One of the most disappointing parts of the 2009 season was watching Nick Markakis’ regression at the plate. I went into it at greater length previously, but the gist of it is that after posting a .406 OBP and walking 99 times in 2008, those numbers fell to .347 and 56 last season. At the former points he was one of the best hitter in the league, and at the latter levels, he was merely OK. What might have caused such a drop?
So this started out as one post, but every time I looked into one point it raised another question. Thus the “Part One”. Because a lot of this is presenting different graphs and tables, I thought it might be easier to handle in smaller chunks. There ended up being five parts, which will be posted throughout the week. Stick around for the thrilling conclusion!
First off, we’d need to know what factors affect how often a player takes a walk. My assumption is most important component of a player’s game for that would be how often he swings at pitches outside the strike-zone. If you take them, then they’ll tend to be called balls and once you take four of them you can jog on over to first-base.
So how linked is walking with swinging at pitches out of the zone? Here’s a graphs for the 2009 batters that qualified for the batting title of O-Swing% from FanGraphs (the rate at which they swing at pitches out of the zone) versus walk rate, along with the regression line (R²=.481) and lines representing the averages of both stats for this group of players. Nick is the red dot:
You would think that Markakis should have been walking more to begin with and – seeing as how this is very close to the multi-year correlations found here* – that pattern was persisted throughout Nick’s career (even in 2008; he “should” have walked more like 14.6% of the time instead of 14.2%). Part of that is likely due to a higher contact rate – those batters in Nick’s quadrant were better than the total average at putting the bat on the ball – since that would somewhat counteract getting into deeper counts.
* Which also found that O-Swing% was the most highly correlated factor with walk rate, as I had assumed.
The batters in the lower right quadrant – guys that swung at balls less than average and walked more than average – had slightly below average contact rates as a group. Just taking the players with better than average O-Swing%, and looking at contact rate versus walk rate:
The R² is only .216, but there is some correlation there. For the whole group, contact rate only had an R² of .096, but once we get some information on the O-Swing% then it makes a bit more of a difference. Since Nick has an above average contact rate, you would expect him to be a bit below the first trendline to begin with, so his walk rate would be a little lower than just his O-Swing% would indicate. Additionally, his contact rate went up to 86.4% in 2009 from 84.6-84.9% in the preceding couple years, so that explains a small part of the decreased walk rate – especially since it went up on pitches both in and out of the strike-zone. Still, the swing rates are probably the biggest part of Nick’s drop in walks, and that’s what I’ll start to look at in part two.