Lou Costello: What's on first?

Bud Abbott: What's on second.

Costello: I don't know.

Abbott: He's on third.

The Pirates' infield is almost that confusing. Just when it appeared the team had settled the question of "which Jose starts," choosing Bautista over Castillo, the man to be moved in that shuffle is hurt.

Freddy Sanchez, who would move from third to second base so Bautista can play the hot corner, may start the season on the disabled list because of a sprained ligament in his right knee. It appears Castillo will start at second for the time being.

Sanchez twisted his knee turning a double play at second March 6, so you can add that to the measure of how far Castillo's stock has fallen. The Pirates would prefer the defending league batting champion take on a more dangerous position than stick with the incumbent, and they can make a good case.

If Castillo wasn't the worst regular second baseman in baseball last year, he was close. Dropping him for Bautista should add a better bat, everything being relative. But even if you believe Sanchez can play second as well or better than Castillo (which I do), there's little chance Bautista will play third as well as Sanchez, who is terrific there.

Let's put the glove aside and start with the bat, though. Castillo topped Bautista in batting average last year, .253 to .235, which would work out to about 10 more hits over the course of a full season. On almost everything else with the bat, Bautista blew Castillo away.

The quick way to show that is with secondary average, an old Bill James stat that measures a hitter's contributions above batting average. Add the extra-base hits a batter gets to his walks and net stolen bases, then divide the total by times at bat. Bautista, in his first full season as a Pirate, had a secondary average of .295. Castillo came in at .195.

Bautista hit a home run every 25 at-bats, Castillo one every 37. Bautista walked every 10.2 plate appearances to Castillo's 17.5. Bautista saw 4.17 pitches per plate appearance compared to Castillo's 3.53. Bautista was also hit by 16 pitches to Castillo's five. Thus Bautista's on-base average was .335 to Castillo's .299, and Bautista topped him in slugging .420 to .382.

That said, Bautista is having a lousy spring and his .755 OPS (on-base plus slugging) last year was still 71 points below the average NL third baseman. He needs to be better for the Pirates to be good.

Now we come to the gloves. This is where the tradeoff may come. That's not because Castillo is great at second, but because Sanchez is tremendous at third, and the left side of the infield stays busy behind the Pirates' left-handed ground-ball pitchers.

Measuring defense is difficult. There's more consensus on global warming. But Castillo was last in fielding percentage in 2006 among the dozen NL second basemen who started at least 100 games. He was also last in Zone Rating, wherein STATS Inc. calculates the percentage of balls a player fields in his defensive zone.

Castillo still had a middling Range Factor, a tally of the number of putouts and assists per game, and was fourth in double plays with 106. But is that due to his skill or to the shortstop and third baseman feeding him the ball? That's where arguments begin.

Castillo's signature skill, turning the DP, deteriorated last year. He went from being baseball's second-best man on the DP pivot in 2005 to 14th in 2006, according to Baseball Info Solutions. He also dropped from a minus-8 on fielding ground balls to minus-14.

Baseball Info Solutions thinks Sanchez was a much better second baseman in less than 19 full games last year. Castillo made 15 fewer plays than the average second baseman given the same chances, while Sanchez was simply average.

Baseball Prospectus says Castillo's defense cost the Pirates roughly 16 runs every 100 games, but it didn't think much of Sanchez at second either. In his limited play there, Sanchez registered negative 15 runs every hundred games, just a hair better than Castillo.

The only unanimity is that Sanchez is a superb third baseman. Baseball Prospectus says Sanchez has been 20 runs better than average in roughly 150 career games at third; he has figured in 43 double plays in that short time. Baseball Info Solutions graded Sanchez 12 plays better than average last year and 14 better the year before. Bautista, meantime, received a minus-9 grade from Baseball Info in less than 31 full games at third last year. Baseball Prospectus had him at minus-1.

"Just from the numbers," said John Dewan of Baseball Info Solutions, "I would try Bautista at second because Sanchez is so good at third."

That's unlikely. Bautista played less than three games at second last year, and a handful more in the low minors in 2003. He has played only one game at second this spring. With Sanchez's injury and these other issues, the infield remains a work in progress. Anyone who thinks it's all set isn't paying attention.