April 30, 2006


A Walk May Not Be as Good as a Hit (ALAN SCHWARZ, 4/30/06, NY Times)

David Neft, a retired vice president for research at Gannett and an editor of several baseball statistical tomes, was looking at OPS recently and realized it needed some updating. When he started his work, he harked back to his Columbia University economics classes.

"You look at it like opportunity cost," Neft, 69, said. "It isn't just what you do with capital, but what you could have done."

When considering the value of a batter's walks, he reasoned, the benefit of reaching first base should first be diminished by the opportunity cost of his power being unplugged.

"Not all walks are created equal — they're batter-dependent," Neft said. "Any manager is less upset when his pitcher walks the cleanup hitter than when he walks the No. 9 hitter. They aren't crazy. They intuitively understand this concept."

Neft prefers to view walks somewhat backward, through the eyes of the pitcher. In figuring what he calls on-base advantage, walks (and times hit by pitch) are weighted not as full-unit successes for the batter, but by their marginal benefit beyond the batter's sidestepped slugging percentage.

For example, walks for Pujols are worth only .110 to him (1 minus his gargantuan .890 slugging percentage entering Friday's games). To a less brawny batter like his St. Louis teammate Yadier Molina, walks are worth .792 (1 minus .208).

However jarring to those riding the modern walk bandwagon, Neft's refinement makes perfect sense. From the pitcher's standpoint, a batter expected to slug 1.000, on average, should always be walked because his average hit is more damaging than a walk.

Meanwhile, walking a player with a .000 slugging percentage is grounds for an early shower, because he is no threat in the first place. The higher the slugging percentage, the less costly the walk.

Neft then adds a batter's on-base advantage to slugging percentage for a refined OPS — call it OAPS — to get a better idea of how dangerous a hitter has actually been. This does not knock Pujols off his perch as the season's best hitter so far, but it does bring him back to the pack somewhat. His 1.385 OPS is 82 percent higher than the National League average of .760; his 1.250 OAPS is 63 percent higher than average.

In contrast, players who rarely walk, like the Blue Jays' Alex Rios and the Rangers' Kevin Mench, move up in the rankings because their slugging is unleashed more often.

While Pujols, Jason Giambi and other players who walk frequently sit in dry dock for a dozen or more plate appearances every month, the likes of Rios and Mench are getting to swing the bat (for now) and have their slugging affect games.

"It doesn't always make much difference in rating hitters, but it's a more realistic reflection of what's going on in the game," Neft said.

The other night Hector Luna stole second with Albert Pujols up with predictable results. Even on a passed ball or wild pitch the runner should stay at first when Albert is at the plate.

Posted by Orrin Judd at April 30, 2006 11:28 AM

So it makes no sense for AL pitchers to have walked Manny, with his .470 slugging %, 20 times so far this season.

Posted by: Jim in Chicago at April 30, 2006 11:41 AM

He's right about the situational value of walks, but his measure is wrong.

What they ought to do is look at the full situation. With a runner on third and one out, an outfield fly is greatly increased in value and a walk (which sets up a potential double play) is diminished. Similarly with a runner on second and 0 or 1 outs.

Also, they have to look at the rest of the lineup. If you're followed in the lineup by .400 hitters, a walk is excellent -- there's a good chance they'll hit you in. If you're followed by .200 hitters, you're likely to be stranded, and it would be better to swing for extra bases.

This sort of analysis is what computers were made for. You have to record every pitch and action, but they're doing that now. It lets you analyze a player's value rigorously (through Monte Carlo simulations), and you can say with confidence how a player's value changes as the lineup he's in changes, as the pitcher he's facing changes, and so forth.

oj, it's wrong to say that a runner shouldn't advance on a wild pitch with Pujols at the plate. The pitcher always has the option of intentionally walking him, so getting the intentional walk doesn't hurt the other team. And if they pitch to Pujols, you're better off with the runner at second than first. I would agree that you shouldn't take risky steal attempts with Pujols at bat.

Posted by: pj at April 30, 2006 12:03 PM

His measure is wrong also because he's not taking outs into account. Even if a batter's slugging percentage is, on average, 1.000, walking him every time is dumb because while the batter will get the same number of bases on average if you pitch to him, he'll still make outs.

Teammates make a difference too, exactly as you say.

Posted by: John Thacker at April 30, 2006 1:06 PM

Many years ago, the Cleveland Indians noticed that a particular slugger (it might have been Boog Powell) had a slugging percentage greater than 1 against Sam McDowell. So they had Sudden Sam walk the guy every time, including leading off an inning.

They only did it for a while. I suspect they got a letter from the league office containing the phrase "a travesty of the game." The Indians used to get a lot of letters like that.

Posted by: Bob Hawkins at April 30, 2006 6:04 PM

If Tony LaRussa really was a genius he would recognize what a good hitter John Rodriguez is and put him in left field in the #2 spot in the lineup against righties the rest of the year. Pujols needs protection in front of him just as much as he does behind.

Posted by: andrew at April 30, 2006 6:55 PM

Taguchi is a better player.

Posted by: oj at April 30, 2006 7:02 PM

Certainly on defense and running the bases Taguchi is better. But offensively Rodriguez hasn't had a run of games to really see what he can do. Plus I like Rodriguez cuz he's the only one on the team who hits well against borderline psychopath Carlos Zambrano. Who "accidentally" threw a ball two feet behind Carlos Lee today.

Posted by: andrew at April 30, 2006 8:59 PM

OPS has been batted around before. (Sorry, just couldn't resist.) In their Total Baseball encyclopedia, John Thorn and Pete Palmer explained how their linear weights formula was just a rejiggering of plain old slugging average - of course, a component of OPS - with new weights and a couple extra terms thrown in.

But then they said that ordinary OPS is pretty good, too. If you want to help your team score runs, just try to get your OPS as high as possible, and let statisticians worry about tiny refinements.

Posted by: Casey Abell at May 1, 2006 11:41 AM