Designated HitterJuly 20, 2006
Let 'er Rip II
By Larry Borowsky

On July 4, Tony La Russa said this to a beat writer from the St. Louis Post-Dispatch:

I disagree with a lot of fans, some experts, whatever. There are times when taking a strike is a good baseball play. There are times when getting deep in a count is a good baseball play. But more often, aggressiveness with the first good strike you see gives you a better chance to be productive offensively. If you give a quality (pitcher) strike one, it's tougher to have a good at-bat.

It wasn't the first time we've heard that from Tony; he gave Buzz Bissinger much the same opinion in 3 Nights in August. It's a little bit jarring, because La Russa's offenses are generally patient and disciplined; they run up good walk totals and have high OBPs, characteristics that seem incompatible with a philosophy that seems to favor hacking at the first ball that catches a piece of the plate.

But is La Russa right? Will an offense get better results if the hitters just let 'er rip? I examined that question last year, looking at 2004 National League data. My study focused on the first pitch of the at-bat -- the 0-0 delivery, which Craig Burley described a couple years ago at The Hardball Times as "the predominant count in baseball." Of course it is: For every plate appearance, there's an 0-0 pitch. And it's a strike about 60 percent of the time.

What a hitter does with that strike goes a long way toward determining the outcome of his at-bat.

The 2004 data seemed to show that La Russa is generally correct: if the first pitch is a strike, the batter is best served to swing. I repeated the study this year, using 2005 data, and got nearly identical results. Take a look:

2004 2005
when batter . . . avg obp slg rc/27 avg obp slg rc/27
takes strike on 0-0 .233 .275 .353 3.4 .237 .271 .357 3.4
swings on 0-0 .270 .300 .439 4.9 .269 .296 .440 4.8

Keep in mind that the "swings on 0-0" section of this table includes at-bats where the batter swings and either misses or fouls the pitch off, resulting in an 0-1 count -- the functional equivalent of a called strike. Such is the result more than half of the time; the first-pitch hacker puts the ball into play on only about 45 percent of his swings.

It's in part for that reason that the conventional wisdom these days favors deep counts and abhors early-resolving at-bats -- the line of thought La Russa was reacting to in his quote to the Post. The reasoning goes something like this: the more pitches you see, the better your odds of drawing a walk or getting a mistake to exploit, hence the better the chances of getting on base or driving the ball for an extra-base hit. But the numbers in this table indicate the exact opposite -- if the 0-0 pitch is a strike, your odds of reaching base or driving the ball are better if you swing.

This would seem to be the proper time to repeat my disclaimer from last year:

I'm not suggesting that hitters should swing at any 0-0 offering that comes within a foot of the strike zone; these are macro figures which mask all sorts of micro situations in which it might make sense to take a strike. If the pitcher breaks one off on the corner or puts a sinker in at the knees and you can't do much with the pitch, might as well take. And if the pitcher is struggling with his control, maybe it's not a bad idea to see if he can follow up strike one with strike two. But it is a bad idea simply to take a first-pitch strike on principle. La Russa's instincts are correct: you ultimately score more runs if you attack the first strike you see.

The Cardinals as a team tend to reflect La Russa's philosophy; in both 2004 and 2005 they ranked among the league's most aggressive 1st-pitch teams, swinging at more than 50 percent of 1st-pitch strikes. But tendencies vary widely from roster to roster. For instance:

0-0 strikes swings swing %
Cubs 3,499 1,877 54
Mets 3,471 1,517 44
Florida 3,474 1,331 38

These three teams saw a nearly identical number of 1st-pitch strikes in 2005 -- about 21 a game. The Cubs swung at 11 of 'em, the Mets only 9, the Marlins just 8.

How did it affect their overall scoring? In this case, not much; the trio finished 7th through 9th in NL scoring last year. Indeed, the strong correlation between 1st-pitch aggression and high run totals that I found in 2004 did not recur in 2005. Whereas the 8 most aggressive teams from 2004 scored, on average, 104 runs more than the least aggressive teams, in 2005 it was a total wash:

Run scoring was dead even (an average of 721 runs per team) on both sides of the median for 1st-pitch swing frequency.

The same variation in 1st-pitch aggression did not apply to pitching staffs. Opponents' swing frequency fell within a narrow range (43 to 49 percent) for 15 of the 16 teams; the Milwaukee Brewers, for some reason, threw particularly juicy strikes on 0-0 and induced a 52 percent swing rate. Opposing hitters' aggression seemed to work to the Brewers' advantage: They allowed the third-lowest OPS in the league when throwing a strike on 0-0. But the correspondence did not hold leaguewide; we can't say that pitching staffs that "induced" more swings on 0-0 got better results. Indeed, it's not clear how you would replicate that tendency over time; the discretion lies entirely with the batter.

