Who Needs Rookies?
The rookies aren’t hitting, and this apparently has Jimah spooked. During spring training, new manager Jimy Williams declared rookie Morgan Ensberg his starting third baseman, but Ensberg stumbled out of the gate. Ensberg is batting .250 with just two extra-base hits in 40 at-bats, although he has half as many walks as hits, giving him a .348 OBP. Given that it’s just three weeks of performance, those numbers don’t mean much. Nonetheless, in the last 10 days, Williams has started career utility player Geoff Blum four times and Ensberg five times at third base.
Similarly, rookie Adam Everett, also named a starter in spring training, has struggled with a .189 batting average and .216 slugging percentage. Like Ensberg, Everett has drawn his share of walks, posting an OBP 129 points above his paltry batting average. Also like Ensberg, Everett has shared starting time recently, with Julio Lugo getting four starts at shortstop to Everett’s five. Everett has since been demoted to New Orleans, presumably for just a couple of weeks, so the Astros can carry 12 pitchers.
A lot of people expected Everett to hit like this anyway, but the only thing 37 at-bats proves is that baseball statistics fluctuate wildly in small sample sizes. If you disagree, consider picking up Robert Fick and his .404 batting average in 52 at-bats for your fantasy team.
There’s no way to know whether Williams might’ve juggled his line-up even if Ensberg and Everett were hitting. With Boston last season, Williams used 93 different line-ups in 118 games, which extrapolates to 128 different line-ups in a full season. Only Mike Hargrove in Baltimore (139), Tony Muser in Kansas City (130), and Mike Scioscia in Anaheim (130) exceeded this figure in the American League. Williams led American League managers with 140 different line-ups in 2000. Williams was in the middle of the pack with 111 different line-ups in 1999.
Line-up usage might not be a function of managers, but of personnel, however. At first base, second base, and in the outfield, the Astros have five players who are clearly better than their available replacements.
Thus, other than injuries and rest, there’s no reason to trot Brian Hunter or Orlando Merced out there. Since catchers can’t play as much as other position players, Brad Ausmus and Gregg Zaun will share time. If having Zaun as his personal catcher floats Dave Mlicki’s boat, more power to him.
For Ensberg and Everett to succeed, however, they’re going to need more than sporadic playing time. Ensberg in particular has shown in the minor leagues that he can be a fine offensive player. Yanking him in and out of the line-up might not be conducive to Ensberg proving that to the Astros. If Williams has a good reason for shuffling the deck, like gaining some kind of platoon advantage, that’s one thing.
But Blum, a switch hitter, is a career .266/.329/.495 vs. lefties and .287/.322/.351 vs. righties, while Ensberg, in the minors, batted .325 against lefties and .299 against rights, so their differentials aren’t large. If Williams is just making changes because that’s what Jimah does, or because he simply doesn’t trust rookies, then that’s not doing Ensberg or the Astros any good.
Intentional Walk-A-Thon
Astros pitchers have intentionally walked 14 batters in 162.3 innings, an average of 0.78 per game. By comparison, consider the intentional walk rates under Larry Dierker:
Year IBB IP Rate ----------------------- 2001 22 1454.7 0.14 2000 25 1437.7 0.16 1999 17 1458.7 0.10 1998 25 1471.3 0.15 1997 25 1459.0 0.15
Dierker was well known to be stingy with the free pass, but even Williams was less generous during his tenure with the Red Sox than he’s been with Houston so far:
Year IBB IP Rate ----------------------- 2001 29 1063.3 0.25 2000 40 1452.7 0.25 1999 25 1436.7 0.16 1998 33 1436.0 0.21 1997 52 1448.0 0.32
In 18 games, Williams has already ordered more than half as many intentional walks as Dierker called for in any of his five full seasons. The Astros are on a pace to walk 126 batters intentionally, which would surpass the record of 116 held by the 1974 Padres. The high rate this early in the season is almost certainly an aberration. Nonetheless, many fans are probably relieved to see the Astros using more “strategy” than under Dierker.
The intentional walk is of limited use in most strategic situations. Intentional walks have raised the OBP against Astros pitchers from .345 to .358. Those are runners with the potential to score runs. How great is that potential? It depends on the base-out situation.
The table below depicts the average number of runs scored after the occurrence of each possible base-out situation for the five seasons from 1994 to 1998.
