In last week's post on the Myth of the RBI Guy, I used a crude method inspired by  Jonah Keri's piece in Baseball Between the Numbers to generate groups of individual seasons by players since 1973 and in 2008 who were overrated by their gaudy RBI numbers. The point of the piece wasn't so much to "expose" individual overrated players, but rather to show that, using traditional linear weights, one can be a pretty terrible hitter and still accumulate big RBI  numbers. For example, in 2003, Tony Batista hit .235/.270/.393 and still had drove in 99 runs.

While I think last week's piece (and those like it) pretty much make the case, there may by some out there thinking that despite the base "three-slash" lines or linear weights numbers some of the guys had, that perhaps such approaches are overlooking the skill these players have with regard to situational hitting, that they have superior skills when runners are on-base. Is there a stat other than the "BA w/ RISP" chesnut that we can look at? Why, yes, yes there is...

If you haven't looked at Part One, I recommend doing so, since I'll be assuming some familiarity with those concepts in this post. Basically, to judge how much a a player's single season performance might be overrated or underrated by RBI, I divided his RBI by his batting runs above replacement. The more RBI per runs above replacement, the more the player's RBI overrate his offensive contribution. The results can be found in this spreadsheet (explained more fully in last week's post). This week, I want to look a a more contextual approach.

RE24 to the Rescue?

At the amazing FanGraphs, they have a stat named RE24. "RE24 is the difference in run expectancy (RE) between the start of the play and the end of the play," they say. Um, what?

Let's back up. Linear weights, such as those I used last week, measure the average runs a player creates (either absolute or above/below average) without considering the specific game context in wihch the player performed. Yes, they can be adjusted for the run environment of the season and park-adjusted, but what they don't take into account is how many players are on base when the event occurs. For traditional linear weights, a home run with 2 outs and no one on base counts the same as a home run with the bases loaded and one out. While certain methods of generating linear weights take into account the average run expectancy of each event over the years, the individual impact of such events are obviously different. How can we measure the differences between events and their impact?

There are 24 base-out states (hence the "24" in "RE24"): eight different combinations of baserunners (e.g., runner on first, bases empty, runners on second and third, etc.) multiplied by the three out states in which hitter might have that situation (no outs, 1 out, 2 outs). RE24 measures is the difference in Run Expectancy from the beginning of the play until the next play. Here's how it works in an RBI situation, quoting the Tango article linked above:

For example, with a man on 2B, and 0 outs, the RE for that situation (the start state) is 1.2 runs . If the batter hits a double, the RE for the end state is of course 1.2 runs. As well, a run scored. So, the run impact of this particular batting event is 1.0 runs. If you have a man on 1B with 1 out, the RE is 0.57 runs. A double-play brings us to the end of the inning, and an RE of 0 runs. The double-play in this case is worth -.57 runs.

The run expectancy for each situation can change by run environment, of course, just as in traditional linear weights, and I believe FanGraphs implementation of RE24 acccoutns for this. In any case, one can see how this might be relevant to "RBI Guys," as it is a stat which takes into account situational hitting by base/out state -- if there are players who have "special skills" with runners on base, RE24 would seem to be the sort of stat that would capture this in a way that traditional linear weights would not.

[Note that this is similar to the way in which the difference between traditional linear weights and another FanGraphs stat, WPA/LI, measures "Little Things" as I discuss in an earlier post. However, WPA and WPA/LI differ from RE24 in that they reflect the game situation (score, innings, etc.), while RE24 takes only the base/out state into account.]

Back to 2008

In Part One, I started by looking at some of the worst 90+ RBI seasons since 1972 before moving on to the worst of 2008. At the suggestion of Tom Tango and terpsfan101 at the Book Blog, I also posted the same rankings using absolute runs rather than runs above replacement (the "BRC" versions of the spreadsheets). While this isn't' a comprehensive statistical approach (nor was last week's), I want to take a similar approach here. Because RE24 isn't something I currently calculate myself, I'll use all FanGraphs statistics to generate my tables. This also means I can't do an extended study over 30 years. Well, I can, but FanGraphs only lets you export one year at a time at the omment... but that would take forever.

Let's start with 2008. I will simply use "absolute runs created" according to wOBA runs.

