Runs Batted In: Opportunity vs Execution

by Adam Peterson on Jul 3, 2009 10:15 AM EDT in By the Numbers Comment 31 comments

Conventional sabermetric wisdom says that the RBI is a "dinosaur" statistic, less important or telling than other, "better" measures such as OBP, OPS, wOBA, or others. This is primarily due to RBI opportunities being out of the batter’s control. Whether a batter comes to the plate with runners in scoring position or with the bases empty is a result of the batting order and the hitters in front of him. As a result, RBI are often dismissed by the community as shallow; a relatively poor indicator of a batter’s true value to a team as a run producer.

What if we could reassess the RBI as a more meaningful statistic by leveling the playing field? In other words, can we normalize each batter’s RBI opportunities to gain a more accurate measure of the batter’s true success when driving in runs? Using 2008 and 2009 (through 7/1) data, I have done just that. You will see a number of familiar names at both the top and the bottom of the leader boards…

Justin Morneau: Elite Run Producer, or lucky to be batting behind Joe Mauer?

Over the past three years, no one on the Minnesota Twins has driven in more runs than the 2006 MVP, Justin Morneau:

2006: 130 RBI

2007: 111 RBI

2008: 129 RBI

That’s a lot of RBIs. Buthow much of this is Morneau the elite run producer, and how much does he take advantage of hitting behind one of the truly elite OBP guys in Joe Mauer? Let’s take a look at the number of RBI opportunities Morneau has had relative to the rest of MLB.

"Expected" Runs Batted In

How do we normalize each batter’s RBI opportunities? First we must create a baseline. I collected data from the entire 2008 and partial 2009 (through 7/1) seasons, counting the total number of RBI for each inning situation (e.g., one out, runner on second base). I used this data to calculate the average number of RBI that one would "expect" a batter to drive in for a given situation. I call this "Expected RBI", or "eRBI". Not surprisingly, bases empty, zero or one out (0.028 eRBI) is the least RBI-friendly situation, and one out, bases loaded (0.766 eRBI) is the best situation for driving in runs. I then added up the eRBI for each player across all of his plate appearances during the season. Who were the leaders in expected RBI?

Rank	Player	eRBI
1	Justin Morneau	99.8
2	David Wright	98.8
3	Garrett Atkins	93.0
4	Ryan Howard	89.2
5	Josh Hamilton	88.6
6	Jeff Francoeur	88.6
7	Raul Ibanez	88.4
8	Carlos Beltran	88.4
9	James Loney	87.1
10	Carlos Delgado	86.5

That’s a familiar name to Twins fans at the top of the list. Justin Morneau had more RBI opportunities than any other batter in Major League Baseball in 2008. That at least partially explains the 129 RBI last year. If Morneau were simply a league average run producer, he would have finished the season with about 100 RBI. This is yet a nother measure of the value of Joe Mauer. A couple of the other names surprise me. Garrett Atkins 3^rd in the Majors? Jeff Francoeur? Three New York Mets in the Top 10? Whatever the reasons, Morneau had the most RBI-friendly batting order spot in the Majors last year.

Opportunities vs. Execution: "RBI Plus-Minus "

Now that we’ve established a normalized set of "Expected" RBI for each player, we can calculate the number of "RBI Added", or "RBI Plus-Minus" by the player over the season. Again, many of the names at the top of the leaderboard are to be expected:

Rank	Player	RBI +/-
1	Ryan Howard	+52.8
2	Josh Hamilton	+41.4
3	Miguel Cabrera	+39.1
4	Kevin Youkilis	+36.8
5	Albert Pujols	+34.5
6	Adrian Gonzalez	+34.3
7	Ryan Ludwick	+33.6
8	Carlos Quentin	+33.5
9	Carlos Lee	+33.0
10	Ryan Braun	+29.9

According to this measure, Morneau comes in 12^th at +29.2 eRBI. Very good, among the elite batters in MLB, but somewhat below his #3 overall rank in RBI in 2008.

