Removing the MLBPA from the equation (show me a union that allows job cuts in a thriving industry, and I'll show you a disbanded union), I think 22 is definitely possible.

Go to a four man rotation - that's one gone - get rid of that last man in the bullpen that you only use when the game is already decided - that's two - then get rid of that last position player on your bench who only pinch hits when the backup catcher is the alternative - that's three.

There is an unintended consequence, though, and that's the utter devastation of the Major League portion of the Rule 5 draft. A player would have to be flat-out, Major League ready before a team would even think about trying to acquire them via that route. I don't know if that's good or bad, to be honest, but it's definitely part of the deal in this scenario.

I say that rosters should be expanded if anything. Teams have been upping the number of pitchers used per game for years. It broke the five-per mark three or four years back if memory serves. I like the idea of Kieschnicks and McEwings finding spots, but I think that's more a function of a limited roster than a throwback to the pre-specialization days. I'm also reminded that the NL had a self-imposed 24-man roster for years supposedly to compensate for the AL's need to carry an extra bat to DH (whereas having the extra bat in the lineup obviated the need for an extra pinch-hitter and long reliever on the bench).

I would say keep it at 25 and add a five-man taxi squad for the emergency QB and kickers, oh and of course, to dump overpriced flops like Mo Vaughn.

Japan uses a 28-man roster, with 25 active players for any one game.

What normally happens is that the starting pitchers take a rest and then there are lots and lots of middle relievers available.

Mike elides to a good point. IMO baseball contracts should not be guaranteed. This would eliminate much of the need for a 25-man roster.

By the way, one could note that Triple-A has become an extension of the bench. Take a look at the average roster size per league since the 1950s:

Decade Lg Avg # Players % Increase
1950s AL 39.33
1950s NL 37.53
1960s AL 38.55 -2%
1960s NL 37.80 1%
1970s AL 37.53 -3%
1970s NL 37.21 -2%
1980s AL 38.43 2%
1980s NL 38.89 5%
1990s AL 42.86 12%
1990s NL 42.31 9%
2000s AL 45.07 5%
2000s NL 44.73 6%

I think AAA is used more because sometime in the '90s they changed the option process from counting total options to option years, so that you can bring a player up and down many times during the same year.

I thought it was earlier than the 1990s when the options rule was changed.

Didn't Dave LaRoche go up and down with the Yankees and Columbus several times in 1981?

Bob,

I'm pretty sure you're right here. I remember players being yanked (excuse the pun) up and down like yo-yos starting in the early Eighties. It probably happens more now, but it was not unheard of then.

A couple of days ago during a Yanks/Blue Jays broadcast, Jim Kaat suggested that the rosters should be expanded to 28, and my head nearly exploded.

Bigger rosters? That's just what we need--an opportunity for Tony LaRussa to trot out even more relief pitchers during the course of a game.

The roster size is just fine. The problem is more with composition, and that's an organization-specific problem in most cases.

I'm fine with 25, but I think that there should be a cap at 10 pitchers to prevent the kind of havoc (if boredom = havoc) that the LaRussas of the world wreak when they use four pitchers in an inning. Maybe we'd get more Kieschnicks, McCartys and Ankiels on the roster, which is much more interesting than watching a team's third LOOGY and 20th best overall pitcher (including minor-leaguers) flail about. And maybe #25 is an emergency catcher only available if the second catcher leaves due to injury.

Oh, and I think it was Dave LaPoint, not Dave LaRoche, to whom Bob T. is referring.

Murray,

Thanks for comparing me to "Kitty" Kaat.

By the way, LaRussa's not the only one. The number of pitchers per game has been on a steady increase since the 1980s:

Decade LG P per G % Increase
1950s AL 2.23
1950s NL 2.36
1960s AL 2.59 16%
1960s NL 2.50 6%
1970s AL 2.34 -10%
1970s NL 2.60 4%
1980s AL 2.58 10%
1980s NL 2.88 11%
1990s AL 3.27 27%
1990s NL 3.39 18%
2000s AL 3.52 8%
2000s NL 3.75 11%

And both leagues reached all-time highs last season:

2004 AL 3.63
2004 NL 3.88

I don't like excessive pitching changes either, but I don't see how to stem the tide.

I'm OK with it the way it is now. It's not like basketball, where there are teams who actually carry less than the roster limit. It's flexible, but not so much so that a manager can do anything he wants day-in, day-out.

Based on my age, I think the options rule was changed sometime in the 90s, even if it was the early 90s ... I don't think I'd remember the change happening if it were any earlier. But I'm just guessing. Perhaps if we had year-to-year as opposed data on roster size as opposed to decade-by-decade we could spot when the change occured. I did a(n admittingly very) quick Google search and couldn't find the year.

If the goal is to reduce the number of pitchers per game, then two solutions would be: (1) limit the number of pitchers per inning and/or (2) require that a pitcher record an out before he is allowed to be removed from the game.

The above proposals would also serve to reduce the time of the game, another trend that has been growing over time.