Simplifying Batter Strike Zone Knowledge for Pitch F/X Data

by Jeff Zimmerman on Jan 30, 2010 11:43 PM EST in Pitch f/x Goodies

I have been working with Pitch F/X data to determine a player's knowledge of the strike zone. Fangraphs.com has some values for hitters and strike zone, but I wanted to control for the strike zone called by umpires and make the data more usable. The following looks at each step in determining a hitter's strike zone judgment and possible improvement methods.

I began by taking the total pitch counts for right handed hitter's (RHH) and left handed hitter's (LHH) balls and strikes for 2007, 2008, 2009. I used the LHH and RHH strike zones I determined previously. I have been working at refining them further, but will use the previous zones for now. Here is a list of descriptions of each pitch as available from Pitch FX:

Pitch FX Description

Ball

Ball In Dirt

Called Strike

Foul

Foul (Runner Going)

Foul Bunt

Foul Tip

Hit By Pitch

In play, no out

In play, out(s)

In play, run(s)

Intent Ball

Missed Bunt

Pitchout

Swinging Strike

Swinging Strike (Blocked)

Unknown Strike

(blank)

I removed the following values as they do not matter in determining the batters judgment or the pitch data is unknown:

Blank, HBP, Intent Ball, Pitchout and Unknown Strike

The following values were combined as they fall into the same general categories.

Balls: Ball, Ball in Dirt

Fouls: Foul, Foul (Runner Going), Foul Bunt, Foul Tip

Swinging Strikes: Missed Bunt, Swinging Strike, Swinging Strike Blocked

Here are the %'s for each category from 2007, 2008, 2009 and all 3 years for both RHHs and LHHs:

RHH Balls	2007	2008	2009	3 years	LHH Balls	2007	2008	2009	3 years
Balls	62.0%	63.5%	62.9%	63.0%	Balls	68.4%	70.0%	70.2%	69.8%
Called Strike	6.1%	5.7%	5.8%	5.8%	Called Strike	4.2%	3.5%	3.4%	3.6%
Fouls	12.3%	11.9%	12.2%	12.1%	Fouls	10.9%	10.5%	10.1%	10.4%
In play, no out	2.1%	1.9%	2.0%	2.0%	In play, no out	1.6%	1.5%	1.5%	1.5%
In play, out(s)	6.7%	6.6%	6.6%	6.6%	In play, out(s)	5.3%	5.0%	5.1%	5.1%
In play, run(s)	1.2%	1.0%	1.1%	1.1%	In play, run(s)	0.8%	0.7%	0.7%	0.8%
Swinging Strikes	9.6%	9.3%	9.5%	9.5%	Swinging Strikes	8.9%	8.8%	8.9%	8.9%

RHH Strikes	2007	2008	2009	3 years	LHH Strikes	2007	2008	2009	3 years
Balls	10.3%	9.3%	10.2%	9.8%	Balls	11.1%	10.7%	10.6%	10.7%
Called Strike	28.2%	28.6%	29.0%	28.7%	Called Strike	28.9%	29.1%	30.1%	29.5%
Fouls	23.6%	24.2%	23.4%	23.8%	Fouls	23.8%	24.0%	23.9%	23.9%
In play, no out	6.8%	6.9%	6.7%	6.8%	In play, no out	6.7%	6.8%	6.6%	6.7%
In play, out(s)	18.7%	18.9%	18.7%	18.8%	In play, out(s)	18.1%	18.2%	17.7%	17.9%
In play, run(s)	3.8%	3.7%	3.7%	3.7%	In play, run(s)	3.6%	3.7%	3.6%	3.6%
Swinging Strikes	8.6%	8.4%	8.2%	8.4%	Swinging Strikes	7.7%	7.6%	7.5%	7.6%

I see no reason not to use the 3 year combined data for comparison.

So far it has been pretty easy.

Comparison of RHH and LHH totals shows values a little off, especially on balls out of the zone. I am not sure the discrepancy is because of differences in quality between LHH and RHH or that the umpires call pitches differently depending on hitter handedness. I don't feel exactly comfortable, but I will combine the numbers at this point or a baseline for both RHH and LHH:

	LHH Balls	RHH Balls	Combined Balls	LHH Strikes	RHH Strikes	Combined Strikes
Balls	69.8%	63.0%	65.8%	10.7%	9.8%	10.3%
Called Strike	3.6%	5.8%	4.9%	29.5%	28.7%	29.1%
Fouls	10.4%	12.1%	11.4%	23.9%	23.8%	23.8%
In play, no out	1.5%	2.0%	1.8%	6.7%	6.8%	6.7%
In play, out(s)	5.1%	6.6%	6.0%	17.9%	18.8%	18.4%
In play, run(s)	0.8%	1.1%	0.9%	3.6%	3.7%	3.7%
Swinging Strikes	8.9%	9.5%	9.2%	7.6%	8.4%	8.0%

I selected two All Star catchers to compare, Joe Mauer and Miguel Olivo (Bengie Molina All Stars). Here are the pair's combined percentages, along with the league's overall percentages:

Pitches out of Strike Zone	Combined	Miguel Olivo	Joe Mauer
Balls	65.8%	48.4%	69.5%
Called Strike	4.9%	3.8%	4.5%
Fouls	11.4%	13.9%	9.8%
In play, no out	1.8%	2.0%	2.5%
In play, out(s)	6.0%	7.3%	7.4%
In play, run(s)	0.9%	1.0%	1.9%
Swinging Strikes	9.2%	23.5%	4.5%

Pitches in Strike Zone	Combined	Miguel Olivo	Joe Mauer
Balls	10.3%	6.1%	15.2%
Called Strike	29.1%	20.5%	37.9%
Fouls	23.8%	27.5%	17.2%
In play, no out	6.7%	6.5%	7.4%
In play, out(s)	18.4%	17.3%	14.6%
In play, run(s)	3.7%	4.8%	3.9%
Swinging Strikes	8.0%	17.3%	3.8%

Now I am at a point where I am not for sure what values are useful/informative to other people. Please let me know what data is desired.

Here is my method for simplifying and improving the data the I see as useful.

Combine Balls and Called Strikes for each category. This would be the Take % for pitches that are supposed to be either a ball or a strike.
Combine the Fouls and all 3 "In Play" categories into a Contact grouping
Combine In Play, no out and In Play, run(s) into a Good grouping.
Combine all groups except, Balls and Called Strike into a Swinging grouping
Divide the Contact grouping by the Swinging grouping to get a Contact %
Divide the Good grouping by the Swinging grouping and to get a Good Contact %

Here are the Take %, Contact % and Good Contact %'s for pitches in and out of the strike zone for the league and the two hitters being compared:

Pitches out of Strike Zone	Combined	Miguel Olivo	Joe Mauer
Take %	71%	52%	74%
Contact %	69%	51%	83%
Good Contact %	9%	6%	17%

Pitches in Strike Zone	Combined	Miguel Olivo	Joe Mauer
Take %	39%	27%	53%
Contact %	87%	76%	92%
Good Contact %	17%	15%	24%

These numbers are fine, but without the combined values, which would be a pain to always provide, the percentages don't mean much. To solve this problem, I converted the percentages to a 100 scale like ERA+ and OPS+. A value of 100 is league average, while a value of 90 mean the player is 10% below the league average and a value of 108 is 8% above the league average. Here are the values again converted to the (+) method:

Pitches out of Strike Zone	Combined	Miguel Olivo	Joe Mauer
Take +	100.0	73.9	104.6
Contact +	100.0	74.1	120.9
Good Contact +	100.0	68.4	180.3

Pitches in Strike Zone	Combined	Miguel Olivo	Joe Mauer
Take +	100.0	67.8	135.0
Contact +	100.0	88.0	106.0
Good Contact +	100.0	89.2	140.9

These values allow a person to know, for pitches out of the zone, that Joe Mauer is taking more than league average and when he does swing he makes good contact. Miguel on the other hand is swinging at pitches all the time and rarely making contact when he does swing.

Let me know what you think. I like the final values, but am pretty sure there is room for improvement.

Tweet Comment 7 comments | 1 recs |

Do you like this story?

Comments

Understanding strike zones

I think you’e on to something but I will have to dig deeper to see what modifications need to be made. Also, you have to account for adjustments by the batter. Sometimes a batter may move off the plate and then lean in to make it look like he’s in the same spot but in reality he’s moved off a few inches so now pitches that are over the inside of the plate are effectively the same as being over the middle of the plate. That is something that will never show up in stats.

by xerac on Jan 31, 2010 2:30 AM EST reply actions