Best outfield arms of 2008
There is something about an outfielder with a cannon arm that captures
our imagination, isn’t there? I guess it’s not too surprising: Plays
involving a crucial throw from the outfield are among the most
important and exciting in the game. I would guess that a runner
gunned down at the plate is one of the most satisfying plays of all
for the defensive side, and their fans. Another thing about throwing
arms: Anybody can see who has a good arm or not. Well, that’s
not quite true, in the sense that strong arm doesn’t necessarily
translate into an effective one. But, still, we all have a
general idea of who is patrolling the outfield with a cannon and who is
trying to control the running game with a pop-gun.
Evaluating outfield arms is also satisfying for the baseball
analyst. It’s a straightforward and clean analysis — I think
Bill James once wrote that all the information we need to evaluate
outfielder throwing is in the recorded data, we just have to pick out
the right pieces. On the other hand, the analysis is not so simple
that you can get the results on your favorite mainstream media web
site. I believe the Hardball Times is still the only place where you
can get comprehensive outfield arm ratings.
The method
This is the fourth year I’ve written about outfield arms and the basic
method hasn’t changed. Here’s a brief summary about how I go about
measuring the value of outfield arms. (If you’re already familiar with
the method or don’t care about the details, skip down to the next section.)
Using play-by-play data from Retrosheet, I
consider five different situations
when a throw from the outfield is
important:
- Single with runner on first base (second base unoccupied).
- Double with runner on first base.
- Single with runner on second base.
- Fly out with runner on third base, fewer than two outs.
- Fly out with runner on second base, fewer than two outs (third base unoccupied).
For those plays, I add up how often the runner (any runner, actually,
including the batter) is thrown out
(I call that a “kill”) or else how often the runner is “held,” i.e., prevented from
taking an extra base (“hold”). I then compare a players number of
“kills” and “holds” with the number that a league-average outfielder
would have gotten, given the same number (and types) of opportunities.
Along the way, I define Kill+ and Hold+ values, which are the rates of
obtaining kills or holds relative to league average (normalized to 100, like OPS+).
Last year I added a new twist: I included in the analysis all assists
that did not fall into one of the five standard categories. Most of
these extra assists involve throwing out a batter trying to stretch
a single or doubling up a baserunner after catching a fly ball. It’s
not easy to define how many opportunities a player had for these extra
assists, so I look at the number per defensive inning and compare to
league average.
Finally, I evaluate the value in runs of kills, holds and extra assists in each of
these situations, which allows to me assign a run value (above or
below average) to all outfielders.
Park effects or “Manny’s arm is above average? Really?”
This year I have finally found the time to sit down and figure out
park effects for outfield throwing. That there might be such effects
is not hard to imagine — the left fielder in Fenway Park will
generally play shallower than in other parks and therefore he has a
shorter throw back to the infield. The Green Monster itself can also
have an effect, posing problems for both the fielder and the
runner. And large parks, like Coors Field, can make throwing tougher,
presumably. Artificial turf may also help outfield throwing, since
balls will get to the outfielder more quickly and they will get truer bounces on
their throws. (Of course, outfielders must play deeper on turf, so
it’s not obvious which effect will win out.)
To measure park effects for throwing, I used the same technique that
is generally used for park effects: for each team, I compared the throwing stats
for the team and its opponents in home and away games. Any
difference is ascribed to a park effect. Actually, I did this for
each of the three outfield positions in each park. I used multi-year
data: For the oldest parks, I used data from 1992 through 2008. For
newer parks, I used all the years in existence for that park.
Finally, I regress to the mean to take into account finite sample
size.
The correction I obtain is given in terms of Runs per 200 opps
(roughly a full season). I found seven fields where fielders are
helped or hurt by more than two runs per season:
+------+-----+------------+ | Team | Pos | Correction | +------+-----+------------+ | BOS | LF | 3.9 | | BOS | CF | 2.6 | | TEX | RF | -2.1 | | COL | LF | -2.1 | | COL | CF | -2.1 | | KCA | LF | -2.7 | | KCA | RF | -2.8 | +------+-----+------------+
These values represent the number of runs that an average fielder
would save playing half his games in that particular park. In other
words, we need to subtract these values from a player’s run value to
correct for park.
So, the Green Monster does help
the Red Sox left fielder (and center fielder, too). I’m not surprised
to see Coors hurting outfield throwing given its size, but I don’t
know why outfield throwing is suppressed in the corners
in Texas and Kansas City.
In any case, in what follows, I’ve applied regressed park corrections,
appropriately scaled to opportunities, for all
players.
