Earlier this week, I looked at the effect that changing leagues has on walk rate for hitters. The results of the study showed that hitters moving from the American League to the National League typically see their walk rates rise, while those moving from the National League to the American League will on average have their walk rates fall.
The logical complement to this study is to investigate the effect that changing leagues has on strikeout rates for batters. To do this, I will employ the same method as the previous study. This means that my sample set will include all non-pitchers who changed teams in the offseason only between 2002 and 2012 and who accumulated at least 200 plate appearances in each of the two seasons in question. In reality, studies show that strikeout rates stabilize a little quicker than walk rates, so I could perhaps lower the plate appearance barrier to entry for this study. However, keeping the minimum at 200 plate appearances does allow me to analyze the same player set as in the study on walk rate, for consistency.
It is not surprising to learn that league strikeout rates have been climbing on average over the last decade. Strikeout rates of non-pitchers in the American League have increased 0.30% per year, with the National League seeing an almost identical 0.29% rise per season during this period. I have removed pitchers from the control groups again, since they will skew the league averages for this study that involves all non-pitchers.
Even with pitchers discounted, the National League has on average posted higher strikeout rates than the American League, with an average delta of 0.25% during this time period.
With these two controls in place, we can adjust any strikeout rate changes observed for players changing leagues to isolate unexpected differences. As in the previous study on walk rates, a potential confounding factor in this study that is not controlled for is the age of the players in the sample compared to the league average age.
|League Switch||Avg Y-Y K% Delta||Avg League Y-Y K% Delta||Avg New League K% Delta||Adjusted Y-Y K% Delta||Sample Size|
|AL to NL||-0.10%||0.29%||0.25%||-0.64%||82|
|NL to AL||0.79%||0.30%||-0.25%||0.75%||92|
|AL to AL||0.15%||0.30%||-0.15%||76|
|NL to NL||0.68%||0.29%||0.39%||100|
Data: Fangraphs, 2002-2012, hitters with >= 200 PA
The results show that after accounting for expected changes in strikeout rate due to year-to-year league average deltas, hitters moving from the American League to the National League have on average seen a strikeout rate decrease of 0.64%. Conversely, hitters coming from the National League to the American League typically have their strikeout rate increase by 0.75%.
The first note here is that these results tend to agree with the assessment from the study on changing leagues and walk rate, in that the American League has been a stronger league. Moving into the American League thus leads to walk rate declines and higher strikeout rates. The inverse is true as well, in that leaving the American League leads to increased walk rates and lower strikeout rates. It is interesting to see that despite rising league-wide strikeout rates, and more strikeouts in general in the National League, hitters moving into the National League actually strike out less than they did in the American League at an absolute level.
The confusing aspect of these results, to me, is that hitters that change teams within the National League have actually had their strikeout rates noticeably rise, even after adjusting for controlled variables. Batters changing teams within the American League have in fact had their strikeout rates rise slightly less than would be expected, given the typical bump in the league-wide rate. The directions of these deltas are both opposite to those for hitters changing leagues. This equates to roughly a 1% difference between players moving from the National League to either of the two leagues over players moving from the American League to either of the two leagues.
Is this just more support for the case that the American League is a far tougher league in which to hit? This idea alone does not explain why hitters changing teams within the National League had strikeout rate increases. From the previous study, we learned that hitters switching National League teams had almost flat walk rates across teams. A case could be made that upon moving to a new league, hitters have to acclimatize themselves to the new set of pitchers that they will face. To a lesser extent, changing teams to another division within the same league could present the same challenges, given the unbalanced schedule in baseball. Perhaps there has been more inter-division movement in the group moving within the National League than in the American League.
I mentioned age as a factor that is not controlled in this study. While the average age for all four league-to-league groupings are close, a quick look shows that hitters moving from the NL to the AL had the youngest average age (30.0 to 31.0), while the group of hitters moving within the NL had the oldest average age (31.3 to 32.3). From this study on aging curves for hitters, we can see that the age 31 to 32 transition is when on average strikeout rates start to visibly rise again after having been quite stable for several years prior. Perhaps then there are more cases of players switching National League teams in the declining phase of their careers, helping to prop up the strikeout rates within this group.
One interesting scenario to observe in 2013 will be the performance of the entire Houston Astros player set, as the team itself moves from the National League to the American League.
In any event, the combination of effects that changing leagues has on walk and strikeout rates appears to indicate that hitters moving to the National League can expect better days ahead for these two measures, while those sliding over to the American League are in for a rude introduction to a tougher league.
You can follow me on Twitter at @MLBPlayerAnalys. Follow @MLBPlayerAnalys
Credit and thanks to Fangraphs for data upon which this analysis was based.