Do successful teams use fewer lineups? (And should La Russa?)
Brian Walton wrote an interesting post over at The Cardinal Nation about Tony La Russa’s lineup combinations for the upcoming season. Walton focused on how the defensive versatility of the players in this year’s spring-training camp affords such combinations. We’d like to offer another angle: Are many different lineups correlated to success? That is, is it necessarily a good thing if TLR trots out 100+ different looks again this season?
To ascertain some idea of the correlation, let’s look at the Cardinals since La Russa took the helm in 1996. To get a truer view of the team’s success, we’ll use Pythagorean Wins instead of actual wins. And, thanks to Baseball-Reference.com, we’ll use the number of different lineups, excluding pitchers (pitchers would add some noise we don’t want). To give you some perspective, the team won 86 games in 2008 (Pythagorean and actual) and used 106 different lineups; in 2001, the team won 94 and used 87 lineups. So as the number of different lineups increases, one might expect to see the P-Wins go down:
Without getting overly technical, this (the R2 of.5159) is not a super-tight fit, but it’s a fairly close relationship (anything over 0.5, generally speaking). Of course, the correlation between using fewer players and winning makes sense, and we need to be careful not to attribute causality: A team who suffers injuries to key players, relies on many inexperienced players and is generally tinkering with its lineup probably isn’t going to win as many games as a team that can go with a consistent group of reliable players. For La Russa and the Cardinals in particular, that’s been the story for the most part since 1996: The major exceptions were the 2000 club, which trotted out 113 different lineups and still managed 91 P-Wins and the 1996 group, who garnered only 86 P-Wins despite a relatively stable lineup (80 variations). The stellar clubs of 2004-2005 also won more games (100 and 98, respectively) than one might’ve expected them to based on their number of lineups (98 and 93, respectively).
Is this a condition endemic to only the Cardinals, or does the use of fewer lineups correspond to wins for the rest of the league, too? We took data for all NL teams from 2006-2008, and here’s what it looks like:
Perhaps unexpectedly, the correlation (R2=.3117) isn’t nearly as strong for the NL over the last three years. It could be that it’s a manager effect; Bobby Cox, for instance, doesn’t play around much in anything, his lineups included, no matter where he is in the standings. It might also be self-determining, in that some bad teams don’t improve precisely because they don’t change their lineups (see the 2007 Nationals, who used only 75 different lineups and got only 70 P-Wins), whereas some good teams might get that way through constant adjustments.
So then let’s look at La Russa. Perhaps he is simply more apt to respond to (or cause) adverse conditions by changing his lineups. For his career (we’ve excluded the 1986 season, when he managed half the season each for the White Sox and A’s):
That is odd, indeed, with almost no correlation whatsoever between the number of different lineups and wins. That is largely due to La Russa’s tenure in the American League, where, if anything, the more lineups he used, the more wins he had:
With those dominant Oakland teams, La Russa won pennants by using tons of combinations: The A’s had a staggering 128 different looks in 1990 and produced 99 P-Wins (and that was on the heels of the ‘89 team, with 112 lineups and 97 P-Wins).
What does all of this mean? Honestly, we’re at a loss. It could be that managing in the senior circuit has proved more of a challenge for La Russa when dealing with unexpected personnel changes. It could be a reflection of the talent and "bad luck" that the Cardinals have had over the years. And, of course, it may be simply be too small a sample to reliably tell us anything about how lineups correlate to wins. But if recent La Russa history is any indication, we’re still hoping that, for 2009, less — as in lineups — means more — as in wins.
February 19th, 2009 at 11:10 pm
The theory at one time was that TLR’s frequent lineup shifts, if nothing else, kept players healthy and rested and ready for August and September, which seemed to be born out by some of his teams roaring through the second half.
Unfortunately, the late swoons of the last three years tend to put a little doubt on that hypothesis.