Managing to Fill Out the Lineup

We've talked about the offensive and defensive moves a manager can make that will affect the outcome of runs, either in possibly 'creating or preventing' them. The last area that the manager affects directly is the lineup.

Game Four of the 2001 World Series demonstrated that one of Bob Brenly's major concerns was scoring first. Now, most of us have read in the past that "the team that scores first wins two-thirds of the ballgames." Right there in the 1998 STATS Scoreboard, there's an essay that reveals that from 1986-1997, the team that scored first won 66.4 percent of the contests.

What exactly does that mean? Does the "first run" actually carry any real weight, or is it just a function of the "better" team being "more likely" to score first? what about first runs scored late in a ballgame? Wouldn't those "first runs" carry more weight than those "first runs" in a slugfest?

Anyway, the way a lot of teams play, they seem to be willing to give up outs in order to score that all-important first run.

An essayist luckily took the question more steps. What about the second run, and the third, and the fourth? Who is most likely to win scoring THESE runs? The bottom line is this: teams over the last decade win more than 70% of the time when scoring the second run of the game, regardless of whether or not they scored the first run. Teams scoring the third run, fourth, and so on, also won over 70% of the time, up till the ninth run of a game.

So it appears that the advantage of scoring first is not as important as scoring often, and that runs scored in bunches are more likely to lead to a win than playing for one run (to the tune of about a 7 per cent chance per game.) Playing "little ball" may increase the chance of scoring one run, but it certainly severely decreases the chances you'll score multiple runs (see: "Managing to be Offensive"). Earl Weaver was right again; you win more games with the big inning than the small one.

Maybe the next time Bob Brenly gets to the World Series (if ever), he won't be so generous with his outs.

So how does this thinking affect lineups?

I've always been fascinated with the theories of lineups. The idea of alternating lefties and righties, the concept of batters offering "protection", speed at the top of the lineup with bat control in the number two hole, etc., etc., etc. Then I read a Bill James book about ten years ago that said it really didn't matter WHERE in the order anyone hit, the number of runs a team scored would be roughly the same over the course of a season. Now, on the one hand, I thought this was pretty cool information...but on the other hand, I really felt pretty bad about this. Could it be true? All of the hand-wringing about who'll protect who, and being top-heavy with lefties or righties, or not having a "good RBI man", or who should bat leadoff, and who should bat ninth?

Well, I admit that once I started thinking about the "protection" thing, that it made sense. After all, I'd watched Dale Murphy have good seasons with Bob Horner in the lineup, and without him. Then an essay in another

Scoreboard made it quite clear that "protection" in the lineup is pretty much a negligible factor in output, no matter how weird it sounds. (I wrote an article about this at the Hangout a couple of years ago called "How are You Fixed for Protection?" It's no longer in their archives because all of my articles have been removed. I'll have to check in my old computer and see if its still there.)

I have managed to put the "protection" argument out of my mind over the years, but this concept of "batting people wherever", I haven't been able to totally embrace. It just goes too much against the old-line thinking I've been privy to all these years...and in some ways, it's just not as much fun.

Still, if we're going to talk about lineups, we have to talk about it. :)

One thing we do know is that the higher a player hits in the order, the more times they will come to the plate (see: Cal Ripken in his last game.) On average, a team's leadoff slot will garner 750 plate appearances over the course of a season, while the number nine slot will come up only 580 times. The averages:

On the face of the matter, it would appear that a team might be best served by batting their leading hitter first, the second leading hitter second, etc, etc, just to maximize the number of at-bats the better hitters get over the course of a season. This might really seem weird on the surface, but it does make some sense.

Before anyone gets all worked up about it, after running thousands of simulations with both the "best" batting orders (such as the one just proposed, in descending order of OPS), the "typical" batting order, and the "worst" possible batting order, the differences in result are not earth-shattering. Usually the simulations come out within 30 runs of each other, no matter what the batting order. In a typical season on a typical team, that might mean three wins.

Three wins is nothing to sneeze at, of course. It's been said many times that the difference between a good manager and a bad manager is probably no more than five wins, so three is a lot. It's just that we often overestimate the value of moving someone up or down in the order, just we usually overvalue the offensive impact an individual player will have on his new ballclub. Since there's probably no more than a one or two win difference between hitting in the "traditional" lineup, and hitting in the "sabermetric" one, one can suppose it's not a subject worth getting all worked up about, although it certainly is interesting.

Occasionally, a team will "accidentally" prove the point through the sudden power splurge of a leadoff batter. In 1997, the Red Sox scored 851 runs with Nomar Garciaparra leading off with a .306/.342/.534, and the 1996 Orioles scored a whopping 949 runs behind Brady Anderson's .297/.396/.637. These lineups became "top heavy" with hitting not by design, but by achievement...but the results were there for everyone to see.

One place that "convention" might actually mean something in the 7-8-9 slots in the sabermetric lineup. Since these are the worst three hitters on the team, perhaps this should feature the "little ball" configuration of the fastest of the three in the number 7 slot so they can run and play little ball with the only two hitters in the lineup bad enough to play "little ball" with.

 What does this tell us, in the final analysis? The most influence a manager can have is not in determining whether to bat a player second or sixth, but rather in deciding whether to play him at all.