Tuesday, July 21, 2009

The Intentional Invitation To Disaster

The other day, when Cubs manager Lou Piniella elected to walk Albert Pujols intentionally in the top of the 1st inning, it made me wonder why Piniella's grasp of arithmetic is so much less profound than his fear of one hitter. I have wondered the same thing often about managers who lean on the intentional walk the way some people lean on canes--more for effect than usefulness.

The pitcher has such a huge natural edge over the hitter that it is almost never wise to walk a hitter intentionally. I mean any pitcher facing any hitter. Albert Pujols is a lifetime .333 hitter. As great as that is, it means that if a pitcher throws strikes to him, two-thirds of the time he'll get an out. Think about the increasing importance baseball pundits give to on-base percentage. It's the fundamental statistic in baseball these days: either you get the batter out, or he gets on base. Walking a batter intentionally is surrendering this fundamental battle without firing a shot.

In 2004, Barry Bonds received a ridiculous 120 intentional walks, more than most good hitters got in their whole career. His on-base% that year was a record .609. Only 14 times since 1900 has a hitter surpassed a .500 on-base%; that is, only 14 times has the batter actually been the favorite to get on base if the pitcher tries to get him out. That's a little hazy, however. In 2004, Bonds also walked 112 times "unintentionally," but we know that many of those involved the pitcher throwing four pitches deliberately out of the strike zone. He also got hit by pitches nine times. If the pitcher threw strikes and got him to swing at the ball, his batting average was a mere .362. That's terrific, but roughly seven times out of every 11 at-bats, he made an out.

Of those 14 .500+ on-base% seasons, Bonds had four, Babe Ruth five, and Ted Williams two. All of them drew a ton of walks, and though intentional walks haven't always been tracked closely, we know that plenty of pitchers simply didn't give Ruth and Williams any pitches worth swinging at. So the question is: is the intentional walk a good strategy, or is it giving the offense something for nothing?

I've had strong feelings on this for a long time, and have steadfastly avoided issuing intentional walks in the various table-top games I've played over the years. I played in leagues with friends in Las Vegas for many years, and as the league statistician I spent some time going through every intentional walk to see whether it helped or hurt the defense. Without a doubt, intentional walks were a recipe for disaster. There seemed to be only one time when it was advisable: walking the #8 hitter with two outs when you were pretty darn sure that the pitcher would take his turn at the plate. The most dangerous times to walk people intentionally were with less than outs and in the middle of the batting order. In other words, just where Piniella walked Pujols, with one out and a runner on third, setting up that potential double play managers crave so much.

I've had this statistic evidence from my own experience plus a lot of intuitive ideas about why intentional walks don't work, but this morning I took some time to see what the serious number-crunchers say. In 2005, Bill Felber's excellent book The Book on the Book was published. In it, Felber takes on lots of baseball conventional wisdom and strategies, and examines whether the numbers justify what the people who run baseball teams think of as "The Book," the accepted wisdom from which you dare not stray lest the press second-guess you. "The Book" tells us, for instance, that you don't let Albert Pujols beat you, even in the 1st inning.

Felber has a section on intentional walks, and lo and behold, he echoes my views. In fact, he goes even further than I do. Citing earlier research by numerous statisticians and analysts who have been looking at the data since the 1960s, he says that there are only a handful of specific situations in which an intentional walk might increase the defensive team's chances of winning, and they all occur no earlier than the bottom of the 8th inning. Even walking the #8 hitter with two outs and a runner on second to get to the pitcher in an earlier inning doesn't make it more likely that you'll win the game. Yes, it slightly decreases the number of runs the offense will score in that inning. But it brings an even larger increase in the number of runs the offense will score (on average) in the following inning, because the leadoff hitter will lead off instead of the pitcher.

Just as my makeshift research of dice-game results showed, Felber concludes that the worst time to walk a batter intentionally is with less than two outs, even if it's Albert Pujols and you're setting up that (elusive) double play. Again, you might decrease the chance that the offense will score one run, but you're increasing the chance for what we call a "crooked number," a multi-run inning which will cost you much more than that single run would. It is simply skating on thin ice to put another runner on base with less than two outs and the middle of the batting order coming up. Those guys can hit, too, and when they do, they'll drive in more runs. So what if Lou Piniella is scared to death of a guy with a .430 on-base%. The next hitter has a .380 on-base%. All this means is that Pujols will get on base one extra time out of 20 (and intentional walks will account for part of that difference). Pitch to him! Two times out of three, he'll make an out.