Okay, now that we’ve got the techie stuff out of the way, who had the
best arm in 2008?
Right fielders
Hunter Pence, that’s who, at least among right fielders, who
traditionally have the strongest arms. The Houston right fielder was
nearly eight runs better than average, tying him with the Orioles’
Nick Markakis. Pence had a handful fewer opportunities, though, so
his rate stat (which I call Runs200, runs above average per 200 opps)
was a bit higher.
Here’s Pence’s performance broken down by situation:
 |
| Hunter Pence , best RF arm of 2008.(Icon/SMI) |
Pence, Hunter HOU RF
Situation | Opps | H K | Hexp Kexp
S-1B | 53 | 32 2 | 31.4 0.7
S-2B | 40 | 17 6 | 13.7 2.9
D-1B | 24 | 13 2 | 14.6 1.1
OF-3B | 14 | 6 0 | 3.2 0.8
OF-2B | 37 | 24 1 | 22.0 1.0
Overall | 168 | 92 11 | 84.8 6.4
Opps - opportunities
H - holds: runner did not take the extra base
K - kills: number of baserunners thrown out
Hexp - expected holds based on league-average arm
Kexp - expected kills based on league-average arm
Here’s how to read this table: focus on the second row of numbers. The
situation is “S-2B” meaning “single, runner on second base”. The
throw is generally going to be at home to nail the runner attempting to
score. In 40 such plays, Pence nailed six runners (the column labeled
“K” for “kill”) while only half that many kills would be expected from
an average right fielder (“Kexp” column). Seventeen times the runner
held third base, while the average number was fewer than 14.
Overall, Pence threw out almost five more runners than expected and
“held” about eight extra runners. Pence chose the right situations to
throw out runners — all those kills at home plate boosted his
run total significantly. In any case, I’ve got a new nickname for the
big Houston right fielder with the high-powered arm: “Big Game” Hunter Pence.
The following table shows the results for 25 right fielders in 2008
(the top 25 in throwing opportunities), ranked by Runs200.
+--------------------+---------+------+-------+-------+-------+---------+ | Name | Team | Opps | Kill+ | Hold+ | Runs | Runs200 | +--------------------+---------+------+-------+-------+-------+---------+ | Pence_Hunter | HOU | 168 | 172 | 112 | 7.7 | 9.2 | | Ludwick_Ryan | SLN | 129 | 126 | 112 | 5.6 | 8.7 | | Markakis_Nick | BAL | 198 | 136 | 112 | 7.7 | 7.8 | | Guerrero_Vladimir | ANA | 99 | 125 | 101 | 2.4 | 4.8 | | Church_Ryan | NYN | 110 | 92 | 105 | 2.5 | 4.5 | | Nady_Xavier | NYA/PIT | 122 | 183 | 102 | 2.7 | 4.4 | | Griffey_Ken | CIN | 109 | 151 | 102 | 2.0 | 3.7 | | Suzuki_Ichiro | SEA | 125 | 103 | 116 | 1.9 | 3.0 | | Francoeur_Jeff | ATL | 153 | 170 | 101 | 2.0 | 2.6 | | Gutierrez_Franklin | CLE | 115 | 104 | 108 | 0.9 | 1.6 | | Ethier_Andre | LAN | 109 | 134 | 92 | 0.8 | 1.5 | | Fukudome_Kosuke | CHN | 129 | 80 | 106 | 0.5 | 0.8 | | Dukes_Elijah | WAS | 96 | 135 | 84 | 0.2 | 0.4 | | Ordonez_Magglio | DET | 165 | 119 | 98 | 0.0 | 0.0 | | Winn_Randy | SFN | 145 | 76 | 98 | -0.3 | -0.4 | | Hawpe_Brad | COL | 180 | 99 | 91 | -1.0 | -1.1 | | Hart_Corey | MIL | 150 | 67 | 100 | -1.5 | -2.0 | | Abreu_Bob | NYA | 171 | 104 | 99 | -3.5 | -4.1 | | Upton_Justin | ARI | 103 | 89 | 91 | -2.1 | -4.1 | | Hermida_Jeremy | FLO | 153 | 70 | 88 | -3.9 | -5.1 | | Drew_J.D. | BOS | 115 | 75 | 82 | -3.4 | -5.9 | | Kearns_Austin | WAS | 111 | 23 | 97 | -3.5 | -6.3 | | Dye_Jermaine | CHA | 157 | 66 | 88 | -5.8 | -7.4 | | Teahen_Mark | KCA | 103 | 24 | 79 | -4.3 | -8.3 | | Giles_Brian | SDN | 152 | 33 | 81 | -10.3 | -13.6 | +--------------------+---------+------+-------+-------+-------+---------+ Kill+ - rate of throwing out baserunners, compared to league average Hold+ - rate of holding baserunners Runs - runs saved above average Runs200 - Runs per 200 opportunities (roughly a full season)
It’s interesting to find Pence followed by Redbird Ryan Ludwick and the
Orioles’ Nick Markakis in the above table. Ludwick, in part-time work over the last
few years, has shown a below-average arm, but he came through big in
2008. Markakis, who really is developing into a fine all-around
player, also improved his throwing significantly this year. Vlad has
never been at the top of my rankings, despite his rep, so I’m glad to
see him pretty high this time around. Andre Ethier
continues
the bewildering pattern of a high
Kill+ paired with a low Hold+.
