linear regression analysis shows how winning percentage is related to OBP, SLG, OPPOBP and OPPSLG (where PIT mean's opponents' stats-data included all teams from 1989-2002)
(1) PCT = .493 + 2.01*OBP + .858*SLG - 2.06*PITOBP - .806*PITSLG
It predicts that the White Sox would have a winning percentage of .548 (using their overall hitting and pitching OBP & SLG listed earlier). That would be 88.82 wins for a 162 game season. The Sox actually won 99 games or 10.18 more than predicted. But how many games should they have won if we take into account their ROB performance?
Below is the regression equation in that case
(2) PCT = .490 + .873*NONEOBP + .334*NONESLG + 1.192*ROBOBP + .551*ROBSLG - 1.104*NONEPITOBP - .177*NONEPITSLG - 1.074*ROBPITOBP - .600*ROBPITSLG
Plugging all the data from above gives the Sox a winning percentage of .571 or 92.48 wins. That is 3.65 more wins than the 88.82 predicted by equation (1), which did not account for ROB performance. It still leaves the Sox 6.52 wins short of their actual total, but doing this eliminates about 36% of the 10.18 gap predicted by equation (1).
Equation (1) had a standard error 4.82 wins for a 162 game season. Since it under predicted the Sox by 10.18 wins, it means that the prediction was off by 2.11 standard errors. The standard error for equation (2) was 4.5. Since it under predicted the Sox by 6.52 wins, it means that the prediction was off by 1.44 standard errors. So the mis-prediction in terms of standard errors falls by about 68% (1.44/2.11 is about .68). I also tried using a team's walk percentage, hit percentage and extra-base percentage instead of OBP and SLG and got similar results.
Have the White Sox maintained their great ability in ROB situations this year? Somewhat. Their hitting OPS in NONE situations was .791 while their pitchers have allowed an OPS of .748. So that differential is .043. In ROB situations, their hitting and pitching OPS are .829 and .743, respectively, for a differential of .086. So the differential rises by .043 in ROB situations this year while it rose .094 last year.
Sources:
The ESPN website
STATS, INC