Of course, I always find it telling to examine the worst players in MLB:

Rank	Player	RBI +/-
1	Ryan Theriot	-27.6
2	Ichiro Suzuki	-27.5
3	Chone Figgins	-26.6
4	Willy Taveras	-26.2
5	Endy Chavez	-23.1
6	Cesar Izturis	-21.4
7	Akinori Iwamura	-18.8
8	Julio Lugo	-18.6
9	Jason Varitek	-18.6
10	Jason Bartlett	-17.8

A lot of slap hitter, leadoff types in this list. Considering that these hitters had relatively many opportunities (especially given Ichiro’s 749 plate appearances), it says a lot about their approach that they drove in so few runs.

Minnesota Twins 2008 "Studs and Duds"

Player	RBI	eRBI	RBI +/-
Justin Morneau	129	99.8	+29.2
Brian Buscher	47	29.9	+17.1
Jason Kubel	78	66.9	+11.1
Craig Monroe	29	20.4	+8.6
Joe Mauer	85	77.6	+7.4
Denard Span	47	41.2	+5.8

Buscher and Monroe have to be considered the surprises here. Performance with RISP is the key for both, as Buscher batted .386 and Monroe had a 998 OPS with RISP.

Player	RBI	eRBI	RBI +/-
Delmon Young	69	82.2	-13.2
Nick Punto	28	40.0	-12.0
Carlos Gomez	59	67.4	-8.4
Brendan Harris	49	57.0	-8.0
Matt Tolbert	6	11.1	-5.1
Mike Redmond	12	15.5	-3.5

Surprise! Delmon Young and Nick Punto at the bottom of a Minnesota Twins leaderboard. Shocking…

Team Performance

In 2008, the Twins led the league by far with a .305 batting average with RISP. One would expect the Twins to be near the top of the leaderboard in RBI +/-. In reality, the Twins were 11^th in MLB with +29.2 RBI Added. Why the difference? I haven’t gone into the data in too much detail, but I suspect this is due to the lack of home runs by the Twins.

Top 5 RBI +/-

Team	RBI	eRBI	RBI +/-
Chicago White Sox	779	679.6	+99.4
Texas Rangers	867	781.2	+85.8
Philadelphia Phillies	756	696.9	+59.1
Detroit Tigers	768	714.3	+53.7
Baltimore Orioles	750	704.2	+45.8

Bottom 5 RBI +/-

Team	RBI	eRBI	RBI +/-
San Francisco Giants	605	709.8	-104.8
Washington Nationals	608	691.2	-83.2
San Diego Padres	612	681.7	-69.7
Oakland Athletics	610	676.2	-66.2
Seattle Mariners	631	696.5	-65.5

What About 2009?

In 2009, Justin Morneau is currently 4th in MLB with 64 RBI, behind league leading Albert Pujols (77). For opportunities, Morneau is 10th in MLB with 43.2 eRBI, behind league-leading Carlos Pena with 46.7 eRBI. In terms of +/-, Morneau is 6th with +20.8 RBI added, behind league leading Albert Pujols, who has an outstanding +33.0 RBI added. It appears that this year, Morneau is outperforming his slugging peers compared to 2008. Due to Joe Mauer’s hot start, he is 8th in MLB with +20.0 RBI added.

At the bottom of the league, Alexi Casilla is the lowest Twin, 15^th worst in MLB with -10.7 RBI added. Even though the rest of his numbers have declined this year, Delmon Young’s -2.5 RBI added is closer to the middle of the pack than to the bottom of the league.

From a team standpoint, the Twins currently sit right in the middle of the pack, 15th in MLB with -5.4 RBI added. Even though we’ve seen a sharp increase in home runs, an even sharper RISP regression (down to .264, 11th in MLB) likely cancels out any advantage.

