RISP for hitters and pitchers (October 13, 2003)
In the BP chat, the following was asked:
Derek Martin (Montreal): I've asked this question a million times to BP with no response. Why evaluate relief pitchers, in a large part, by their inherited runners scored, while eschewing batters success with RISP? Seems like a direct contradiction.
The response from Adam Katz (Sosa's agent?!?) ended with:
RBI is a team-dependent stat. Stranding inherited runners is not.
I'm with Derek Martin, 100% here. (Him being from Montreal is a bonus.)
See, the first question to ask is: what are you trying to measure? A player's ability (future) or a player's value (past)?
So, to the extent that a pitcher (or hitter) does have a certain skill with men on base, you do want to know how they perform with men on base. However, this skill is probably so elusive to establish that in the end you don't need to worry about this skill (for the hitter or pitcher). I will concede that perhaps there are some pitchers that may have this skill more apparent, perhaps because their mechanics/approach changes completely.
However, what ARP does is clearly a "value-added" stat, and from that perspective, how a player performs with men on base, whether the manager selected him to pitch with men on base, or whether a manager played the odds by putting such a player in the cleanup spot to maximize his chances, is important for both the hitter and pitcher.
And inherited runners is a team-dependent stat, as well as a manager-dependent stat.
***
I loved Gary's response to another question on BP: It's a perfect Valentine's gift if you're in a relationship you want to get out of.
--posted by TangoTiger at 03:47 PM EDT
Posted 7:06 p.m.,
October 13, 2003
(#1) -
bob mong
I agree with you 100%, Tango - I just read this and was abouts to email you asking you to post it for Primate Study - but you beat me to it :)
I don't really see how the BP Boyz can tout ARP endlessly when discussing relievers, on the one hand, and completely ignore situational hitting when discussing hitters (or, to complete the analogy, pinch hitters), on the other hand.
I can see disregarding RBI, since it is a counting stat - but if ARP is so great, why don't they compile and present AVG/OBP/SLG for all runners in all base-out states, and rank hitters based on their ability to hit with RISP, for example?
This came up a bunch when the Mariners traded Jeff Nelson for Armando Benitez; Jeff Nelson, despite a reasonably pretty ERA and peripherals at the time of the trade, had a gaudy ARP (it was negative, I believe) - so lots of people argued that the Mariners got a far better reliever, since the ARP totals were so different. I argued that no one had ever shown that ARP has any predictive ability, and that traditional methods of evaluating pitchers (ERA, K:BB, etc) could certainly be applied to relievers as well.
Posted 8:23 p.m.,
October 13, 2003
(#2) -
David Smyth
A batter is generally in there for all his ABs--RISP or bases empty. But some relievers are brought in primarily because there are RISP, and in a fairly close game. With such a more 'concentrated' situational usage pattern, it is entirely reasonable to examine the performance of relievers in those splits. I have always believed that it is more likely that there are clutch pitchers than that there are clutch hitters--simply because the pitcher is the initiator, and they are involved for several consecutive PAs (where any particular mental set is likely to persist).
So, I think (or assume that) there may be a 'qualitative' difference between pitchers and hitters with RISP. If such is not the case, it should easily show up in the stats, properly analyzed. That BP is apparently assuming such a difference means either that they have studied and confirmed such, or that they are simply following the conventional line of thinking.
Posted 8:39 p.m.,
October 13, 2003
(#3) -
Patriot
Remember too that a reliever gives up runs that aren't incorporated in his run total. Including some kind of inherited runners condition(or going further like ARP, although it's a lot further) is necessary. Of course, the whole concept of actual RA or ERA lines up better with a batting stat that includes situational performance.
On the whole, I agree with the general point of Derek et al. If you're going to trumpet a value added method for pitchers, you should do it for hitters too.
Posted 11:46 a.m.,
October 14, 2003
(#4) -
impatton
While there is ARP, which considers sintuations like inning,runner,out,etc. for pitchers, why isn't there any metric like that for hitter? Except Win Percentage Added in "Paths to Glory." However, I haven't seen anything like that provided regularly.
Posted 12:44 p.m.,
October 14, 2003
(#5) -
tangotiger
(homepage)
If you guys keep clamoring for it, I might decide to get off my lazy a-- and do it seriously.
You can check out the above for what I did a few months ago.
Posted 1:41 p.m.,
October 14, 2003
(#6) -
Anonymous
.
Posted 2:56 p.m.,
October 14, 2003
(#7) -
Alan Shank
It seems to me that the most important aspect of ARP is to divide the responsibility for runs allowed up in a rational way. I wrote a program many years ago that used the base-out table from "The Hidden Game of Baseball" to assign "runs" to pitchers based on the run-expectation value of the situation when they came in, the number of runs scored while they were in the game and the run-expectation value when they left. This was not to focus on RISP situations, but just to get an "ERA" that was more reflective of how well pitchers prevented runs than the official ERA.
Cheers,
Alan Shank
Posted 3:48 p.m.,
October 14, 2003
(#8) -
tangotiger
Alan, I agree that that ARP process is sound, and I whole-heartedly support it. I just happen to support it to the same extent that I would for hitters, too.
I would support it more for pitchers than hitters, but not much more, if research shows that pitcher's have more variability in their performance men on base v bases empty than hitters too.
Posted 5:19 p.m.,
October 14, 2003
(#9) -
studes
(homepage)
I guess I'm missing something. We assign runs to pichers when we use ERA, and I thought the point of ARP was to correct ERA for situations in which another pitcher impacts the stat. This particularly impacts relievers, I believe. Virtually all fans use ERA to assess pitchers, so this makes something like ARP more imperative for pitchers than hitters.
I don't believe we assign runs to batters in the same way. Yes, we do count runs and RBIs, but they are not primary stats for evaluating hitters (at least, not for educated fans). Speaking for myself, I pretty much ignore them. And most fans do look at other stats, such as BA, HRs, etc.
If you develop a stat that assigns run values to hitters, similar to ERA for pitchers, then I agree that you should correct for base/out/inning situation.
Posted 5:52 p.m.,
October 14, 2003
(#10) -
tangotiger
What happens with ARP is that if you happen to be brought in alot with men on base, and you happen (by luck or design) get yourself out of those jams (but not necessarily your own jams), you will get a very favorable ARP.
In essence, ARP allows you to leverage your talent on the mistakes of others. I whole-heartedly support this.
At the same time, Eddie Murray leveraging his talent by having alot more men on base to work with is just as valid.