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

by Adam Peterson on Jan 31, 2009 9:34 PM EST Comment 34 comments

I've finally completed my initial analysis of 2008 offense, writing software that uses MLB Gameday play by play data and accounts for every run, breaking down each play to a level where offensive runs can be assigned to hitters, baserunners and the opposing team's fielders.

I've posted a team by team spreadsheet with total "expected runs" for baserunning, hitting, etc to Google Docs here.

I also posted a detailed description of the method behind this accounting to BTB here, so I won't go into a bunch of detail on how these numbers are calculated. I do have a number of interesting results from a Minnesota Twins perspective.

We all know the Twins are considered by "baseball people" to play the game the "right way", moving runners over, bunting, taking the extra base, etc. The question many of us have asked ourselves is, what is the value of doing these "little things"? Does this give the Twins a significant advantage over other teams, relative to their "standard" performance metrics such as OBP, SLG, wOBA, WAR, etc?

Well, I found that it does. In 2008, by doing these "little things", the Twins helped themselves to the tune of about 27 runs compared to the next best team in baseball, much less the average team (42 runs) or other AL Central contenders such as the White Sox (65 runs).

In this context, what are the "Little Things"? In my analysis, I consider baserunning (stolen bases, caught stealing, taking the extra base, not getting thrown out on the base paths) to be one component. In baserunning, the Twins came out 8.44 runs above the average MLB team, primarily due to taking more extra bases, more often than any other team. Because the Twins' runners were also thrown out, they came out 7th in MLB in baserunning, behind Philly (#1 at +25.15 runs compared to average). By comparison, the White Sox were third worst in the majors (-11.12 runs) due to their general station to station style of play.

The second component of the "little things" is hitting the ball in locations where the baserunner is more likely to advance an extra base. In the spreadsheet, I note "Standard" batting as the outcome of the batted ball, assuming a standard advance of one base for a single, two for a double, etc. I then determined the percent chance (using all games from the 2008 season) where a runner would advance an extra base (e.g., first to third) based on the type of batted ball (GB, FB, LD, bunt, pop fly) and location (position that fielded the ball). For example, a runner is more likely to advance to third on a line drive to the right fielder than on a line drive to the left fielder (47% versus 18%). I assign the batter credit for the "expected" chance that a runner would advance. Then I assign the baserunner credit or blame depending on whether he actually advanced. In the case of a first to third single on a LD to RF, the batter would get 47% of the credit for advancing the runner to third (and all of the credit for getting him to second and himself to first), and the baserunner would get credit for the other 53%. The additional 47% is considered "Other" Batting, part of the "Little Things".

This is where the Twins blow away all other MLB teams, coming in at +33.78 compared to league average. The next best team in the majors is Boston, at +16.19, and the White Sox come in at -11.97, 8th worst in MLB.

The good news here (from a Twins perspective) is that these "little things" should be repeatable. They are based on aggressiveness and speed on the basepaths, and hitting the ball the other way. Considering that the Twins are 27 runs better than any other MLB team (at least in 2008) at doing these "little things", this corresponds to +2.7 value wins, or equal to Prince Fielder and Vlad Guerrerro, or just slightly greater than Alfonso Soriano and Pat Burrell (+2.6), or Torii Hunter (+2.5) in total value wins above replacement, according to Fangraphs. In other words, the equivalent of signing a $12.2M FA.

10 recs | Comment 34 comments

Story-email Email Printer Print

Comments

WOW!!!

Adam,

What a lot of work. Thanks for your hard work.

I’m not a big stats guy. Yet, I am always curious as to how one might quantify the the uniquely human aspects of any endeavor.

Very interesting.

Thanks,

I don't suffer from insanity...I relish every moment of it!

by the Dragon on Jan 31, 2009 10:56 PM EST reply actions 1 recs

Thanks for posting this

I have to read this again when it’s not 2 in the morning but I definitely like to read quantitative analysis that says that the Twins are doing something right offensively besides just dumbly cashing in with runners in scoring position. Also, it’s cool to read something different than the typical Aaron Gleeman line about how the Twins don’t really do the fundamentals right. In the morning, I’ll re-read your post, but for now thanks for doing something creative with stats and sharing it with us.

"I don't care about feelings." - Lou Piniella

by natetheskate on Feb 1, 2009 7:08 AM EST reply actions 0 recs

Gleeman and fundamentals

I’ve heard similar statements about the Twins not doing the “fundamentals” right. The real question is, how do we define “the fundamentals”? I suspect our below average defense and aggressive (easily the most aggressive in the majors) baserunning, resulting in a bunch of outs on the basepaths, makes it appear we are less fundamentally sound than we are.

by Adam Peterson on Feb 1, 2009 1:10 PM EST up reply actions 0 recs

very nice

One of the biggest parts of analyzing data is just accumulating it in the first place, and you’ve certainly done that AP – well done.