Conclusions

We see that many of the league's top RBI men are also at the top of the list in terms of RBI efficiency. This is not surprising, as the league's top sluggers bat in the cleanup spots behind great OBP guys for a reason. Managers (in general) aren't dumb. In the case of Morneau, the fact he has more opportunities than any other hitter has to be taken into consideration, but even with a more average number of eRBI, he'd only be losing 10-15 RBI over the season. Another reason Mauer and Morneau are a great combination for Twins fans.

Note

Sorry I've written so few articles so far this year. This picture helps to explain why I've been busy with other things during the first half of the season. My son, Chase Von Peterson, was born on April 30th. His sisters, Audrey and Olivia love having a little brother and everyone's doing great. Mommy would have been in the picture, but she doesn't have a Twins jersey yet (Orioles Fan...).

0 recs | Comment 31 comments | Share

Story-email Email Printer Print

Comments

Great stuff

Does this shed light on whether “clutch hitting” exists?

And congrats on the baby

by wcooley on Jul 3, 2009 10:52 AM EDT reply actions 0 recs

I'm not sure I would consider this "clutch hitting"...

Since for many of these guys, they’re simply well above average hitters who you would expect to drive in more runs than “expected”. What I believe I have shown is the extent that RBIs are “opportunity” based (the primary knock on RBI as a stat) and how much is due to the hitter’s performance.

A number of other “clutch hitting” studies have been conducted to look at whether a player improves his performance in the clutch (relative to other times), and whether this is repeatable from year to year (indicating a “skill”). Results have shown that if clutch hitting truly exists, it is minimal. Of course, a reasonable argument can be made that “anti” clutch hitting exists simply by watching Alex Rodriguez…

by Adam Peterson on Jul 3, 2009 11:17 AM EDT up reply actions 0 recs

Great post Adam

nice pic of the kids too. :)

PS I knew BUTCHER WAS CLUTCH

by ajmargarine on Jul 3, 2009 11:38 AM EDT reply actions 0 recs

Thanks for this interesting piece

One thing that comes to mind, particularly in regard to the White Sox, is that their ballpark is a homerun magnet. Thus, if someone like Morneau played half of his games there, one would expect that he would have more homers, and thus, a higher RBI efficiency.

Also, with the way that teams play so many games against others in their division, the relative quality of pitching in a division seems like a major factor in these numbers as well.

Who wouldn’t like to play 18-19 games against some of those NL team staffs?

by Old Twins Cap on Jul 3, 2009 12:12 PM EDT reply actions 0 recs

One would suspect that

but the apparent correlation between HR and a higher RBI efficiency is simply a supposition on my part, based on looking at the teams on top of the list. I should probably verify the extent to which they are in fact correlated.

by Adam Peterson on Jul 3, 2009 2:09 PM EDT up reply actions 0 recs

Craig Monroe

Makes me feel even better about despising last year’s promotion of Randy Ruiz – whose skills make him an incredible AAA offensive threat, which is nothing to be ashamed of, but nothing more – at the expense of Monroe.

Bloggin' the bloggers since 1938.

by Johnny Safron on Jul 3, 2009 1:08 PM EDT reply actions 0 recs

Monroe was nothing special last year

and Ruiz was tearing up Rochester. I was fine with giving him the opportunity. Yes, Monroe was pretty efficient at driving in runs. He wasn’t terribly productive overall though, and in general an abnormally high AVG/OPS with RISP is going to level out.

by Adam Peterson on Jul 3, 2009 2:11 PM EDT up reply actions 0 recs

Ruiz is tearing up...

…Vegas this year. But that’s what he does. He’s a career AAA player. He tears up AAA. He’s not a big-leaguer.

It’s the little things that make a difference. Had the Twins picked up Hawkins rather than Guardado and kept Monroe and forgot about Ruiz, they would have been in the AL playoffs last year.