No matter how you slice it, the intentional walk is counterproductive. (If you want the exact stats, e-mail me or get yourself a copy of Felber's book.) And sometimes it hurts more than other times.

I wonder how Greg Maddux felt back in 2001 when he set the National League record for consecutive innings without issuing a walk--72 1/3--only to have the streak end when manager Bobby Cox ordered him to issue and intentional walk. It was the 3rd inning, he was already losing 4-0, and there was a runner on second with one out. For some reason, in what was hardly a crucial spot in the game, Cox had Maddux throw four wide ones to the immortal Danny Bautista, a .272 lifetime hitter who drew exactly one intentional walk in 2001. It set up a double play, but when Maddux got the next batter to hit a ground ball, the only out was at first base. That left runners on second and third, and Cox ordered another intentional walk! This was one I might even have gone for, since it brought up pitcher Albie Lopez, who was 0-for-19 for his career when he went up to face Maddux. He grounded out; Maddux lost anyway, 9-1. Without the forced passes, he would have tied Bill Fischer's major-league record of 84 consecutive walkless innings. Sorry, Greg.

I recently experienced Exhibit A in the argument against intentional walks. I've been playing in another table-top league, this one based on Statis-Pro Baseball as reinvented and improved by Ty Waterman. Ty is the commissioner of the Great American Fantasy League (GAFL), and I'm managing the all-time Mets team. Last week my Mets played at St. Louis, and trailed 3-1 in the top of the 6th inning. I got runners to second and third base with nobody out, Jose Reyes up, and Mike Piazza (batting 3rd in the order) on-deck. The Cardinals manager decided to walk Reyes intentionally. I didn't say anything, not wanting to give him any reason to change his mind, and licked my chops at the thought of having the bases loaded with nobody out and the middle of my order coming up. He was hoping to get a double-play grounder from Piazza that would make the score 3-2 but kill my chances for doing much beyond tying the score.

Indeed, Piazza did hit a grounder, but Scott Rolen kicked it, and everybody was safe. Now the intentional walk changed from a calculated gamble to an invitation to disaster. David Wright singled in a run to tie it 3-3. Darryl Strawberry singled in a run to give the Mets the lead. After a pop-up for the first out, Lee Mazzilli walked, forcing in a run. By now we were up to the third Cardinals pitcher of the inning. It didn't matter. He walked Mookie Wilson (unintentionally) to force in the run that made it 6-3 Mets. And the carnage was just beginning. I sent up Kevin McReynolds to pinch-hit for Lenny Dykstra, who had pinch-run for Cleon Jones, whose pinch-hit single had led off the inning. McReynolds doubled in two runs, and it was 8-3.

That brought up Howard Johnson with runners again on second and third. Did the Cardinals manager learn from his not-so-long-ago intentional-walk disaster? Nope. Still with only one out, his craving for the double play somehow increased, and he walked Johnson intentionally to bring up Jose Reyes. Can you guess what happened next? Al Hrabosky hit Reyes with a pitch. 9-3. Piazza singled in a run, ending a futile five-hitter debacle for the "Maddened Hungarian". Jesse Haines came in and got Wright to pop up, but thanks to the two intentional walks (which according to every baseball percentage would have resulted in at least one out by the batters), the inning wasn't over. Nor was the carnage. Strawberry singled in two runs, another scored on an Edgardo Alfonzo single, and Mazzilli capped the outburst with a three-run homer.

The final tally: a 15-run explosion fueled by two intentional walks. Of course, this isn't going to happen to your favorite pitcher the next time his manager makes him walk someone intentionally. But it doesn't take a ridiculous number of runs to beat you. It might only take two or three, and the intentional walk only increases the chance of a crooked number. So unless it's one of the extreme situations cited by Felber and the analysts he cites (for instance, a tie score and a runner on second with one out in the bottom of the ninth), intentional walks will do nothing but produce more runs. The way they're overused by today's managers--and it's useless to point to what happened in this or that instance, because we're talking about the long haul, the overall real percentages--they are an open invitation to disaster.

No comments: