Measuring Swing Effectiveness

by Joelestra on Jul 24, 2009 1:41 PM EDT

Hand-eye coordination and reflexes are imperitive to a hitter's effectiveness. As metrics such as BABIP take note of, there is a significant amount of luck involved with whether or not a hit ball goes foul, is a grounder, flies out, turns into a triple, etc. The batter's essential job is to not swing at bad pitches (unhittable balls) and to make contact with good pitches (strikes).

In an attempt to determine who was good at discerning pitches to swing at as well as ability to hit those pitches, I took a stab at making a "Good Swing":"Bad Swing" ratio as a metric.

What is classified as a "Good Swing" or a "Bad Swing" I am sure is up for debate and I hope to get some discussion around the subject. For this initial look I considered the following as "good":

1. A ball that was not swung at

2. A foul ball, including those taken on bunts and pitchouts - this was a limitation of my data...obviously a ball that is swung at and hit foul was a worse outcome than taking the ball, and you would hope once someone goes for the bunt they are successful

3. The swing put the ball into play

I considered these outcomes as "bad":

1. A strike

2. A missed bunt

For the 2008 season I summed the "good" and "bad" swings for every player and determined the ratio of good swings to bad swings. I then removed anyone with less than 502 PA's, the cutoff for a batting title.

The majority of players are group somewhere between 2.7 and 3.3 good swings for each bad swing but as with any metric we see more extreme cases in the tails. Below are the players with the best and worst ratios:

Player	Ratio
Yadier Molina	4.38742
Vladimir Guerrero	4.37867
Brian Giles	4.16421
Bengie Molina	4.06831
Magglio Ordonez	4.02472
Casey Kotchman	3.93643
Yuniesky Betancourt	3.88281
Yunel Escobar	3.87707
Jeff Kent	3.83776
A.J. Pierzynski	3.82952
Albert Pujols	3.77439
Brian McCann	3.74558
Robinson Cano	3.70175
Cesar Izturis	3.68047
Carlos Guillen	3.67233

...

Player	Ratio
Geovany Soto	2.50702
Cody Ross	2.49033
Curtis Granderson	2.48263
Dan Uggla	2.45921
Jeremy Hermida	2.44196
Casey Blake	2.41342
J.J. Hardy	2.41342
Franklin Gutierrez	2.39855
Jason Bay	2.37115
Chris Young	2.36842
Bobby Abreu	2.33779
Bill Hall	2.32929
Jayson Werth	2.23719
Jack Cust	2.23636
Mark Reynolds	2.07593

I think this is a good start but I'd like some better data on whether or not certain pitches were going to be balls or strikes before the swing was taken, as well as opinions on whether I have accurately defined "good" and "bad" swings.

Tweet Comment 5 comments | 0 recs |

Do you like this story?

Comments

Divide foul ball outs and foul ball not outs

Batters can fight off certain pitches, but foul pop up is bad.

On the same lines, maybe include infield popups as bad numbers as almost always these are caught.

Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.

by Jeff Zimmerman on Jul 24, 2009 1:53 PM EDT reply actions

There is some difficulty in the classifications in your list. In general, it tends to rewards players with high contact rates and severely devalues the Three True Outcomes players. The issue with that is that often times it’s those players who see more pitches that have the better plate approaches in terms of seeing balls but don’t have great contact rates and are hurt in this list by that. Essentially, one group is a contact group and the other appears to be a non-contact group, which somewhat misses the point of the study.

For example, I give you two players on the list, one from the best and one from the worst. Data from FanGraphs:

Player 1: 30.3% OSwing%, 74.2% ZSwing%, 96.2% ZContact%
Player 2: 19.5% OSwing%, 65.0% ZSwing%, 84.6% ZContact%
League Average: 25.0% OSwing%, 65.7% ZSwing%, 87.8% ZContact%

Just looking at those numbers, who do you think has the better approach? One guy swings at 5% more pitches out of the zone (bad),10% more in the zone (good) than average, but makes massive contact with the ball. The other guy makes less contact, takes more pitches on average outside of the zone while swinging at about the average rate. I’ll let you guys try to guess out who is who for a while.

I think distinguishing between outside contact and zone contact would help, maybe weighing outside contact less. I think there is something to be said about making contact outside of the zone, but in general I’d say those swings are less likely to produce hits for players. It’s hard when you have guys like Guerrero who show no discipline and find success as opposed to True Outcomes guys who work the count and such but don’t receive benefits due to low contact rates.

I also second everything Jeff said.

(EDIT: And I finally realize as I finish this that you already mentioned everything I was on a diatribe about. Bah. You noticed what you needed to notice, which is a good thing certainly. In zone or out of zone contact/swing data is needed, like you said. I’m seconding that.)

Marlin Maniac, a Florida Marlins blog

by SFiercex4 on Jul 24, 2009 2:36 PM EDT reply actions

If anyone knows where that data can be accessed I think it is a good next step. Essentially every pitch needs to be considered as “did the batters swing affect the play positively or negatively relative to not swinging?”. An easy one would be that a foul ball on what would have been a ball is probably not a good idea since you chance the ball being caught (although, on the other hand, one has to consider whether the foul is due to a bad swing or bad luck as a foul ball might be three feet away from being a triple). A harder one would be a foul on what would have been a strike…by making contact you have a huge improvement in the odds of a hit, but also chance being caught and out whereas the alternative is to take a strike. Perhaps I should first determine the likelihood of a foul being caught and whether or not some players are truly better than others at not hitting foul.

Also, as was pointed out, the “good” in terms of hitting the ball probably needs to be weighted player-by-player based on how god their hits tend to be…

by Joelestra on Jul 27, 2009 10:19 AM EDT up reply actions

A pitch F/X database would have all the information

Do might want to look at installing SQL and downloading the Pitch F/X dataset?

Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.

by Jeff Zimmerman on Jul 27, 2009 12:17 PM EDT up reply actions

You’re including some linear weights analysis in the idea you brought up. That would be interesting but I’m guessing vastly time consuming. You’re basically considering the run value of a given swing based on the result, it sounds like.

Also, I don’t think you’ll be able to determine the type of foul ball exactly until there’s some nice Hit f/x work available. I’m not up to date on Hit f/x as of right now, so I wouldn’t know if that’s findable (not a word?).

Marlin Maniac, a Florida Marlins blog

by SFiercex4 on Jul 27, 2009 2:26 PM EDT up reply actions