I’m puzzled by Jermain Dye — I
mean, you look at the guy and he just looks like he should have
a strong arm. And, in fact, he did have a good arm, he just doesn’t
anymore. Since 2002, Dye has been below average every year except
one, 2005 when he rated +1 run. A noodle arm seems to be the least of
Brian Giles’ worries at the point, but I will note that his -10.3 runs
is the second worst performance for a right fielder in the last 50
years. Owie.
 |
| Matt Kemp, best CF arm (probably best OF arm) of 2008. (Icon/SMI) |
A caveat on Kill+ and Hold+: These measures should be used as a rough
guide, but they are not super-accurate for two reasons: 1) they don’t
include the park effects and 2) they do not take in the extra
assists. That’s why you can’t infer Runs from Kill+ and Hold+.
Center fielders
The best center field arm in 2008, on a per-opportunity basis, belongs
to 23-year-old Matt Kemp who patrols the middle garden at Chavez
Ravine. Here’s how Kemp did it:
Kemp, Matt LAN CF
Situation | Opps | H K | Hexp Kexp
S-1B | 39 | 25 1 | 27.1 0.4
S-2B | 29 | 9 0 | 5.7 1.4
D-1B | 9 | 4 1 | 2.1 0.5
OF-3B | 11 | 4 1 | 1.6 0.3
OF-2B | 13 | 10 1 | 8.5 0.2
Overall | 101 | 52 4 | 45.1 2.8
Kemp was better in both Holds and Kills in nearly each of the five
situations. Additionally, Kemp had six extra assists, which boosted
his run total significantly. All told, Kemp was 5.7 runs above
average in a partial season (101 opportunities). He also saved 3.6
runs in 68 opps in right field. Excellent work.
Here are the other center fielders:
+-------------------+------+------+-------+-------+------+---------+ | Name | Team | Opps | Kill+ | Hold+ | Runs | Runs200 | +-------------------+------+------+-------+-------+------+---------+ | Kemp_Matt | LAN | 101 | 142 | 117 | 5.7 | 11.3 | | Upton_B.J. | TBA | 168 | 190 | 98 | 6.6 | 7.9 | | Victorino_Shane | PHI | 149 | 111 | 123 | 5.0 | 6.7 | | Suzuki_Ichiro | SEA | 93 | 147 | 113 | 2.9 | 6.2 | | Kotsay_Mark | ATL | 101 | 68 | 123 | 2.5 | 5.0 | | Ankiel_Rick | SLN | 108 | 94 | 126 | 2.3 | 4.3 | | Jones_Adam | BAL | 169 | 65 | 120 | 3.1 | 3.7 | | Taveras_Willy | COL | 140 | 98 | 107 | 2.6 | 3.7 | | Rowand_Aaron | SFN | 200 | 105 | 118 | 3.6 | 3.6 | | Cabrera_Melky | NYA | 120 | 118 | 108 | 1.8 | 3.0 | | Gomez_Carlos | MIN | 193 | 132 | 103 | 2.3 | 2.4 | | Ross_Cody | FLO | 102 | 150 | 94 | 1.1 | 2.2 | | Bourn_Michael | HOU | 140 | 170 | 94 | 1.3 | 1.9 | | Gathright_Joey | KCA | 100 | 114 | 96 | 0.7 | 1.4 | | Hamilton_Josh | TEX | 136 | 45 | 109 | 0.5 | 0.7 | | Beltran_Carlos | NYN | 189 | 107 | 106 | 0.0 | 0.0 | | McLouth_Nate | PIT | 196 | 72 | 104 | 0.0 | 0.0 | | Patterson_Corey | CIN | 96 | 84 | 103 | -0.2 | -0.4 | | Cameron_Mike | MIL | 119 | 66 | 97 | -0.4 | -0.7 | | Granderson_Curtis | DET | 179 | 60 | 97 | -1.6 | -1.8 | | Wells_Vernon | TOR | 105 | 130 | 80 | -1.0 | -1.9 | | Young_Chris | ARI | 183 | 59 | 100 | -3.0 | -3.3 | | Hunter_Torii | ANA | 152 | 43 | 95 | -4.1 | -5.4 | | Milledge_Lastings | WAS | 152 | 22 | 98 | -4.4 | -5.8 | | Sizemore_Grady | CLE | 164 | 21 | 96 | -6.7 | -8.2 | +-------------------+------+------+-------+-------+------+---------+
Most of the other top names here are familiar: I wrote about
B.J. Upton and Shane Victorino last year and they kept up the good
work in 2008. Ichiro was fine in a half-season in center (and above
average in right) as was
veteran Mark Kotsay in his Atlanta stint.
