Statcast and Twins OF Defense

[Eris]

Inspired by this Statcast article on mlb

 

http://m.twins.mlb.com/news/article/221699212/minnesota-twins-2017-statcast-preview/?topicId=27118388

and

http://m.mlb.com/news/article/221576208/statcast-ranking-of-top-defensive-outfielders/

 

I visited https://baseballsavant.mlb.com/catch_probability_leaderboard?min=25&year=2016 downloaded data and did a few calculations in an excel file.  The calculations were based on giving a player the inverse of the catch probability for making the catch and substracting the difference from 1 for a missed opportunity. Being that these are bucketed into 5 brackets I took the average for that bracket. For example, for the 5 star bucket, I used 0.875*# catches and subtracted 0.125*#misses.  For the 1 star buckett, 0.07 was used for a catch and 0.93 for a missed opportunity.   This was done across all buckets and divided by the total opportunities.  

 

Several issues with the method.  Being that averages are used a large number of data points are needed to get decent data.  The range of a 5 star opportunity is 0 - 25% (or 1 to 0.75 for making the catch if the methodology could applied for each specific incident.  Also, the statcast current methodology does not take into consideration, direction of running and proximity to the wall.  This will significantly skew the results for several types of plays (such as balls near the wall or areas with short foul distances.  For example, I believe that if the sample player played LF and RF in Fenway Park, his stats in LF would be significantly worse.

 

Sorry, I don't have permission to upload this excel file.

 

In the calculations on the spreadsheet, 0 would represent an outfielder who matched the probability expectations when summed across the entire set of opportunities. I have pasted some poorly formatted data below.  The first number, is the overall ranking across all outfielders with at least 25 opportunities in 2016 (there were 156 players who met this criteria).  The first number after the name is the Calculation described above, the last number is the % balls caught.  (Other factors such as arm strength is not considered)

 

The truly elite outfielders (Hamilton, Kiermaier) catch 80% of the opportunities.

 

Byron Buxton is very good, but not elite.  The reason for this is that he only converted 38.5% of his 4 star opportunities into outs.  This is exceptionally low and stands out.  Hamilton converted 75%, Kiermaier converted 87%.  Although limited to a part time role for the Mets, Juan Legares converted 100% of his opportunities that were 4 star or less.

 

Kris Bryant is a very decent outfielder.  

 

Kepler and Rosario only converted about 60% of their opportunities.  They show up in the mid-tier of good, but not great outfielders.  For Kepler this is because he failed to convert a number to opportunities 3 stars and below.  For example, Kepler missed four 1 star opportunities.  If you correct for these, Kepler is almost a good as Lorenzo Cain.  Rosario is harder to evaluate.  He had 27 five star opportunities and only converted 2 of them.

 

Statcast is not very kind to Aaron Hicks.  He is only slightly above average.

 

Danny Santana shows up in the range of OF like Denard Span, Carlos Gomez and Mike Trout (a little more than slightly below average or about as below average as Rosario is above average).  Interesting for Santana the issue is that he only converted about 55% of his 2 star opportunities.  This kind of reflects the Santana we have come to know, great play one minute and basket case play on the next.  

 

As an outfielder, Miguel Sano profiled very similarly to Oswaldo Arcia.  

 

Robbie Grossman was/is horrible.  Readers of TD know this.  Only Daniel Nava and Mark Trumbo were worse.  For Grossman, the data is reflected in the significant number of 1 and 2 star opportunities that he missed.  He only converted 75% of  this 1 star opportunities.  He had 6 fielding errors last year (after having 5 with the Astros the previous years).  As another reader on a different thread had pointed out, he wasn't this bad with the Astros.  He also failed to convert none of his 19 five star opportunities.

 

Other notes.  Shane Robinson is a decent outfielder.

 

Statcast is not kind to  J.B. Shuck and Drew Stubbs.  This was a surprise to me because if the Twins are bringing in players to evaluate as a 4th outfielder, the minimum criteria should be that they are above average defensively.  Neither is the case, if fact both rank near the bottom using this methodology.  (Of course they didn't make the team either and Statcast suggests that was a really positive development)

 

 

1 Desmond Jennings 18.8 82.7
2 Juan Lagares 16.8 80.0
3 Billy Hamilton 16.4 77.8
4 Kevin Kiermaier 15.5 83.7
5 Jake Marisnick 14.3 67.9
6 Travis Jankowski 13.5 78.2
7 Andrew Toles 13.3 73.3
8 Ender Inciarte 12.8 72.0
9 Keon Broxton 12.4 82.1
10 Peter Bourjos 11.9 65.6
11 Jason Heyward 11.8 71.9
12 Albert Almora Jr. 10.4 88.0
13 Byron Buxton 10.0 70.5
14 Mookie Betts 9.6 71.2
15 Kris Bryant 9.6 71.1
16 Adam Eaton 9.5 71.5
17 Enrique Hernandez 9.0 72.4
18 Shane Robinson 8.7 73.5
19 Jarrod Dyson 8.6 77.3
20 Kevin Pillar 8.2 68.8
24 Lorenzo Cain 7.2 67.7
45 Ichiro Suzuki 3.5 61.4
46 Max Kepler 3.1 59.8
53 Eddie Rosario 2.7 60.9
59 Ben Revere 1.2 67.1
62 Aaron Hicks 0.9 66.7
64 Alex Gordon 0.4 60.9
91 Mike Trout -2.8 60.8
94 Carlos Gomez -2.9 57.5
97 Denard Span -3.1 58.8
99 Danny Santana -3.4 55.6
135 J.B. Shuck -9.1 57.4
141 Oswaldo Arcia -10.9 52.5
143 Miguel Sano -11.0 55.6
144 Danny Valencia -11.1 51.4
146 Drew Stubbs -11.9 46.2
154 Robbie Grossman -14.9 42.7
155 Daniel Nava -16.5 44.8
156 Mark Trumbo -28.8 32.8

 

 

[Vanimal46]

If Desmond Jennings rates that well, and was just released by the Reds, maybe he's a guy to look at acquiring for this team.

[Eris]

Jennings hit .200 last year and has a below average arm.  The analysis is on catch probability only and doesn't take into consideration any other factors.  

 

http://www.fangraphs.com/statss.aspx?playerid=1965&position=OF#fielding

 

I think the Twins player who is may profile similar to Jennings and Lagares is Zack Granite.  The difference is that there is hope that maybe Granite can hit for a decent enough average to be a positive contributor offensively.  Whereas Jennings has already demonstrated that he can not.  Some place on TD there is a thread on Granite based upon KATOH.  I can't find the thread, but the original article is

http://www.fangraphs.com/blogs/the-2017-all-katoh-team/