The DH and Hit Batsmen, Again

As some of my readers may know, I have done some research on the causes of hit batsmen differences between the NL and the AL. I certainly didn’t start the discussion in the economics literature, but if you read the papers linked below you can find its origins. As the graph below demonstrates, from the early 1970s until the early 1990s, the AL HBP rate exceeded the NL HBP rate by an average of almost 20 percent.

It just so happens that the higher HBP rate in the AL compared to the NL coincides with the introduction of the DH in the AL. There have been a few theories as to why this is. The first is that AL pitchers do not have to bat, where they might face retaliation for plunking, and therefore shirk in their pitching. NL pitchers bear a greater cost from hitting batters than AL pitchers and therefore hit less batters. In terms of the law of demand (a phrase that could ruin any good discussion), as the price of hitting batters falls, pitchers will hit more batters (or consume behavior that leads to more hit batters). But don’t get caught up in the economic terminology, this is traditional baseball folklore. When the pitcher doesn’t have to face the music, he may be a little less careful pitching high and inside.

However, the higher hit batter rates in the AL do not necessarily prove this theory. A competing explanation is that because pitchers are such poor hitters, the NL hit batter rate is lower than the AL. Hitting a pitcher puts the easiest out on the team (almost always) on base. So pitchers in the NL are extremely careful not to hit pitchers when they bat. AL pitchers do not have to be as careful against the DHs. So, we have a competing theories that both predict higher hit batter rates in the AL.

But, things get trickier. In 1994, the NL hit batter rate rose above the AL for the first time in two decades. In fact, in four of the past ten years the NL rate has exceeded the AL rate. What the heck is going on here?

Well, my colleague, Doug Drinen, and I think we have found some possible answers. Using two datasets from Retrosheet event files and game logs, we have been able to examine the causes of hit batters with the DH and without. The disaggregated data allow us to control for many factors to distinguish between alternate hypotheses. For example, we can look at the impact of the quality of the hitter on hit batters. If, after controlling for the better batting lineups created by the DH, DH games are still associated with more hit batters, then something else must be going on. Our first paper on the subject uses play-by-play data from the late-1960s/early-1970s and 1989-1992 to examine the likelihood of a batter being hit, controlling for game situations. From this data we determined that the lack of fear of retaliation among AL pitchers explained between 60-80% of the difference in hit batters between leagues during these time periods. Additionally, we found evidence of pitchers being hit in the half-inning following a plunking.

Unfortunately, the available play-by-play data exists only for the period before the league HBP rates began to narrow in the mid-1990s. So, to look at the more recent data, we used game-by-game data from the Retrosheet game logs, which extend all the way through the 2003 season. Using the data in a second paper we find that the incidence of hit batters is about 8% higher than games with no DH, controlling for many relevant factors. From 1973-2003 the hit batter difference between leagues was about 15%; therefore, the DH explains roughly half of the difference. In interleague games the difference is even more pronounced, with the DH being associated with 11% more hit batters than in non-DH games. We also find evidence of retaliation for hit batters. The more batters a team hits in the game, the greater number of batters on that team are hit themselves.

But still, what about the 1990s? Though we show that the difference in hit batsmen is still real when the more recent data is added, it still seems odd the the NL’s HBP rate should rise relative to the AL. Well, it turns out there is are two explanations for this. First, the NL expanded in 1993. This would not be problem except for the fact that the expansion draft rules allowed AL teams to protect more players from the draft than NL teams. Therefore, the talent dilution affected both leagues, but hurt the NL more than the AL. Talent dilution ought to increase the number of accidental hit batters from inexperienced players. From 1993-1997, the NL HBP rate exceeded the AL rate in 3 of the 5 years. In the 1998 expansion, there was no such asymmetry in the expansion draft, and only in one year since expansion (2000) has the NL rate exceeded the AL rate. Expansion is also consistent with the rise of hit batters in both leagues. Second, in 1994 MLB adopted the double-warning rule for hitting batters. The double-warning rule requires the umpire to warn both teams following a hit batter (or a nearly hit batter) that he deems to be intentional. After the warning, any future hit batters result in the ejection of the offending pitcher and manager. Why is this important? Well, it significantly raises the cost of retaliation. If a pitcher hits a batter, he knows that retaliation will be very costly for the other team. Therefore, pitchers in the NL could feel a little more like their AL counterparts in knowing that retaliation is less likely. The narrowing is expected.

So, there you have it. I could say a lot more, but already this post is very long. Read the papers if you are interested in this. Doug and I welcome thoughts and suggestions on the papers.

Comments are closed