My first attempt at sabermetrics

1960-Program-320As of this writing, the Kansas City Royals are currently tied with the San Francisco Giants in the World Series at two games a piece. That’s a wonderful and miraculous thing for me to be writing in the first place, given the Royals’ postseason drought since 1985. After a painful game last night, however, I’ve been looking for signs of hope as we head into the home stretch. This led me to commit my first act of sabermetrics.

I’ve been intrigued by sabermetrics (the art and science of baseball statistics) ever since first reading about it in the essays of Stephen Jay Gould. I can’t claim anything like expertise in it — there are plenty of baseball stats used by folks like my friend Ben that I can’t even explain — but I always liked the idea that this great big data set was there for anyone to use and play with.

The Royals’ present dilemma gave me the perfect chance. Today’s Royals are a defensive team known for impossible catches and even more impossible-to-beat pitching. They do well holding on to a one-run lead. So it’s not surprise that the Royals edged out the giants 3-2 in Game 3. Our margin in Game 2 was more respectable — we won 7-2.

But in the two Giants victories — one in KC and one in San Francisco — we got clobbered. The first game was 7-1. Last night’s loss truly stung, with the Giants beating us 11-4. Each of these games left Royals fans wondering whether we truly deserved to be in the Series at all.

But runs don’t win the series — games do. Hoping to take away the pain from last night’s loss, I spent a few hours today running an analysis of the number of runs versus final victory in all of the World Series matchups since 1903. I hoped to answer the question, “How often does the winner of the World Series score fewer runs than the loser?” It’s probably a question someone out there has investigated before, but it was fun for me to try on my own… and the results are also encouraging for the Royals fan at this moment. Some of my conclusions:

1. The lopsided run result is more common than you might think. In the 109 World Series matchups since 1903, it has occurred 23 times — a little over 21 percent of the time. (I’m including 1948, when the Cleveland Indians beat the Boston Braves with both teams scoring 17 runs in the series.) It comes as no surprise that all these examples come from World Series that went to six or seven games, since this outcome is technically impossible when there are only four games and extremely unlikely when there are five.

It’s also not surprising that in series of this type, the overall numbers of runs scored by both teams are closer to each other than in average championships. The losing team tends to have scored about four more runs than the winner (and the average goes down to three when you exclude a notable outlier that I’ll discuss in a moment). By contrast, in World Series games overall, the winning team leads the losing team by an average of about 6.8 runs.

I also noticed that the lopsided series tend to have a higher overall number of runs than the average World Series. However, I suspect that has more to do with the higher number of games played than anything else.

2. The lopsided run data seems to bode well for the Royals in this series. The two teams that have been on the losing side of a lopsided run result most often are the Giants and the Yankees — both four times. But given the Yankees’ obnoxious habit of making it to the Fall Classic again and again, these losses represent a much small percentage of their World Series appearances — just 10 percent. They also won with fewer runs than their opponents in 1962, 1977, and 1996.

The Giants, on the other hand, have lost about 21 percent of their championship appearances in this way. In fact, a full third of their losses have occurred with more runs than the winning team. And they’ve never won with a lopsided result. (My father-in-law, a long-time Giants fan, dismissed this result by saying, “When you’ve lost twelve World Series over the years, you find all sorts of ways to lose.”)

Of course, this stat has little predictive or strategic value in the middle of a World Series — overall, scoring more runs all the time is a much more effective path to victory. But it’s a nice reminder that double-digit run totals like last night’s don’t necessarily foretell a Giants victory.

3. If you’re still worried about this series, have a look at 1960.

When putting together my data, one year clearly stuck out. In most years with a lopsided run result, the losing team had just a few more runs than the winners. But the Pittsburgh Pirates beat the New York Yankees in 1960 despite the fact that they scored 28 fewer runs than the Yanks.

It’s worth sharing the scores of the individual games to drive this point home.

Game 1: Yankees 4, Pirates 6

Game 2: Yankees 16, Pirates 3

Game 3: Yankees 16, Pirates 4

Game 4: Yankees 2, Pirates 3

Game 5: Yankees 2, Pirates 5

Game 6: Yankees 12, Pirates 0

Game 7: Yankees 9, Pirates 10

Heading into Game 5 this year, the Giants had scored 22 runs and the Royals 15. Even if the Royals lost another game by seven runs then won the remaining two games by one run each, the difference between run totals would only be (22 – 15 + 7 – 1 – 1), or 12. That’s still less than half the number of runs that divided the Pirates and the Yankees before the big swingers ultimately went down in 1960.

All stats from the Baseball Almanac site. Math errors are mine. I’ve provided the spreadsheet I used to figure all this out here. It didn’t occur to me to track the number of games in the series until I was about halfway done, but I went back and added the results for the lopsided series. Other potential inquiries: does it matter if the winner is from the AL or NL? In which game of the series are the losers likely to score all those runs? How does all of this correlate with teams’ other offensive stats?

Postgame update: Shutout? Meh. If the Pirates could come back from 12-0…

Here’s to Game 7 in KC!

Post-series update: One thing my father-in-law observed was that in the end, the run totals between the Royals and Giants were closer than most World Series — proof that the teams were equally matched. It’s a nice thought, though I don’t quite have the energy to investigate it right now. Here’s to 2015!

Advertisements