A howitzer. A cannon. A rifle. A gun. There’s a reason we use firearms terminology when talking about the strength of an outfielder’s arm. We recently talked about Roberto Clemente’s throwing as a signature skill, and when a player is described as having a “right fielder’s arm,” we immediately think of a Clemente, Barfield, or Evans. Of pinpoint lasers vaporizing yet another foolish runner.
However, assists, baserunner kills in the parlance, are only part of the story. Sometimes they are downright misleading. And they mean different things to each of the three outfield positions. Today let’s take a quick look at throwing arms and then speculate on what they might mean for a few HoME bubble candidates.
Reading right to left
Baserunner kills by outfield position are revealing in their confusing mix of information. We only have play-by-play data on by-position outfield arms since 1953, so here’s the top 5 at each outfield station since then:
RANK RIGHT CENTER LEFT ----------------------------------------------------- 1 Clemente (254) Griffey (141) Yastrzemski (177) 2 Aaron (178) Lofton (138) Ba. Bonds (158) 3 Staub (159) W. Davis (134) J. Rice (132) 4 Callison (159) Finley (127) R. Henderson (124) 5 Evans (155) Otis (118) Raines (123)
Other than Rusty Staub, right field looks pretty solid. Full of great throwers. In center, we all remember Junior Griffey’s arm, some of these others don’t really come to mind as great arms. On the other hand, these guys were all burners who perhaps could have used their speed to field the ball quickly and keep opposing baserunners from advancing. In left field, neither Rickey nor Raines were ever thought to have good arms, while Yaz and Jim Ed both “coincidentally” patrolled left field in front of the Green Monster.
In other words, as we move from right to left, as the throwing requirements become lesser and lesser, things get more jumbled.
A call to arms
BBREF’s player pages have a “fielding” sub-page. Here we can see what they calculate an outfielder’s arm to be worth. This helpful stat, unhelpfully called Rof, tell us how many runs above or below average a player’s outfield throwing was worth. It combines the rate of assists with the rate of holds, which are how often the outfielder managed to prevent advancement merely by the threat of his arm.
So I took BBREF’s list of the top 27 career totals in assists at each outfield position. For each player, I noted the number of assists, his total career Rof, his career Rof at the position in question (pRof), the number of times he led the league in assists at his position, and the number of times he finished among the top 5 in assists at this position.
Here’s the averages of the 27 fellows at each position. Because the Green Monster dramatically increases the chances for an assist or a hold, I’ve included one extra column, that removes the Fenway guys from the left field average.
RIGHT CENTER LEFT NO FENS ----------------------------------- A 128.0 102.9 100.6 92.7 #1 2.5 2.1 2.2 2.0 T5 5.5 5.2 5.4 5.3 Rof 26.1 12.8 1.9 -0.9 pRof 24.7 12.9 3.0 0.4
We should expect the sheer number of assists to decline as we move from right field to left field. The averages for finishing first or within the top five stay exactly the same. But look at the differences in the arm ratings! Between each position the average outfielder loses about a dozen runs. And once you pull Fenway out of the left fielders, youch! Fenway’s Green Monster gobbles up extra bases.
Next I ran correlations against assists, Rof, and positional Rof. Correlations tell us what relationship exists between two sets of data. When the correlation is a positive number, they rise or fall simultaneously. When it’s negative, they are inversely related. When it’s close to zero, there is no relationship. And the closer you get to 1 or -1, the more significant the finding is. Reminding ourselves that these are very small samples of the greater population of outfielders…here we go.
Correlation against career assists RIGHT CENTER LEFT NO FENS ----------------------------------- #1 0.65 0.37 0.49 0.47 T5 0.51 0.56 0.38 -0.08 Rof 0.50 0.49 0.43 0.19 pRof 0.59 0.51 0.49 0.27 Correlation against career Rof RIGHT CENTER LEFT NO FENS ----------------------------------- #1 0.53 0.41 0.66 0.58 T5 0.13 0.09 -0.20 -0.25 pRof 0.96 0.95 0.97 0.97 Correlation against career pRof RIGHT CENTER LEFT NO FENS ----------------------------------- #1 0.60 0.31 0.69 0.60 T5 0.18 0.22 -0.17 -0.23
Let’s put this into bullet points. In our sample:
- Leading the league in assists at your position is far more strongly correlated with effective throwing than finishing in the top 5, which has little or no relationship.
- High career assists totals are moderately weakly correlated with runs saved by throwing
- There’s not much to the relationship between left field throwing and assists, unless you play at Fenway.
Obviously, small sample size caveats apply, but at the HoME we’re only interested in the top 200-odd guys anyway, and that’s about sixty outfielders.
Armed and dangerous?
What good is this information? For pre-1953 outfielders, it gives us some insight. If you look at Harry Hooper who is sixth all time in outfield assists. He led the league in assists four times and placed in the top five nine times, you can find comparable post-1953 right fielders and see what it meant to their throwing performance. You might be able to make a mental adjustment to Hooper’s fielding numbers if you think they are low. DRA’s arm ratings are definitely low because they don’t account for holding runners, only for kills. So you could ratchet up its arm component.
If you were smarter than me, you’d find an effective mathematical way to do it.
You might also figure out the riddle of Enos Slaughter. He is frequently cited in contemporary and latter-day accounts for his arm strength, and he led the NL in assists by a right fielder twice and placed in the top five in right and left field five times. But his post-1952 throwing numbers are poor (-8), reflecting drop offs in assists, speed indicators, and playing time after his 1953, age-37 season. Did he simply get old and slow, sending his throwing off a cliff? Or was he never that great to begin with?
How about Sam Rice who led AL right fielders in assists five times? Was he the Johnny Callison of his day (+37 Rof) or the Rusty Staub of his day? (+4 Rof). Both led their leagues five times as well.
Or “Indian” Bob Johnson. He led his leagues in assists six times from left field. Should he get a little more credit than the 39 runs DRA’s arm rating gives him? After all, Clemente led the league six times from right field. (Answer: probably not, Barry Bonds also led his league six times and finished with about the same number of career Rof as Johnson has DRA arm runs.)
Anyway, that’s me and Miller’s job to figure this stuff out for ourselves and arrive at our own decisions. Most guys wouldn’t benefit enough from a deeper analysis to make it useful to do. But then again, there’s only a few guys close enough to in/out line whose cases could hinge on judgments like this.