 |
| Alfonso Soriano, best LF arm of 2008. (Icon/SMI) |
Curiously, Carlos Beltran and
Mike Cameron, who are considered to possess strong arms, showed as
merely average. Torii Hunter’s arm continues to decline (-0.6,
-3.6, -4.1 runs the last three years). And if you’re the kind of
person who can’t stand perfection in others, well, you’ll be pleased
to know that Grady Sizemore, who seems to be good at everything, is
quite noodle-armed. This was his worst season to date, but he’s been
below average every year of his career.
Left fielders
Alfonso Soriano has done it again. For the third straight season,
Soriano takes top honors in throwing for a left fielder. As in past
years, Soriano excelled in nailing baserunners.
Soriano, Alfonso CHN LF
Situation | Opps | H K | Hexp Kexp
S-1B | 30 | 28 0 | 25.1 0.3
S-2B | 19 | 7 1 | 6.0 1.4
D-1B | 14 | 6 2 | 8.5 0.5
OF-3B | 13 | 1 4 | 2.5 0.6
OF-2B | 17 | 14 0 | 15.5 0.4
Overall | 93 | 56 7 | 57.6 3.2
Soriano was hurt for part of the 2008 season and only accrued about a
half-season’s worth of opportunities, but he made the most of them.
He impressively threw out four runners trying to score
from third base on a fly ball. Curiously, Soriano’s hold rate is only
about average, despite his proficiency in gunning down opponents since
he became an outfielder.
You’d think that runners would start wising up.
Here’s how other left fielders around the leagues fared in controlling
the running game (or not) in 2008:
+-----------------+---------+------+-------+-------+------+---------+ | Name | Team | Opps | Kill+ | Hold+ | Runs | Runs200 | +-----------------+---------+------+-------+-------+------+---------+ | Soriano_Alfonso | CHN | 93 | 218 | 101 | 3.7 | 8.0 | | Lewis_Fred | SFN | 118 | 212 | 101 | 4.4 | 7.5 | | Boggs_Brandon | TEX | 89 | 219 | 105 | 2.6 | 5.8 | | Jackson_Conor | ARI | 77 | 146 | 107 | 2.1 | 5.5 | | Ramirez_Manny | LAN/BOS | 122 | 158 | 108 | 2.5 | 4.1 | | Young_Delmon | MIN | 185 | 111 | 102 | 3.6 | 3.9 | | Burrell_Pat | PHI | 155 | 146 | 106 | 2.9 | 3.7 | | Francisco_Ben | CLE | 103 | 97 | 102 | 1.6 | 3.1 | | Quentin_Carlos | CHA | 139 | 69 | 108 | 2.0 | 2.9 | | Ibanez_Raul | SEA | 205 | 112 | 100 | 2.8 | 2.7 | | Holliday_Matt | COL | 195 | 120 | 101 | 2.5 | 2.6 | | Anderson_Garret | ANA | 85 | 194 | 85 | 0.9 | 2.1 | | Braun_Ryan | MIL | 155 | 113 | 98 | 0.9 | 1.2 | | Willingham_Josh | FLO | 110 | 94 | 104 | -0.1 | -0.2 | | Crawford_Carl | TBA | 100 | 53 | 109 | -1.5 | -3.0 | | Scott_Luke | BAL | 144 | 59 | 100 | -2.5 | -3.5 | | Cust_Jack | OAK | 76 | 137 | 86 | -1.4 | -3.7 | | Harris_Willie | WAS | 76 | 42 | 92 | -1.4 | -3.7 | | Damon_Johnny | NYA | 78 | 41 | 102 | -2.2 | -5.6 | | Dunn_Adam | ARI/CIN | 139 | 42 | 93 | -4.2 | -6.0 | | Headley_Chase | SDN | 90 | 29 | 97 | -2.7 | -6.0 | | Bay_Jason | BOS/PIT | 210 | 68 | 92 | -8.1 | -7.7 | | Lee_Carlos | HOU | 112 | 56 | 93 | -4.7 | -8.4 | | Pierre_Juan | LAN | 84 | 0 | 96 | -3.6 | -8.6 | | Gonzalez_Luis | FLO | 76 | 0 | 105 | -3.6 | -9.5 | +-----------------+---------+------+-------+-------+------+---------+
I must confess I have not had a chance to see Fred Lewis or Brandon
Boggs in action much, but I’m going to be paying attention this
coming season. Manny Ramirez showed an above-average arm this year,
even taking into account the help he got from playing most of the
season in Fenway Park’s left field. Now that we have the park
corrections figured out, we can say that in his Boston years Manny’s
arm was just below league average: I have him at -3 runs in almost 1000
opportunities.
