Who is the Oakland Athletics's best outfielder?
The A's have had one of the better outfields in the game the last few years. In 2012, the A's were 5th in total OF fWAR ( the fangraphs version of wins above replacement) behind the Angels and Cardinals. In 2013, they slipped to #11 behind the Red Sox and Angels. So far in 2014, the A's are #9 behind the (you guessed it) Angels and Rockies. They have been very good but not great.
Unlike most teams known to have a great OF, the A's do not employ a stars and scrubs approach. They employ a mix and match approach. With the addition of Craig Gentry, the A's are projected to be 5th in CF WAR, 6th in RF WAR and 7th in LF WAR in 2014. No one is a big star. No one is a scrub.
|CF||projected to be #5 in MLB
|RF||projected to be #6||Link|
|LF||projected to be # 7||Link|
Each player has different strengths that work together, and the A's use that approach to gain advantages beyond what will ever show up in WAR totals. Josh Reddick, Brandon Moss, Craig Gentry, Yoenis Cespedes and Coco Crisp all should do certain jobs at certain times. Sometimes that job is to be the DH so I include that in this breakdown.
But who's the A's best outfielder? I wanted to visualize each of their strengths today, so I created a set of graphs pitting each OF against each other in 5 different category's in 2012-2014. UZR/150, OF vs R, OF vs L, DH vs R and DH vs L. Of course there will be a poll.
I will mainly be using and combining two stats today. wRC+ and UZR/150.
In there essence, they do the same thing. Count runs created or count runs prevented. If you see that a player has an UZR/150 of ten (10), he saved ten more runs than average over a year. If you see a wRC+ of 110, he created 10 percent more runs (about 10 runs) than average over that year. These numbers have already been adjusted for league, park and position.
I will also be using a version of wRC+ splits that have been (base running enhanced)* today to help compare A's outfielders offensively. With speedy players like Josh Reddick, Coco Crisp and Craig Gentry in the Oakland outfield, there was way too much offensive information being missed by using strait wRC+ alone.
* base running enhanced. More on wRC+BsR at the end of this article. When you see "wO+B" on the graph , think wRC+.
For now, let see how individual A's outfielders compare to each other with creating runs. First off in the battle will be Josh Reddick vs Craig Gentry.
vs left-handed pitching...................................vs right-handed pitching
wRC+ and therefore wO+B, is set to 100 as league average. A wRC+ of 120 is 20% better than league average and one of 80 is 20% below league average.
Craig Gentry (green) vs Coco Crisp (red).
The first things we see from the four graphs above is Gentry has been far better vs left-handed pitchers ( LHP). The next thing we see is Gentry and Reddick are roughly equal vs right-handed pitchers (RHP), with Gentry better the last year or so. Last, Crisp is hands down best vs RHP.
Coco Crisp (R) vs Brandon Moss (G)
They are interchangeable at DH. Speed or power gets it done.
Cespedes (R) vs Gentry (G)
Pretty equal here as well. Cespedes has a slight edge vs LHP and also RHP (thanks to 2012) at creating runs over Gentry. Gentry has been better vs RHP lately.
The above graphs are strictly offense. So it shows us who should DH vs what hand. In Moss' case he will be at 1B much of the time, so just go down the line to the next best player.
Ok, now that we have compared all the OF by runs created, lets see how they do on defense. This is the last 2.2 years of OF UZR/150 on the left. I include the individual player vs player UZR/150 graphs at the end. One of those graphs is included now on the right.
Finally, lets combine ( base running enhanced wRC+) and UZR/150 from 2012-2014 to give an idea of who should play where in the OF. 100 is still average.
As far as the A's are concerned, a solo home run robbed is equal to a solo home run hit. To help keep it simple*; one (1) point of UZR/150 ( one run saved) equals one (1 ) point of wRC+ ( one percent created). (* this isn't saber metrically clean and exact, but me making it more exact will take 2000 more words and only get us 5% closer and still not exact.)
The two charts above show the total offensive and defensive contribution by handiness. This is what all the other charts are feeding into. These two are the main two of this article. Look it over again. There may be just a few surprises.
I now want to use this to get a general idea of each players total strengths vs each of the other players on this team. In the chart below, Players that are better vs RHP get more points than those who are best vs LHP, because they will face RHP 60-70% of the time on average. This isn't the most rigorous of point systems, but it ought to ball park it for our purposes. We need to pick our OF idol.
