Does changing players change a team?

In a recent interview, Lance Berkman told one of the Astros beat reporters that this year’s team would be composed largely of players new to the team and he would have to see how this affected its chemistry, and by inference, its ability to win. This comment was met with:

1. Unhappiness by many Astros fans who were delighted to see new bats replacing Craig Biggio, Jason Lane, Adam Everett, Chris Burke and Luke Scott and delighted to see an almost entirely new bullpen and

2. The usual amusement of fans who believe “team chemistry” to be a antiquated term that attempts to define an amorphous feeling that can neither be explicitly quantified nor accurately defined.

If the term “team chemistry” were bandied about by only the media, then I would consider it merely a media invention. But best I can tell, its use by the ballplayers themselves predated its use by the media. Therefore, I’m not surprised that Berkman, the new de facto “team leader” would be concerned about this. A baseball team is composed of 25 men who spend the better part of seven or eight months together. Some groups of men work better together than other groups. Groups can divide into rival cliques with rival leaders, all vying for more power. That can significantly interfere with working relationships.

The Astros’ 2007 season marked only the team’s second losing one since 1991, and as the farm is virtually empty, the owner and GM had the choice of entering the 2008 season with essentially the same players and hoping for improved performance from everyone, or removing much of that team’s roster and replacing it with new players. They chose the latter.

Of the 25 players who were Astros on Opening Day 2007, only seven or eight will be Astros on Opening Day 2008. Of the players who were on the Astros’ 25-man roster as of Aug. 30, only 10 remain, plus two who were with the Astros organization at that time. This means that roughly half the team will be completely new to the organization.

I was wondering what happened in the following year to teams who replaced half of one year’s roster. Now I know very well that if you replace five Mario Mendozas with five Albert Pujols types and half of your pitching staff with aces, the club will improve. However, that never happens, not even to the Yankees or Red Sox. I also know that there are different umpires, different strike zones, different baseballs, different opposing pitchers, different opposing hitters and nothing stays exactly the same from year to year.

I looked at teams since 1994, and the percentage of plate appearances in a given year that were contributed by players who were on the team in the previous year. In 2008, The Astros will receive fewer than 48 percent of PA from players who were on the team in 2007. I compared them with the following teams from 1995 to 2006 who received fewer than 48 percent of PA from players who were on the team the previous year. I also looked at their before-and-after winning percentages, as well as noting whether, like the Astros, they changed managers from the beginning of one year to the next:

Year    Team    Turnover     Previous W%   W%            Manager change?
1995    TEX     45%          0.456         0.514         Yes
1997    SF      45%          0.420         0.556         No
1999    PIT     46%          0.486         0.424         No
2000    CHC     39%          0.414         0.401         Yes
2000    COL     47%          0.444         0.506         Yes
2002    NYM     48%          0.506         0.466         No
2002    SD      46%          0.488         0.407         No
2003    CHC     47%          0.414         0.543         Yes
2003    COL     45%          0.451         0.457         Yes
2003    MIL     43%          0.346         0.420         Yes
2003    SD      43%          0.407         0.395         No
2003    SF      44%          0.590         0.621         Yes
2003    TAM     33%          0.342         0.389         Yes
2004    COL     45%          0.457         0.420         No
2004    NYM     41%          0.410         0.438         No
2005    LAD     41%          0.574         0.438         No
2005    ARI     42%          0.315         0.475         Yes
2006    FLA     35%          0.512         0.481         Yes
2006    LAD     36%          0.438         0.543         Yes
2006    SD      42%          0.506         0.543         No

Ah, you all noticed that I have listed exactly 20 teams and only two are in the American League. In fact, 70 percent of the NL teams are from the West. 11 teams changed managers, and of those that did, seven had better records the following year. Eleven of the 20 teams had improved records the year after rolling over more than half the players.

I didn’t know if the acceleration of the rollover rate was a recent phenomenon, so I examined the rollover rate since division play began in 1969. Well, until 1995, only the following 18 teams, 10 of which are AL teams, rolled over fewer than 48 percent of their players from one year to the next:

Year    Team    Turnover     Previous W%   W%            Manager change?
1971    CHW     47%          0.388         0.565         No
1972    CLE     44%          0.370         0.462         Yes
1973    CLE     43%          0.462         0.438         No
1975    SF      48%          0.444         0.497         Yes
1977    CHC     46%          0.463         0.500         Yes
1978    OAK     20%          0.391         0.426         Yes
1981    CHW     32%          0.438         0.509         No
1981    CHC     41%          0.395         0.369         Yes
1981    SD      47%          0.471         0.353         Yes
1982    CHC     44%          0.369         0.451         Yes
1982    TEX     44%          0.543         0.395         Yes
1983    SEA     41%          0.469         0.370         No
1985    MON     45%          0.485         0.522         Yes
1989    CLE     44%          0.481         0.451         No
1990    CLE     44%          0.451         0.474         Yes
1992    KC      42%          0.506         0.444         Yes
1992    NYM     44%          0.478         0.444         Yes
1993    SP      48%          0.506         0.377         Yes

Of these 18 teams, 13 changed managers, and of those, eight increased their winning percentage. I’m not surprised that slightly more than half of manager changes resulted in increased wins—that happens more often than not regardless of the team’s record the previous year. Overall, nine of 18 teams improved their winning percentage the year after rolling over more than half of the team.

So, what can I conclude from these numbers? Well, if a team changes managers, it is more likely than not likely to improve its winning percentage, but a team that rolls over more than half its players has approximately a 50-50 chance of bettering its record, regardless of manager change.

So, based on these numbers, can I reasonably expect the Astros to improve their record this year and, by the way, exactly what does all this have to do with chemistry? If I had to guess, I’d say it’s baseball, so youneverknow.

As for the chemistry, it’s simple: If the team improves, it had better chemistry. If the team does worse, the chemistry was worse. At least that is how the team leader can explain success or failure to the media, who will accept that explanation happily, even if I won’t. But I still have no more idea than when I started if or how chemistry affects pitching, hitting or fielding because there is a factor you can’t find a formula for: the mind of the player. As a wise woman once told me, people don’t believe what they see, they see what they believe and the mind is a very powerful thing.