A retro look at predicting the NLCS

Back in the 1980s, one of my first purchases every spring was the Bill James Baseball Abstract. The stuff was so cutting edge back then, and it was the precursor to today’s explosion of sophisticated statistical analysis of baseball.

Since I did so great in predicting the winners of the National League division series last week _ I was 0-for-2 _ I thought I’d try and track down a formula Bill James used back in the ‘80s for predicting postseason winners. Sad to say, I don’t have those books anymore, but through the magic of the Internet, I found the formula. It might be considered quaint by today’s standards since there’s no mention of OBP, OPS, UZR, BABIP or any of that other neat stuff. But it’s something, and I thought I would give it a whirl here at the corner saloon in advance of the Phillies-Dodgers NLCS, which opens tomorrow.

(Don’t forget, we still got action, as Dusty used to say, on yesterday’s blog about trade rumors and the Arizona Fall League, so check that out, too.)

Anyway, here was James’ formula, which I remember him exclaiming in one of his Abstracts, “The damn thing works.” So let’s put it to the test, and we’ll see how we do when the NLCS is over, and maybe we’ll try it out for the World Series:

1. Give 1 point per half-game difference in record.
2. 3 points to team with more runs scored
3. 14 points to team with fewer doubles
4. 12 points to team with more triples
5. 10 points to team with more home runs
6. 8 points to team with lower batting average
7. 8 points to team with fewer errors
8. 7 points to team that's turned more double plays
9. 7 points to team whose pitchers have walked more men
10. 19 points to team with more shutouts
11. 15 points to team with lower ERA or if World Series team whose ERA is lower
than the league average by more
12. 12 points to team in postseason play more recently. If both were in same
year, give the points to the team that did better
13. 12 points to lead that lead in head-to-head competition

First the answer: The formula predicts a Dodgers victory as they get 71 points to 41 for the Phillies. It breaks down like this:

--The Dodgers get 4 points for finishing 95-67 compared with 93-69 for the Phils.
--The Phils get 3 points for outscoring the Dodgers 820-780.
--The Dodgers get 14 points for hitting fewer doubles, 278 to 312.
--The Dodgers get 12 points for hitting more triples, 39 to 35
--The Phillies, as expected, get 10 points for hitting more home runs, 224 to 145
--The Phils get 8 points for having the lower batting average, .258 to .278.
--The Phils get 8 points for having fewer errors, 76 to 83
--The Dodgers get 7 points for turning more double plays, 134 to 132
--The Dodgers get 7 points for their pitchers walking more men, 584 to 489
--Shutouts are a wash, with each team having 9.
--The Dodgers get 15 points for the lower ERA, 3.41 to 4.16.
--The Phillies get 12 points for winning the World Series last year.
--The Dodgers get 12 points for edging the Phillies 4-3 in head-to-head play this year.

I don’t remember Bill’s methodology for this formula. You had to figure the Phillies would lead in homers and runs because of their hitter-friendly ballpark and that the Dodgers would lead in ERA because of spacious Dodger Stadium. I think I remember reading Bill saying, "Never bet a dime on a team that hits a lot of doubles," or some such.

I thought I’d throw some other numbers out there not considered in this formula:

--Run differential: Dodgers were plus 169. Phils were plus 111. Dodgers had a Pythagorean record of 99-63, while the Phillies were at 92-70.
--OBP: Dodgers led .346 to .334.
--SLG: Phillies led .447 to .412.
--OPS: Phillies led .781 to .758.

If anybody’s got a different formula, post it here. Or maybe you remember this one. Given my record of predictions this fall, who am I to go against the Bill James formula?

Dodgers in seven great games.