Bloggin' the bloggers since 1938.

by Johnny Safron on Jul 3, 2009 5:02 PM EDT up reply actions 0 recs

Painful but true...

by Jesse on Jul 3, 2009 5:03 PM EDT up reply actions 0 recs

Better yet

Better how, how about taken Chad Bradford of waivers instead of… doing nothing…

"You can't sit on a lead and run a few plays into the line and just kill the clock. You've got to throw the ball over the damn plate and give the other man his chance. That's why baseball is the greatest game of them all."
~ Earl Weaver
"In God we trust. All others must provide evidence."
~ Billy Beane

by AdamOnFirst on Jul 3, 2009 6:55 PM EDT up reply actions 0 recs

Or

If that flyball to center would have been 5 feet deeper

by wcooley on Jul 3, 2009 9:37 PM EDT up reply actions 0 recs

When you don't know what you're talking about...

…speak in generalities, rather than specifics.

But then, I don’t need to tell you that.

Bloggin' the bloggers since 1938.

by Johnny Safron on Jul 4, 2009 1:16 AM EDT up reply actions 0 recs

Some thoughs, questions, observations

- How did oyu create the eRBI stat (I’m interested in more specifics). For creating this data in general, did you manually go through the whole season manually for all these guys at bat by at bat, or was there a list somewhere cataloging the number of times they came up in each situation (like, 1 out, 1st &2nd, 35 times, etc)?

- I would like to see an RBI+% stat similar to OPS+. Morneau (or another high RBI+/- guy) has their RBI + pushed up by the numebr of chances (ie, Buscher’s 17 is super impressive considering that’s out of very few chances.) So Morneau’s RBI% last year would be right at about 29%. Basically, we’re just adjusting it to a rate stat to remove playing time from the equation. Guys on first are really different stories from second and third, and might be looking into alone, since you have to actually slug the ball to score them.

- In general, I’d like to learn a lot more about how you created the data to form this stat, where you got it, etc, and see a lot more explanation of it. Not because I don’t trust it, but because I think it’s a really useful something that ost people would understand easily, and doesn’t really make up any numbers about “runs created” that some traditional people don’t like. IE, I think this creation is better than what baseball prospectus has for this right now (which is simly % of runners on base driven in) and deserves to be used in more than just one article, and more than by just you. This could “catch on” if you tell us more about how to create and find the data.

- “Why the difference? I haven’t gone into the data in too much detail, but I suspect this is due to the lack of home runs by the Twins.”
I would generally agree, the same with the slap hitter guys. A lot of those guys are good hitters, and somebody you would want up with RISP, but they aren’t going to drive in anyone off of first base, and will miss out on maybe a dozen RBI a year jut from not hitting many solo HR’s or HR’s with guys on first all year. Slugging in those guys from first is such a nice luxury…

by AdamOnFirst on Jul 3, 2009 2:15 PM EDT reply actions 0 recs

Oh

Oh, and if we want to get really intense, we could do things like eRBI-Created (see how many of thsoe RBI opportunities are Joe Mauer’s doing, and how many are like, Denard Span being good or something), eRBI-expected (ie, I dunno, the average number of eRBI for hitter at each spot in the batting order, probably separated by league to adjust for the pitcher, we could call the eRBIe) and other fun things.

I’m excited, this is a relatively easy to understand, palatable for traditional fans, way to look back and very well answer a sinple and important question: when this guy came up, how many did he drive in compared to other players? This could become a TOP stat to look at for MVP awards (of course, then we’d want peopel to look at it by position, which we could potentially adjust in the stat, RBI+%ORP, or we could just see where a player stands compared to other players of their position on a list, ie, oh look, Morneau has the highst RBI+% in the league, but he’s barely ahead of several first baseman, and, say, Curtis Granderson is like 3 times ahead of the next center fielder and plays great D out there, maybe he’s more of a standout. That leaves it open to debate about how important different positions are as well, instead of trying to end debate with a final number.

