:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/31083881/gyi0064192939.0.jpg)
One of the best parts about writing for Beyond the Box Score is the comments section. The feedback is almost always constructive and the commentators raise interesting questions. The comments section from my post last week about Freddie Freeman's batting eye inspired (I think) Scott Lindholm's article about first-pitch hitting, so it's only fair that the comments section on that article inspired me to write this piece.
adenzeno wrote:
The other side of this coin is what happens when the batter swings and does NOT put the ball in play- he is now down 0-1…and THAT is not a good situation for a hitter. So when that is factored in, does it make sense to be aggressive on the 1st pitch? One year I kept track of what the result was when the batter swung and made the count 0-1 vs taking strike one, and the results were better when they swung, but it was a 30 game High School season, and the more aggressive hitters in HS are often better hitters so I don’t know what validity that would have being such a small sample size..Great stuff here BTW- Thank you
So I went back to the PITCHf/x database and found all those plate appearances from last season where a hitter fell behind 0-1, either because of a called strike, swinging strike, or foul ball. (I excluded failed bunt attempts.) I then compared the expected and actual matchup on-base percentages using the tried and true logged odds ratio method. A reminder: this method can be used on any percentage statistic (OBP, K%, BABIP, etc.) by first converting the statistic to an odds ratio (OR = p/1-p) and then plugging the respective odds ratios into the equation
expected OR = (batter OR * pitcher OR)/league OR.
Using this, we can set up a logit regression to compare the actual outcome for each plate appearance that started 0-1 with the expected outcome. To eliminate the effect of batters with 1.000 OPS, I only looked at matchups where the batter had at least 100 PAs and the pitcher had at least 100 BFs.
Before we get into that, though, let's see what we can observe from the data alone. Compared to the league average OBP of .318, batters in an 0-1 count find themselves at a considerable disadvantage, getting on base at a .263 clip after that point. But we can see that batters who took strike one (as opposed to missing or fouling it off) did end up with a higher OBP than those who didn't. Looking at the average OBP allowed by the pitchers in those matchups, though, we get a sense of the real culprit: the pitchers who got a first-pitch swinging strike tended to be a little bit better pitchers, at least by OBP against.
| called | swinging | |
|---|---|---|
| Matchup OBP | .273 | .270 |
| Pitcher OBP | .313 | .311 |
| Batter OBP | .328 | .328 |
| Total PA | 44,587 | 20,727 |
The logit regression confirms this suspicion. Once we account for the difference in pitcher quality, there is no substantial difference between the OBP of hitters who swing at strike one and those who watch it.
| Swinging vs. Called Strike OBP | Estimate | Std. Error | z-value | Pr(>|z|) |
|---|---|---|---|---|
| (Intercept) | -0.29698 | 0.0302 | -9.834 | 2E-16 |
| ln(Expected Odds Ratio) | 0.92353 | 0.03887 | 23.76 | 2E-16 |
| Swinging Strike (dummy variable) | -0.01005 | 0.019 | -0.529 | 0.597 |
This conclusion makes intuitive sense. Major league hitters are tough to fool, so a pitcher with good enough stuff to get swinging strikes -- especially on a count when most batters aren't necessarily looking to swing -- is probably good enough to get a higher proportion of those hitters out.
. . .
All statistics courtesy of Baseball-Reference. PITCHf/x data courtesy MLB Advanced Media and available through Baseball Heat Maps.
Bryan Cole is a featured writer for Beyond the Box Score, and is always looking for new article ideas. You can follow him on Twitter at @Doctor_Bryan.