[The next few lines are very simple math, but if it bores you, fell free to skip it.]

Recall from last week's article that to get a player's runs created above/below average, one simply uses

((wOBA-lgwOBA)/wOBAScale)*PA

[wOBAScale is usually around 1.15]

This is what FanGraphs calls wRAA (I called it BRAA in my version, which is basically the same thing).

To get total runs created, by the player (what FanGraphs calls wRC), one simply adds in the league average runs per PA to the runs above/below average:

(((wOBA-lgwOBA)/wOBAScale)*PA) + (lgR/lgPA)

To we get a total runs created number for RE24, which is also an above/below average stat, then, we can simply do

wRC-wRAA+RE24

for each player. Let's call this 24RC.

Here is the spreadsheet I'll refer to for 2008. The first sheet is a replay of the reworked version of last week's info on 2008. We aren't as concerned with the rankings as the general group of players over- or underrated by RBI totals. I've only used those who had 90 RBI or more last season.The highlighted columns are the ones relevant for the sort order -- basically, they are sorted in descending order of the yellow column, in this case, RBI per wRC (wOBA Run Created). Mike Jacobs is in the "lead" here.

Now look at the second sheet. It focuses on 24RC, and is sorted by RBI per 24RC. It's basically the same group, but with Jose Guilen at the top.

The initial thought is that, while there are some ordering differences, it's the same group of guys -- there isn't much difference here. Keep in mind that just because a guy  is "overrated" by his RBI according to this stat, that doesn't mean he's was a bad hitter in 2008. Mike Jacobs was above average according to both wRAA and RE24. Ryan Howard is a good hitter. THe point is that compared to their raw production, just looking at their RBI would overstate their real offensive value. Indeed, the arbitrariness of the list of the guys at the top is revealing in itself says something. Good hitters like Ryan Howard, 22 runs above average according to wRAA and 38.8 according to RE24, are in the same general "rank" as below average hitters like Jose Guillen and Garrett Atkins.

You may also have noticed the "Situational" column to the right. Like my  attempt to quantify Little Things, this number attempts to get at if the player is underrated by traditional linear weights as opposed ot his RE24 number, to see if he added more value in situations with guys on base, etc. If you look at the numbers for Jose Guillen, Bengie Molina, and Mike Jacobs, you might get that idea, although the difference isn't big. However, notice that there are some guys up top whose "situational" numbers are negative -- Atkins, Ordonez. What's the deal?

Keep in mind how RE24 works -- aside from the run expectancy before and after the play, it also gives a run for each run driven in. Since these guys have a lot of RBI, their RE24 is actually being helped. And how did hitters like this get more RBI per run created than guys at the "bottom" of the list like Pujols, Berkman, Sizemore, etc. The old-fashioned way -- they were in the right spot in the batting order, as Tom Ruane shows (Tom Tango also directed me to this article on RBI by Ruane which covers similar ground to what I've covered here and comes to similar conclusions).The "situational" difference between RE24 and wRAA in this case doesn't show much more than the "RBI" guy being in the right place at the right time, and in some cases (Atkins, for example), he didn't even take advantage of that.

Final Thoughts

I realize that this post has been a bit more scattershot than last week's, but hopefully it gives an idea of how RE24 might be useful, even if it doesn't vindicate bad-hitting "run producers," which it shouldn't anyway. In any case, I want to leave you by looking at four of my "favorite" 90+ RBI seasons from last week (switching back to my own BRAA) and seeing if RE24 is friendlier to them:

I was a bit shocked when I saw the "situational" numbers, and Carter's 1990 RE24 numbers. And I may have misanaluzed the results above. Still, keep in mind that RE24 gives the player credit for the runs he drives in -- and in all four of these seasons, Carter and Batista were put in  position to drive in more runs than the average player. And note that only one of these seasons has an RE24 above average -- RE24 is an above/below average stat. So while further research on situational hitting is needed, even counting RBI in, RE24 shows that one can still be a far below average hitter and get to 100 RBI.

Thanks to Tom Tango and terpsfan101 for their comments and suggestions at The Book Blog, and to Sky Kalkman for suggestions and encouragement. I am, of course, solely responsible for the "quality" of the piece.