That level of absolute certainty, ie, a number like VoRP being the final word on offensive production, adjusted for all possible considerations, is maybe mot troublesome to traditional columnists because it makes their job pointless. how are they supposed to write multiple 500 word stories a week for months on end if there is no debate to be had between considerations because we’ve already quantified those considerations and answered the questions in an exact and absolute fashion? Columnists don’t liek their ability to pump out their quota of words being threatened.

by AdamOnFirst on Jul 3, 2009 2:24 PM EDT up reply actions 0 recs

I was thinking the same thing. Seems like it would be relevant to take into account the skill of the baserunner. Even if Mike Redmond gets on base 40% of the time, it’s harder to drive him in than Carlos Gomez.

by ben2 on Jul 3, 2009 2:37 PM EDT up reply actions 0 recs

Definitely harder

which is why the RBI-based stats aren’t as effective as the other run-based “little things” stats I created before the season. In those cases, the hitter gets credit for how far an “average” baserunner would have advanced, considering location and type of batted ball (GB, LD, FB, etc). The baserunner gets credit (or blame) for advancing the extra base or not.

by Adam Peterson on Jul 3, 2009 2:40 PM EDT up reply actions 0 recs

Wasn't even

Oh gosh, I wasn’t even TALKING about that. i didn’t think of runner speed. You’re right. Another thing to consider, though that’s a LOT tougher to deal with.

by AdamOnFirst on Jul 3, 2009 4:43 PM EDT up reply actions 0 recs

When I originally set out with this effort

after the 2008 season, my goal was to account for each and every run scored and allowed by breaking up each play into pieces that could be allocated to the batter, runner, fielder or pitcher. I’m done with offense and still working on defense/pitching. Distilling down to a single number, based on runs, allows this type of discussion to occur.

by Adam Peterson on Jul 3, 2009 2:38 PM EDT up reply actions 0 recs

Answers

1. I’ve written my own software to read and manipulate MLB Gameday data. This is what I used before the season for analysis of the “Little Things” in baseball, including base running. Basically, I wrote this piece of software to run through each and every plate appearance and count the number of RBI for each situation. That’s how I got the baseline. Then, I updated my season totals-generating utility to go through and calculate eRBI for each of his PA, summing over the entire season.

2. After I calculated the +/- data, I noticed the lack of normalization for number of PA. This would be a solid next step…

3. Again, I spent much of the last offseason working on software to read the MLB data. I’ve been kicking around the idea of posting a number of utilities (here of course ;) to view these type of stats on a per game basis. I’ve also considered posting libraries and an API to allow other software-minded folks to write their own analysis tools. I’ll get there, but it will take some time.

4. Being able to slug guys in from first helps, you’re right. With no one out and a guy on first, eRBI for the situation is 0.092. What does this mean? Every time a batter comes up in that situation and fails to drive in the runner, he loses 0.092 in +/-.

by Adam Peterson on Jul 3, 2009 2:35 PM EDT up reply actions 0 recs

Wo wrote?!

You wrote a program? You are way advanced into this. I don’t know if any organizations do anything like this internally, but you could literally publish this and potentially affect the baseball player evaluation industry. Please, if you have time, write longer things that do more to explain the stats you’ve created and their process, instead of just answering a specific question (ie, how good is Morneau) with them, because this is a really big deal, what you’re doing, and it deserves to be understood/known about by even more people.

by AdamOnFirst on Jul 3, 2009 4:48 PM EDT up reply actions 0 recs

I've written a few articles

here and elsewhere about the software, and I plan to do so more in the future.

“Total Run Accounting”, or The Value of Doing the “Little Things”

“Total Run Accounting” – 2008 Player by Player

“Total Run Accounting” – Team Offense

Total Run Accounting – Teams, Players and “tORA”

by Adam Peterson on Jul 3, 2009 10:00 PM EDT up reply actions 0 recs

Lots of data...

Nice work sifting through all that data. I think bb-ref has splits for outs and splits for all the base combinations but not all the base-out combinations. Maybe some other place does. That would save some time.

