Monday, June 21, 2010

To K or Not to K?

A few days ago (before the Reds' absolutely dreadful performance against the Mariners), I was wondering what to make of Drew Stubbs.  He strikes out a ton, and Adam Dunn showed that Reds' fans have little tolerance for strikeouts - even from good players.  It's pretty easy to see why Stubbs strikes out so much.  According to fangraphs, only five qualified players have lower Contact %.  While Stubbs struggles to make contact, his judgment of the strike zone is adequate enough.  His O-Swing % of 24.5 is good, and his Z-Swing % of 62.4 isn't that bad.  His struggles really do just come down to being unable to make consistent contact.

This prompted me to wonder whether understanding the strike zone or being able to make contact is the more important skill to possess.  I've done some digging, and I propose that "what separates the men from the boys" is being able to swing with authority at pitches inside the strike zone.  (Everything I'm about to present is the result of pulling data from fangraphs on May 18 for the 2010 season only.)

Batters who make contact at a high rate tend to swing at fewer pitches inside the strike zone.  In the chart below, I've plotted two series.  Each data point is a player, with the Reds players as the red series and all other players as the blue series. The x-axis is Contact % and the y-axis is Z-Swing %, both from fangraphs.  I've included a trendline which shows that Z-Swing % decreases as Contact % increases.  For what it is worth, the Reds players in this exercise are Johnny Gomes, Jay Bruce, Joey Votto, Brandon Phillips, Orlando Cabrera, Scott Rolen, and Drew Stubbs.  Cabrera, Rolen, and Stubbs swing at roughly 60% of the pitches inside the strike zone, while everyone else is closer to 70-75%.  Cabrera and Rolen make contact at a high rate, while Stubbs' 71% contact rate makes him the lower-left red dot.

At first, I thought this suggested that batters who are more confident in their ability to make contact are more likely to "wait for their pitch."  In effect, I thought that high contact rate led to a low Z-Swing %.  After looking into it further, I now think something else is at work.  Before I get there though, some more findings...

Contact % has little effect on whether batters swing at pitches outside the strike zone.  Again, I'll use a similar chart to illustrate, this time with O-Swing % on the y-axis.

There is still a very slight downward trend, but it is nothing like the previous chart.  So, we've established that hitters who make good contact are less likely to swing at pitches inside the strike zone, but not really any less likely to swing at pitches outside the strike zone.  If good contact hitters were "waiting for their pitch", they'd be swinging at fewer pitches outside the zone than their more aggressive (inside the zone) counterparts.  What's the explanation then?

Better hitters swing at more pitches inside the strike zone.  This time, I'm comparing the Z-Swing % to wOBA, again from fangraphs.  I've moved the Z-Swing % to the x-axis and put wOBA on the y-axis.

There is a fair amount of variation in this chart, so the trendline isn't a great fit, but you can see that the trendline does tend to rise as Z-Swing % increases.  As for the Reds, we can see that Scott Rolen bucks the trend here, putting up a high wOBA despite a low Z Swing %.  Only Stubbs and Cabrera produce a lower wOBA than their Z Swing % suggests among the Reds. 

We've now established that batters with higher wOBAs tend to swing at more pitches inside the strike zone, that batters who swing at more pitches inside the strike zone tend to make contact less frequently, and that how often a batter makes contact doesn't really influence how often the batter swings at pitches out of the zone.  This all suggests that better hitters (as measured by wOBA) take a more aggressive approach on pitches inside the strike zone.  They swing at pitches in the zone more frequently, and they have a higher wOBA, so they must hit them with more authority, right?

Hitters who swing at a lot of pitches inside the strike zone also hit the ball with more power.   This chart shows ISO from fangraphs on the y-axis.

Again, there's a lot of spread here, but there's also a definite trend.  The Reds story is much the same, with Rolen significantly outperforming the typical player with his Z-Swing %, Cabrera underperforming - is there no spot on the bench for him?!  Brandon Philips actually has a lower ISO than the typical player with his Z Swing %, but I suspect the change from batting cleanup to batting second has something to do with that, as he's clearly changed his approach at the plate. 

The difference in ISO between a typical batter who swings at 60% of the pitches in the zone and a batter who swings at 70% of the pitches in the zone is .050.  This suggests that batters who swing at more pitches in the zone also swing harder.  Batters who swing harder also miss more, which explains the first chart above.  Batters who swing at more pitches inside the strike zone miss more pitches too.

To summarize, while major leaguers need to be able to make contact with regularity and lay off pitches outside the strike zone, really good hitters swing harder and more often at pitches that are in the strike zone.  One final chart to drive the point home...  In this chart, I grouped players based on their wOBA, then plotted their Contact % vs. their Z Swing %.  There is a clear tendency for the dark blue and pink points (higher wOBA) to fall on the upper left and the yellow and light blue points (lower wOBA) to fall on the lower right.  Better hitters swing at more good pitches and hit fewer of them.

Drew Stubbs may very well strike out a lot because he swings and misses a lot.  As opposed to worrying about making contact more though, if Stubbs is like most players, he'd be better served by swinging (and swinging hard) at more of the strikes that he's watching go by.


  1. I think there is a major omission to the analysis if you do not separate out counts with 2 strikes since even the best will feel the need to "protect the plate". I am curious whether your graphs would be different for 2-strike counts than for the other counts.


  2. Good question. I'd love to generate the first four charts for each possible count. I'm not sure where to get the data though. I just pulled this stuff off of fangraphs, and I don't think I have a way of getting at the count-specific stats there. If you've got any ideas where to get the data, I'd be glad to take a look.

    Still, I think it is probably overkill to say that leaving the 2-strike counts in is a "major omission". It would be interesting to find out if these characteristics change as the count changes - they probably do - but that wouldn't invalidate the more general findings that I've posted here.