123 --0-- --1-- --2-- --- 0.544 0.292 0.111 x-- 0.940 0.563 0.245 -x- 1.185 0.720 0.347 xx- 1.569 0.969 0.470 --x 1.437 0.992 0.393 x-x 1.839 1.193 0.539 -xx 2.044 1.405 0.616 xxx 2.349 1.618 0.804
The left column indicates which bases were occupied. The right three columns show the number of outs and the average number of runs. For example, after situations with bases loaded and no outs, an average of 2.349 runs were eventually scored. A team enters an inning with an average potential of 0.544 runs.
As the table makes clear, via the intentional walk a team never decreases the average number of runs expected to score. The number of additional runs expected by the issuance of an intentional walk in each possible situation are summarized below:
123 --0-- --1-- --2-- --- 0.395 0.272 0.134 x-- 0.630 0.405 0.225 -x- 0.384 0.249 0.124 xx- 0.780 0.650 0.334 --x 0.402 0.201 0.145 x-x 0.510 0.425 0.265 -xx 0.305 0.214 0.188
In other words, the least costly times to distribute the free pass are as follows:
with a runner on second and two outs (+0.124);
with no runners on and two outs (+0.134);
with a runner on third and two outs (+0.145);
with runners on second and third and two outs (+0.188);
with a runner on third and one out (+0.201); and
with runners on second and third and one out (+0.214).
It’s interesting that perhaps the classic intentional walk scenario, with runners on second and third and one out, is not the best bet mathematically. Despite setting up the double play, this situation also forces the pitcher to throw strikes, since a walk will force in a run. Not only must the pitcher be accurate and try not to get hit, he must either coax the double play or retire two batters.
What these tables ignore is the one-run situation. Sometimes the expected number of runs is irrelevant, because one run will end the game. To analyze this situation, a table showing the frequency with which at least one run was scored is more useful. Here is that data for the same period as above:
123 --0-- --1-- --2-- --- 0.295 0.174 0.074 x-- 0.447 0.289 0.136 -x- 0.638 0.419 0.230 xx- 0.651 0.433 0.239 --x 0.852 0.673 0.276 x-x 0.870 0.655 0.295 -xx 0.871 0.688 0.276 xxx 0.886 0.678 0.336
In other words, with the bases loaded and no outs, teams scored at least one run 88.6 percent of the time. In a one-run situation, for example, as the visiting team in a tie game in the ninth inning or later, there are two situations when, on average, intentionally walking the batter decreases the frequency of at least one run scoring. The percentage change for an intentional walk for all base-out situations is shown below:
123 --0-- --1-- --2-- --- 0.152 0.116 0.062 x-- 0.204 0.144 0.103 -x- 0.013 0.014 0.009 xx- 0.235 0.244 0.097 --x 0.017 -.018 0.019 x-x 0.016 0.023 0.041 -xx 0.015 -.010 0.060
The times to walk the batter intentionally are with a runner on third and one out, which reduces the scoring frequency by 1.8 percent, and with runners on second and third with one out, which reduces the scoring frequency by 1.0 percent. All other situations yield at least a tiny percentage gain, although with a runner on second and two outs, the increase is a mere 0.9 percent.
These tables are aggregate data and are thus the average of all batters in those situations. With a weaker batter, the expected runs and scoring frequency should decrease. With a stronger batter, conversely, they should increase. But what the tables reveal is that intentional walks on average should be confined to particular situations and are most appropriate in certain one-run situations.
“I know it’s early, but …”
The Astros have stumbled out of the blocks at 8-10, leaving many of their supporters apoplectic. How short are the memories of baseball fans. The 2001 Astros at 10-8 weren’t significantly better at this point, and that team failed to get permanently above .500 until its 34-33 mark on June 18.
The Cardinals, the biggest threat in the Central Division, haven’t exactly set the world on fire. After jumping out to 8-3, St. Louis has dropped seven of eight games, putting them just a half-game ahead of the Astros at 9-10. That record is exactly where the Cardinals stood at this point last year.
Ignore all the proclamations that begin, “I know it’s early, but …” The baseball season is a marathon, not a sprint. Making judgments based on the first three weeks of the season, indeed, just about any three weeks of the season, is foolhardy. There is nothing that says good teams will play well all the time.
Right now the Pirates at 12-5 sit atop the Central Division. Maybe Pokey Reese will bat over .300 all season. Maybe Mike Williams will maintain his pace and save 76 games. Smart money wouldn’t take these bets, though. I could be wrong, but I doubt the Pirates are going to win 114 games.