One final note: in 3 Nights, Bissinger claims that La Russa specifically wants his guys swinging away in RBI situations. Turns out that Tony is far from alone in that regard. I broke out the 1st-pitch data for RISP situations and found that all 16 teams swung at 1st-pitch strikes more often in those situations than otherwise. Leaguewide, the frequency increased by about 9 percent -- from 44 percent in non-RISP situations to 53 percent with RISP. That only makes sense: With runners in scoring position, there may be a reward for simply putting the ball in play; you can move a guy up or knock him in with a groundout or a flyball. A strikeout, conversely, gets you nowhere -- so big-league batters, as a group, alter their approach accordingly, expanding their swing zones from the beginning of the at-bat.

There's no universal formula here; as La Russa himself says, "There are times when getting deep in a count is a good baseball play." But there are also times -- lots of them -- when a guy's got to just take his hacks. Times, in fact, when it's the smart play.

Larry Borowsky writes about the Saint Louis Cardinals at Viva El Birdos, one of the blogs in the Sportsblog Nation family.

Comments

Quick question:

Is there any way to divide the numbers of players who swing and either miss or foul it off (getting that same 0-1 count as if they had just taken it) from the players who put the ball into play? If players who get an 0-1 count fair the same regardless of how they got there (swinging or taking), then the above differences in RC/27, etc. are just a product of players who are able to put the ball into play on the first pitch doing better than players who start a count out 0-1.

I think the key point is the phrase "the first GOOD strike you see". The issue is not whether a batter should take the first pitch regardless, but whether he should swing at any strike or at a hittable strike. I do not see La Russa saying that batters should "let her rip". Rather, he seems to be saying that batters should look to swing at a hitter's pitch, regardless of the count, rather than allow the pitcher to get into a count where he is forced to swing at a pitcher's pitch. I think Ted Williams would agree with that.

Baseball seems a sport of disciplined aggression, a balancing act. Thus it makes sense to defend the plate against pitchers on an 0-0 count, not let them feel they can get an easy strike and gain an advantage. Batters still have to balance aggression with judgment and try to selectively choose which pitches can be best hit and which are best left alone. Beyond the 0-0 count issue, managers would like to encourage controlled aggressiveness by their players in a general sense, including baserunning and fielding. So getting a green light on 0-0 seems in keeping with an overall aggressive approach to the game.

responding to mraver --- yes, the success of 1st-pitch swinging does rely on putting the ball in play. if you swing or foul it off, you're no better off than you would be if you took the pitch.

but you can't put it in play if you don't swing --- that's the whole point.

what's the best-case scenario if you take the pitch? you're ahead 1-0 in the count. from 1-0, batters hit .271 and slugged .439. but when they swung on 0-0 --- when they swung, not when they put it in play -- batters hit .268 and slugged .440. the only thing gained by taking a pitch and having it called ball 1 is a vastly improved chance of drawing a walk. you're no more likely to get a base hit or an extra-base hit --- so if what your team needs as a hit and not a walk (ie, it's an RBI situation), you better go up there swinging.

bob r --- good clarification. it's a matter of looking for a certain pitch in a certain zone --- recognize it, swing at it.

that's why i sometimes argue that 1st-pitch swinging requires more discipline than deep-count hitting. the batter is not just being asked to discern between balls and strikes; he's got to distinguish between pitches within the strike zone, ie distinguish between hittable strikes and "pitcher's pitches."

once you get behind in the count, especially when there are two strikes, such distinctions are out the window. if it's close, you almost have to swing; matter of survival. there's a lot less discretion involved.

Unfortunately, drawing any conclusions from that table is reliant on the assumption that 1st pitch strikes are evenly distrubuted among hitters.

Say, for example, the league gets 60% FPS, but Pujols and other good hitters only gets maybe 35% FPS (pitchers trying to pick at the corners more). That skews the overall totals toward the lower end of the hitters' spectrum. There are simply a lot of at-bats that are not included in the FPS table, and it's not necessarily random.

And if certain types of hitters react differently to the same situation, it's wrong to draw a single conclusion for all batters. Maybe pitchers (as batters in the NL) tend to take 3 strikes and sit down. Granted, thier performance couldn't get any WORSE if they swung away at the first decent pitch, so I suppose by default you could say it gets better, but I wouldn't go so far as to apply that reasoning to ALL hitters.

