Is baseball a team game?

Is baseball a team sport? Obviously, it's a sport played between two teams, but is it a team sport? That is, how much of baseball is played by the team as a whole, and how much of it is the sum of a series of individual contests? Baseball, maybe more so than any other sport, falls somewhere in between, which makes the statistical analysis of it so complicated but possible.

Individual sports are easy to identify. If there is no team playing, then it is impossible to be a team sport. Golf and tennis are America's favorite individual sports. When an individual wins an event in one of these sports, there is no question as to who is responsible for the game's outcome.

Team sports, on the other hand, present special problems. Let's say that a true team sport is one where each side is composed of multiple individuals, and that the performance of each individual is heavily influenced by the performance of their team mates. American football provides good examples of this. Many a promising young quarterback has come undone when their line fails to protect him. The best of cornerbacks cannot cover even average recievers forever if the line does not provide a decent rush. The only individual who disproportionally affects the outcome of a football game is the quarterback.Even the best of running backs suffers when the linemen do not provide space for the back to run.

When a running back gains 1500 yards in a season, it's impossible to give all the credit to just the back, even if we do. A great deal of the credit has to go to the line. But how much? And to which lineman? These questions may be impossible to quantify, at least with the data we have presently at hand. I think it's fair to say that a game where it is impossible to quantify the true contribution by any one individual to be a true team game.

Similar arguments can be made to greater or lesser degrees by basketball and ice hockey. There are sports which are team in name only such as Ryder Cup golf. The game is contested by two teams, but is really a series of individual match play events. The doubles events, like fourball and best ball, involve a modicum of team play, but the majority of the contest, the singles matches, have (or at least should not have) any bearing on one another.

Baseball, on the other hand, sits squarely in the middle. While championships are awarded strictly on team play, no other sport offers so many individual statistics, both traditional and modern, to evaluate the performance of individuals on each team. And for good reason: during any one at bat, two figures, the pitcher and the batter dominate the outcome of the at bat. While there are runners on the bases and fielders in the field, the actions of these two individuals overwhelmingly account for what happens and the success of each atbat.

While traditional individual baseball statistics have served us well in identifying good players, they are less useful for truly quantifying a players contributions to the team's overall success. Some traditional stats are particularly deceiving in this regard; the RBI for batters and pitching wins (and most recently, saves) are the most egregious. There was a time when Randy Johnson would go a month without giving up a run, yet did not earn a single win. The number of RBI a player earns is directly proportional to the number of runners on base when the player gets up to bat. These statistics more closely measure team performance rather than that of any individual.

Modern baseball analysis attempts find measures genuinely measure each individuals contribution to a team's success. To do this, it looks for measures that can 1) be applied to both a team or to an individual, and 2) explains the number of runs a team actually scores. It can be shown, of all published baseball stats, that the combination of on-base average and slugging average bests explains the number of runs a team scores. Intuititively, this makes sense: an efficient way to score runs is to get one or more runners on base, then knock them in with a double or a home run. Note that it requires the combination of on-base average and slugging percentage. Pure slugging results in a batch of one run homers, while pure on base play leads to a glut of runners left in scoring position. This sounds a lot like an endorsement of the old three run homer strategy of the old Baltimore Orioles, at the expense of small ball. There's no doubt that the sacrifice bunt and stolen base have their place in unique situations, but the numbers show that offensive production is explained very well by on base average and slugging percentage, and very poorly bunts and running.

Now the critical assumption in modern baseball analysis is what is good for the team is good for the individual. Since good teams are characterized by high on base and slugging averages, then good players are also the ones with these high averages. So by these standards the best hitters are those who combine good on base and slugging average. Those who do just one well are not as highly valued as those who do both well. But couldn't a pair of players, one who got a single or walk every time up, who was followed by a player who homered every time, be just as effective two players who both homered every time up? Can't a team be composed of a mix of players, some of which specialize in getting on base, followed by others who specialize in knocking runs, whose team production is greater than the sum of its parts? If so, we could conclude that baseball is a true team sport. After all, isn't this the way that lineups are formed, where the best on base batters (sometimes the best base stealers) get up first, while the best sluggers bat clean up and later? The revered three spot in the order is reserved for the batter who best combines the virtues of high on base and slugging average.

The evidence suggest the opposite, however. Most tellingly, the number of runs a team actually scores is closely related to the sum of individual contributions as predicted by the player on-base and slugging averages; there seems to be very little value-added for the "team effect". Additionally, there are few players who can be genuinely characterized as pure sluggers or pure on base players. The chart below compares the on base and slugging percentage for all big leaguers with 95 or more atbats as of July 12. While the correlation is not perfect, in general, players with high on base percentages also have high slugging averages.

It's impossible to look at the chart without commenting on the obvious. In a galaxy of points, Barry Bonds stands out by himself as the supernova in the upper left hand corner. Love him or hate him, natural or artificial, he is in an entirely different league. It's also interesting for as anomolous as he is, he does not stand out as either a slugger or as an on base guy. He is extremely typical in his ratio of on base to slugging percentage; his numbers are both double the league average.

So what does this mean with regard to team construction? At the risk offending purists, it suggests that the best team is the best collection of individuals, and that mixing on base and slugging specialists is suboptimal. Lineup construction should make use of the resources at hand. No lineup will ever consist of 9 clones, all equally capable. So to the extent that you have on base and slugging specialists you may as well order them to take advantage of their strengths. This should not be confused, however, as the best way to build a team. The more good individuals you can load up, the more runs you will score. I dare say this strategy has worked out very handsomely for the team whose payroll is as anomolous as Barry Bond's stats.

An interesting side question to all of this is: What is the most team oriented game? I can't answer that question quantitively or even accurately, but my hunch is that football, soccer, rugby, and ice hockey are good candidates. Basketball is an interesting case, where 5 players on each team are all active at any one time, yet it's the actions of 2 or 3 players on each side that most affect the game. No NBA team wins on the basis of just one star player, yet its rare when all 5 participate more or less equally in the result. I may be biased, but you have to go back to the 69-73 Knicks before you can find a basketball team where five players contributed so evenly to their championship.