So it's the first week of the season and you're already itching to ditch Pirates manager Lloyd McClendon? You hated the experiment of batting Tike Redman third? Just generally unconvinced about McClendon's decisions?

	Peter Diana, Post-Gazette Pirates manager Lloyd McClendon doesn't need to worry about being replaced by a computer. Click photo for larger image.

When there's a swell of discontent with a manager or coach, the voice of reason needs to ask who would come in as a replacement and do a better job.

Here's a thought to download:

What if a well-programmed computer, or a robot with a computer brain, took the place of a major-league manager for a season?

We could call such a contraption a cyber manager. Cyb for short.

Think about it.

Baseball in particular is managed on matchups, steeped in statistics, preoccupied with playing the percentages. Teams already rely more and more on computers to keep track of such things.

If you loaded all sorts of data into Cyb's memory bank and kept it updated to the minute, he could compute the best possible lineup every day.

He could take into account not only a hitter's past against a pitcher, but also break it down to include variables such as home or road, day or night, grass or artificial turf.

Taking it a big step further, Cyb could then manage the game.

For every pitch, he could come up with the best odds on when to bunt and when to swing away, when to change pitchers and who to bring in, where to position the fielders and every other aspect a programmer could imagine.

People play chess against computers. Why not a baseball team against a circuit-head?

Of course, that would take the fun out of it for those fans who love to bellyache about the skipper. It would be difficult to second-guess Cyb knowing that he had all the facts and figures and cold, hard logic at his disposal.

McClendon can relax. Two local computer whizzes with connections to Carnegie Mellon don't see Cyb in a dugout soon.

"It's a very interesting idea," said Raul Valdes-Perez, a former professor at CMU whose specialties are artificial intelligence and data mining.

He's now CEO of Vivisimo, a Squirrel Hill company that operates www.clusty.com, an Internet search engine.

"But to take the idea to the extreme, I think it wouldn't be viable. Baseball is a very human thing. Emotions play a strong part in the game -- pride, loyalty, team spirit. It would be demoralizing. And there are intangibles. You might say that all those intangibles are baloney, but the number of intangibles is so big."

Handling a player who's in a slump, for example.

At CMU, Valdes-Perez developed a program that, given any player in history and a year, would come up with a grammatical sentence about that player.

For instance, in 1946, Joe DiMaggio had the lowest strikeout rate -- 5 percent -- of the 402 post-war Yankees outfielders.

He couldn't market that program. He doubts Cyb is scouting the marketplace.

"I think sports teams will use data more and more to guide them, but designing a system to really take over is just hard to do," Valdes-Perez said.

If Cyb would come along, he's much more likely to tweak conventional wisdom than to redesign managing, said Michael Trick, a Bosch professor in CMU's Tepper School of Business who along with his colleagues this year produced the first computer-generated master schedule for Major League Baseball.

Trick said computer models exist for certain sports scenarios, but they aren't always followed.

"If you look at hockey, every model says coaches do not pull goalies early enough. Most models say you should do it with 3 to 4 minutes to go," he said. "In football, the model says coaches punt too often on fourth down.

"Either our models are invalid, or the coaches are incorrect."

Regardless, Trick said, crunching the numbers -- whether by computer or human -- doesn't necessarily translate into success.

Take McClendon's choice to bat Redman third in the opener Monday, a 9-2 Pirates loss to Milwaukee.

"If that was the wrong decision, then we were just as likely to lose, 8-3," Trick said. "Then, the manager's decision is moot."

In fact, he thinks that's the case most of the time.

"Most of the models say managers don't really make a difference," Trick said. "All you do is change the odds of winning."

And not by all that much.

Trick estimates that a baseball manager's decisions change a loss into a win perhaps as little as 2 percent of the time.

Over 162 games, he figures as an aggregate a team might get as few as three and probably no more than 10 of its victories based on the manager's moves.

Three to 10 extra wins a season could mean a lot to the contenders, but not a whole heck of a lot to struggling franchises like the one at PNC Park.

So for most teams, it's all about the players.

Gee, who would have computed that?