Maybe there are a few hacktastic guys in the NL who also happen to be good hitters (like Vlad in the AL, but I don't know many NL hitters as well). Thier approach may not work for everybody.

Also, a batter may get some benefit to swinging at the first strike because there's an aggressive pitcher who wants to get ahead early. If everybody started swinging at the first strike, pitchers would make an adjustment in response. In other words, it may only be an advantage right now to do so because under 50% of the batters do it. Lose the rarity, lose the advantage.

I know, these are general observations, so you aren't saying everybody should swing at the first pitch. I also know that every hitter can improve; nobody is perfect up there. But there are enough doubts in my mind that I question whether the 100 point jump in OPS is merely a product of the current environment or whether it can actually be realized (even to a lesser degree) by very many batters.

If I'm the manager, I tell the guys that they are MLB hitters, so they must be doing something right. Get comfortable up there and do your thing.

true --- not all batters are the same. brian giles for example thrives on deep counts. i wouldn't advise him to change his approach and start hacking at 1st-pitch strikes.

that's not really the point of the article. it's not prescriptive; it's descriptive. while the infinite variation in batter abilities, game situations, etc often dictates the take-a-pitch approach, the aggregate data comes down very strongly on the side of 1st pitch swinging. so while there is great variation, the preponderance of it falls on the side of 1st-pitch aggression.

that runs counter the opinions la russa attributes to "a lot of fans, some experts, whatever" -- ie, it's at odds with the sabermetric view, which broadly speaking tends to hold patience and deep counts as virtues.

i'm only saying they're not virtues in all cases -- in fact, they may not even be virtues in a majority of cases. they are important and they do have their place; but perhaps there is a tendency to overstate their value.

la russa's philosophy might be phrased as "don't take a strike if you don't have to" --- strike as early as you can. before seeing this data, i would have thought such an approach, in the aggregate, decreased overall offensive production. in fact, the opposite is true.

am i the only one who's surprised and a little intrigued by that?

There was a section of Moneyball about Scott Hatteburg, where coaches in Boston told him he had to swing earlier because he batted around .500 when he swung at the first pitch. Hatteburg replied that the reason for the stat is that he only swung at the first pitch when it was too good to pass up.

Comparison between the swinging and taking groups is based on the two groups getting similar pitches. However, all of these guys "are MLB hitters, so they must be doing something right" (Thanks, Aaron). Batters will self-select into one of the two groups; they are going to make sure anything too good to pass up ends up in the "swinging" group, which would certainly improve the batter's stats. Secondly, this self-selection is likely also indicative of the pitcher. Pitchers who start with pitches in the zone that the batter can't offer at are likely throwing better than their counterparts, above and beyond giving up more first-pitch hits.

This can be investigated by comparing the "taking" group to swing-and-misses and foul-offs. If there is no difference, then the swing or take distinction is only a question of whether the first pitch is the best you'll get. If there is, then this gets even more interesting. Thanks for posting this.

thanks ryne. actually, the swing/foul group is slightly less productive than the "take" group, which i attribute to this: a "swing/foul" result is more likely to occur against a very good pitcher. the batter likes the pitch, swings, but doesn't make solid contact; that more often happens against very good pitchers than against mediocre ones.

i don't entirely agree with the premise that "the swing or take distinction is only a question of whether the first pitch is the best you'll get." some batters may be taking pitches they ought to be swinging at, in hopes of getting a better pitch later in the count. they may be overly judicious, making their "swing" zones too small --- they may be letting the "best" pitch go by on 0-0.

indeed, we've all seen it happen many times. and the numbers are certainly at odds with the idea that hitters who take strike one are likely to get a "better" pitch later on in the count.

the only instances in which a "better" pitch would ensue are when strike 1 is such a bitch of a pitch that it's almost unhittable --- ie, would reduce the average hitter to a .200 success rate. called strikes account nearly 1/3 of all 0-0 pitches; i can't believe that all of them are such great pitches.

i don't know how we would test for this however.

This is an interesting article and does stand to reason, but ignores the ancillary benefit of an organizational approach to multiple-pitch at-bats. The chief goal is to get the starting pitcher out of the game and feast on what is historically the soft underbelly of the staff - middle relievers and to a lesser extent setup guys.

I suppose the argument could be made that you could knock the pitcher out of the game earlier simply by creating an extra 1.5 runs per game with good swings on first pitch strikes, but as it's unclear WHO is pitching those strikes, I'm not sure that statement can be validated.

In other words, how many of these extra runs are created against relievers' first-pitch strikes, when the approach is no longer to run up the pitch count? That to me is a crucial part of this analysis.