A couple of questions.

Why is it easier to get RBI’s with the bases loaded with 1 out as opposed to none?

I found a similar study from the 90s here:

http://www.baseballthinkfactory.org/btf/scholars/ruane/articles/rbi_production.htm

And they found the same thing. I’m thinking there must be an easy explanation but I’m missing it.

Second question: What do you do with BB & HBP? Excluding IBB’s from the analysis sounds very logical, but the unintentional free passes is another story. Some would like them excluded (pitcher being extra-careful, added baserunner) but others might disagree (passing the RBI “buck” to the next batter).

by DavidRF on Jul 3, 2009 3:35 PM EDT reply actions 0 recs

Bases loaded, one out versus none

I’m missing it as well. There’s a slight difference between the two situations, but I can’t put my finger on the reason. Perhaps the fact that a team may be more likely to play at DP depth and go for a DP may make a difference, but it seems the effect would be opposite…

by Adam Peterson on Jul 3, 2009 10:04 PM EDT up reply actions 0 recs

BB and HBP

I include all plate appearances in this analysis, regardless of outcome. IBB, BB, HBP, they’re all in there. IBB should probably be excluded, you’re right. I’d probably want to keep BB in there since an opportunity existed to drive in a run, but the batter didn’t get the RBI. May or may not be his fault though.

by Adam Peterson on Jul 3, 2009 10:06 PM EDT up reply actions 0 recs

I think walks should be left out

after all, do we expect the batter to chase after 4 or more pitches outside of the strike zone just because there are runners on base? we go ape sh*t mad when we see gomez and young do that, are we now saying that morneau should too?

Will the Real Thor Please Stand Up ... ?

by the Real Thor on Jul 4, 2009 8:31 AM EDT up reply actions 0 recs

OUTSTANDING work, Adam!

This is why I’m glad you’re part of this team, this is the kind of stuff I could never do. You’re basically inventing stats, aren’t you? And not made up stuff either, it’s easily understandable and tells a story…exactly what stats should be.

Can you do this again at the end of the year? It’d be a great piece of year-end review.

by Jesse on Jul 3, 2009 4:58 PM EDT reply actions 0 recs

Sure thing

Yeah, I think I’m inventing some of these stats, and more to come in the future. If anyone has any ideas, let me know and I may take a crack at it.

by Adam Peterson on Jul 3, 2009 10:01 PM EDT up reply actions 0 recs

Bases loaded, one out versus none

Nice article Adam. I think the reason it is easier to get RBIs with 1 out and the bases loaded rather 0 outs is that with 0 outs there is incentive to take the out at home rather than trying for the double play. This would happen in every bottom of the inning tie game situation 9th inning or later automatically, but also would be the better play in many close game situations.

AdamOnFirst – You can use Retrosheet PBP info to do what Adam does for any completed season. Adam’s Gameday programming gives him the advantage of doing studies on current season data as well. Adam’s program is very useful but not unique. Many amateur analysts have written them and I am sure most teams have there own.

by pjensen on Jul 3, 2009 10:51 PM EDT reply actions 0 recs

Thanks

Good point on the bases loaded situation. I was thinking along the same lines myself.

I considered using Retrosheet, but in the end I preferred to create a general API I could use for a variety of purposes. Accessing the data was relatively easy. It was breaking down each play into a series of “transitions” (not really applicable to this article) and creating the various league baselines (hit chart by zone, probability of a runner advancing, etc) that took the most time and effort.

by Adam Peterson on Jul 5, 2009 7:16 AM EDT up reply actions 0 recs

I would make one suggestion

when placing the order of the league bottom dwellers, I would put them in a reverse order ,so the 10th worst would be at the top, and the absolute worst would be at the bottom. makes it a little easier to read.

Will the Real Thor Please Stand Up ... ?

by the Real Thor on Jul 4, 2009 8:34 AM EDT reply actions 0 recs