|Best at or equal DH vs L||.5 points|
|Best at or equal DH vs R||1.0 points
|Best OF vs L|| 1.0 points
|Best OF vs R
|Equal OF to Best player vs R
|Equal OF to Best player vs L
|Reddick is equal to Gentry at DH vs R||1.0|
|Reddick should OF vs Crisp vs L||1.0|
|Reddick should OF vs Moss vs L||1.0|
|Reddick should OF vs Cespedes vs R||1.5|
|Theme: Reddick should 4th OF||4.5 points|
|Cespedes should OF/DH vs Reddick vs L||1.5|
|Cespedes should DH Reddick vs R||1.0|
|Cespedes should OF/DH vs Moss vs L||1.5|
|Cespedes should DH vs Gentry L and R||1.5|
|Cespedes should OF/DH vs Crisp vs L||1.5|
|Theme: Cespedes should OF vs L and should DH vs L & R||7 points|
Coco Crisp best spots:
|Crisp/Gentry OF are equal vs R||1.5|
|Crisp should DH vs Gentry vs R||1.0|
|Crisp should OF vs Moss vs L||1.0|
|Crisp/ Moss are equal at DH vs L||0.5|
|Crisp should OF/ DH vs Cespedes vs R||2.5|
|Crisp should OF/ DH before Reddick vs R||2.5|
|Crisp should DH before Reddick vs L||0.5|
|Theme: Crisp should OF/ DH vs ( L & R )||9.5 points|
|Moss should OF vs Crisp vs R||1.5|
|Moss/Crisp are equal at DH vs L||0.5|
|Moss should OF vs Reddick vs R||1.5|
|Moss should DH vs Reddick vs L and R||1.5|
|Moss should OF/DH vs Cespedes vs R||2.5|
|Moss and Gentry are equal in OF vs R||1.5|
|Moss should DH vs Gentry vs R||1.0|
|Theme: Moss should OF/DH vs R||10 points|
Gentry should OF/DH before Crisp vs L
|Gentry/ Crisp OF are equal vs R||1.5|
|Gentry should OF/DH vs Reddick vs L||1.5|
|Gentry should OF vs Reddick vs R||1.5|
|Gentry/ Reddick are equal at DH vs R||1|
|Gentry should OF vs Cespedes vs L||1|
|Gentry should OF vs Cespedes vs R||1.5|
|Gentry should OF/DH vs Moss vs L||1.5|
|Gentry / Moss are equal in OF vs R||1.5|
|Theme: Gentry should OF/DH vs L, AND "equal OF vs R"||12.5 points|
Free Craig Gentry !
Regardless of what pitcher is on the mound, a case can be made for Craig Gentry to start in the OF. All things considered, a case can be made for Craig Gentry to be the best OF on this team.
Moss is also surprisingly valuable. A case could be made that with a better 1B, Moss still plays most days in the OF. Move Moss to 1st as usual, and it may become apparent Gentry wins today's OF idol. It's easier to see it if reversed. There are only three obvious places to not play Gentry. And they are all at DH.
1. Cespedes should DH vs Gentry L and R
2. Crisp should DH vs Gentry vs R
3. Moss should DH vs Gentry vs R
What do you think Nation, has Gentry done enough? Who is your choice for "the best" ( not just your favorite) OF on the A's?
Edit: I have added WPA (win probability added) and "clutch" graphs covering 2012-2014. Cespedes leads both. On the whole team. I also added 2012-2014 playoffs wRC+ in the reader comments.
The two graphs above show Cespedes is who you want up at the critical time.
- Extra graphs and notes I used to help me with this article.
- A more detailed explanation of base running modified wRC+.
- Billy Burns Bonus.
Coco (red) vs Gentry (Green) .....2012-2014 UZR/150....Reddick (red) vs Gentry (green)
Cespedes (red) vs Gentry ( green).....................................Cespedes (red) vs Crisp ( red)
Coco vs Moss...........................................................................................................Final
2012-14 WAR per plate appearance * 100......................How will Burns compare?
Above: Slow down Craig! Craig being a late inning replacement / pinch runner helps him too much here so I never used this chart. However, even taking that into account, Gentry would lead by a large margin.
Assumption: Billy Burns wob modifier will be 10, his UZR/150 will be 10 and he will have around an 90 wRC+ vs both L and R. This would make Burns most valuable to the A's as an OF vs LHP in place of Crisp/Moss/Reddick. He could also hold his own as an OF vs Cepedes and Reddick vs RHP at times.
*wRC+BsR. I am calling this Beta version "weighted offense including all base-running"
Or (wO+B) or "wob" and it is still a work in progress. As one quick example, Josh Reddick's 2013 wRC+ was 92. Adding his wob enhancement raised it (4) points to 96. I will explain the exact formula at the end.
wO+B (weighted offense including all base-running) or wO+B or "wob" is a modifier equation put together to capture not only the stolen base/caught stealing aspect of a player that wRC+ misses, but also any extra bases (score on a sac fly, 1st to 3rd,...) that a player provides that help create runs.
When wRC+ was first created, it included SB/CS information, but it has since been modified to measure only offense created with the bat. It has never captured other runs created by speed.
Wob is used the exact same way as wRC+ and is based on the exact wRC+ number. In a speedy players case, the base running enhancement can add up to 10 points of wRC+ to a players original wRC+ number. A plodder will be penalized. Moss had a negative (-3) modifier in 2013.
the equation is this:
500 / PA * BsR = wRC+ modifier
example: Craig Gentry. 2013.
BsR 5.8 wRC+ 108 P.A. 287
wRC+ vs L is 119, wRC+ vs R is 95
(500/287) * 5.8 = +10
+10 is Gentry's modifier
108+10 = wRC+BsR or wob of 118
95+10 = wob of 105 vs RHP
119+10 = wob of 129 vs LHP
To find a players BsR and wRC+, go to fangrapsh.com and look on the right side of a players stats. The reason wob is a work in progress is that wRC+ is a rate stat and BsR is a counting stat. It's a bridge. The 500/pa gets us ~90% of the way to accurately capturing this BsR value. The other 10% is the really hard part.