The Pirates entered last night's game clinging to the division's best record since the All-Star break (28-26), but they had been only 21-23 against teams not named the Cubs.

Better teams await, and the Pirates need a strong finish from their starting pitchers if they are to play spoiler in the final three weeks and offer hope for next season. You know about the lack of a solid fifth starter, but the corollary is the need for one of these young pitchers to grow into an ace.

How likely is that?

Pirates fans learned from Kris Benson, Oliver Perez and Kip Wells that young pitchers can easily flounder after early success. This season, Paul Maholm and Zach Duke have fallen after phenomenal starts in 2005. They're doing OK, but how realistic is it to think Maholm, Duke, Tom Gorzelanny or Ian Snell will be a star?

Ron McClure of Murrysville was wondering the same thing, and so he looked at 153 pitchers whose careers began in 1969 or later, and had won at least 100 games by 2004.

McClure found that only 18 percent of these pitchers had what he calls "an erratic or underwhelming" early career.

"I was surprised at how many guys had really strong starts to their careers," he said. "Quite a few were better during their first three or four seasons than they were the rest of their career."

That doesn't mean a shaky early career is a death knell. It doesn't even mean that we can say any part of this quartet is off to a bad start; they haven't pitched enough to have established a career track. But it does mean a shaky start is very much the exception for long careers as a starter, not the rule.

McClure's criteria for an erratic or underwhelming start are these: at least 200 innings across the first three or more seasons; at least two seasons where the pitcher's "earned run average plus" was below his career number; an overall ERA+ in that time no more than 85 percent of his career number; more losses than wins.

A definition is required, so stick with me here. ERA+ is a number used by baseball-reference.com to compare a pitcher's ERA to the league ERA, making adjustments for the ballpark. Anything above 100 is above average. For instance, Duke's 1.81 ERA in about 85 innings last season gave him an ERA+ of 236. Perez plummeted from an ERA+ of 139 in 2004 to 73 in 2005.

Only 28 of the 153 pitchers in McClure's study reached 100 career wins after an erratic or underwhelming start. Among them were Tom Glavine, Jerry Reuss, Jamie Moyer, Dave Stewart, Al Leiter, Darryl Kile, Denny Neagle, Mike Scott, Jose Rijo, Pete Harnisch and Bob Walk.

I see only one potential Hall of Famer there, and not all were aces, but all made an All-Star team at least once. Let's look at three.

Glavine had four seasons under his belt through age 24, going 33-41 with an ERA+ of 89 in 646 innings. In his fifth season, Glavine went 20-11 with an ERA+ of 153, winning his first Cy Young Award.

Moyer took longer to get going. He was 34-54 with an ERA+ of 86 through his first 700 innings at age 28. In years after, through 2005, Moyer went 171-98 with an ERA+ of 112.

Walk was 33-32 as a starter and reliever through age 29. His ERA+ was 86 through those first seven seasons. From age 30 on, he was 72-49 with an ERA+ of 94.

There is not enough data on the Pirates' pitchers to know what's to come. None is older than 24, and all four should finish with an ERA+ above what Glavine, Moyer or Walk managed in their first couple of seasons. What that projects is another question.

I don't have up-to-date career ERA+ numbers, but the league ERA this season is 4.49, up from 4.22 last season. Compare that to the career numbers of the four starters through Sunday.

That's not too bad, though none is setting the world afire. Gorzelanny (3.74) alone is below the league ERA this season, albeit in just eight starts. The others are worse than league average, though reasonably close (Duke, 4.78; Snell, 4.88; Maholm, 4.90).

There may be better criteria, but I suspect McClure's conclusion about strong starts foretelling the vast majority of good careers is true. The Pirates need at least one starter to break big in 2007 and have to hope none stray too far in the wrong direction. Guys with good careers generally don't tarry as long as Glavine, Moyer or Walk.

"If a guy doesn't do reasonably well [fairly close to league average] in his first two full seasons, or two of his first three, don't expect him to be an impact pitcher," McClure said. "It may happen, but it ain't likely."