Manny’s replacement in Boston, Jason Bay, is a terrible thrower, at
least he was last season. Curiously, almost all of those negative 8
runs were “earned” in Pittsburgh, Bay was slightly above average in
Boston (with 61 opps).
For Juan Pierre there was good news and bad
news in 2008. The good: he was nowhere to be found at the bottom of
the list of center field arms, a place he had inhabited for several
years running. The bad: Juan has now established residency at the
bottom of the left fielders’ list. I’m not counting Luis Gonzalez,
who, as far as I can tell, has been playing with a paralyzed throwing
arm for the last several years. Note the Kill+ of zero for these two.
Visual aids
A good way to grasp quickly who has an effective outfield arm is to
view the plot on the right. Each point in the plot represents a
single player, with the color denoting the position played, red for
right field, etc. The horizontal position of the point gives the
player’s hold rate (Hold+) — recall that above 100 is good and
below 100 is bad. The vertical axis represents the kill rate, where,
for example, a kill rate of 150 means the player threw out 50% more
baserunners than average. The dotted gray lines show the average
values (100) of hold and kill rates.
So, players that are above average in both Hold+ and Kill+ appear in
the northeast quadrant of the plot and indeed here we find “Big Game”
Hunter Pence, Matt Kemp, Shane Victorino, etc. Soriano, as I mentioned,
has great Kill+ but average Hold+ and you can see him near the top of
the plot.
The worst throwers are in the southwest quadrant (Giles, Teahen,
Pierre). Players in southeast quadrant are good at holding runners,
but not throwing them out. Mark Kotsay and Josh Hamilton are in this
group.
That raises the question of what makes a fielder good at
holding runners? Since it’s the runner who decides to run or not, the
fielder doesn’t have complete control over his hold rate. Of course,
if a player has a strong reputation, runners will be more cautious,
leading to high hold rates. Looking at the plot, though, you can see
that there isn’t a strong correlation between hold and kill rates.
Clearly, a fielder can help his hold rate if he generally gets to the
ball early and efficiently puts himself into throwing position,
perhaps discouraging runners from taking the extra base.
Dwight Evans, the strong-armed right fielder of the Boston Red Sox
back in the ’70s and ’80s said that he made it a point to make several
serious warm-up throws
before each game in clear view of the opposing team’s dugout: Dewey
wanted to make sure potential baserunners got a good look at his
throwing prowess.
Final word
A few things to keep in mind when evaluating outfield thowing. First, there is quite a bit of
year-to-year variation in these results, since they are based on a rather low number of opportunities. Do I know that Matt Kemp or Hunter Pence are among
the best-throwing outfielders in the game? I do not, although I do know that in 2008 they were among the best. (I’m pretty sure Soriano is exceptional at controlling
the running game, since he has been great for several years.)
Another thing to note is that for the majority of outfielders throwing is a relatively minor contribution to their total value. In 2008, only thirteen players saved or cost
their teams more than five runs with their throwing arm. Obviously, offensive and non-throwing defensive contributions are more important. On the other hand, according to the economics of the game today, five runs cost over $2 million on the free agent market.
And at the extremes, there can be wide swings in value: replacing Brian Giles’ arm with Matt Kemp’s would result in two additional team wins. So, let’s not overestimate the importance of outfield throwing, but we
shouldn’t underestimate it either, especially when it comes to evaluating the guys with particularly good or bad throwing arms.
Well, that about wraps it up for this year’s annual “Best arms”
review. We will be updating the arms stat on the site in the near
future, so you’ll have access to all players.