I’m going to back Aaron on the fundamentals line: they don’t field like they used to in TK’s time, and imagine if their bunt attempts were successful at a reasonably higher % (are you listening, Carlos?). I don’t believe people (including AG) are saying that the Twins don’t do the fundamentals right, more like they are no longer the fundamentally sound team they were pre-Gardy. I’ll take on anyone who can say that the Twins’ fundamentals are as good as they were 8 years ago, and haven’t atrophied in those 8 years.

by Rhubarb_Runner on Feb 2, 2009 6:57 PM EST up reply actions 1 recs

Fundamentals != Little Things

For the non-nerds in the crowd, “!=” is “not equal” in various computer languages.

There’s not really any reason both can’t be true. Certainly fundamentals implies “doing the little things” in the modern baseball vernacular, but I don’t know that it necessarily correlates to what Adam measured here. Taking an extra base is certainly an exercise in awareness of the ball’s location and trusting your third base coach, but I don’t know that it’s as fundamental as getting a bunt down or taking a good route to a ball in the outfield. Directional hitting, the other major component of these numbers, is again a fundamental, but if you’re attributing the focus on fundamentals to TK over Gardy, that can likely be ascribed more to the hitting coach than the manager.

Also, I’d like to reiterate everyone’s kudos to Adam – this is one of the most useful blog posts I’ve ever read, and I agree with others’ comments that it’s a legitimately important new step in sabermetrics. I’d also love to see this for other years to see how repeatable these results are, although I don’t know whether the data is available.

"There are only two things that are infinite, the universe and human stupidity, and I'm not sure about the former." - Albert Einstein

by BeefMaster on Feb 3, 2009 10:29 AM EST up reply actions 0 recs

That might be true, but Gleeman conflates the terms

and Howard Sinker does too. Check this excerpt from Gleeman’s 5/18/08 post (not sure how to make quotation blocks look nice):

“As evidenced by the previous bullet point this site certainly tends to be far more critical of the Twins than coverage that you’ll find in newspapers and on television or radio, so even if they’re late to the party it’s always nice to see the local mainstream media picking up on some points that I’ve been harping on here. For example, here’s Howard Sinker of the Minneapolis Star Tribune on the Twins’ reputation for "doing the little things” right:

Anyone who prattles on about the Twins “doing the little things” right is living in the past. It is a catchphrase of the national media, which tends to live a few years behind reality when it comes to teams not playing on the Coasts or in Chicago, and of local loyalists who need vision and comprehension checks. It’s just not happening any more and the sum total of the Toronto sweep should drive home that point to anyone still doubting it.

The former manager, Tom Kelly, imbued his teams with the fundamentals and Gardy’s division-winning teams were in large part the result of Kelly’s ways of doing things. The current group has lost the right to carry that banner.

And that was before committing three ugly errors Sunday. I’ve been trying to point out the long-expired nature of the Twins’ “doing the little things” reputation for several years now, yet the national media and much of the local fan base continues to treat it as gospel."

"I don't care about feelings." - Lou Piniella

by natetheskate on Feb 3, 2009 3:00 PM EST up reply actions 0 recs

even more explicit

Here’s Gleeman on 9.1.2005:

If you’re going to hit like it’s the 1960s, you need to play the rest of your game that way too. Instead, the Twins neither hit for power or “do the little things,” and the end result is really ugly to watch.

Remember all that stuff we used to hear about the Twins “playing the game the right way”?

"I don't care about feelings." - Lou Piniella

by natetheskate on Feb 3, 2009 3:07 PM EST up reply actions 0 recs

It will be interesting to see if Gleeman picks this up in his Link-o-Rama

He’s a lot more open to different perspectives than he was, say two years ago. So, he might look at this and say, “Hey, I didn’t know that was measurable. Now that it is, let’s put it into the toolbox and use it for analysis.” I won’t presume that he’ll reject this analysis just because “doing the little things” has never been measured and it’s a Bill James core principle that it doesn’t make a hill of beans. So I’ll just watch and hope he has an open mind about it.

I for one am convinced. It reminds me of the old days when the best defensive metric out their was range factor and RF reports featured as many anomalies as truths. At the time, I said you just couldn’t measure defense because there are too many variables. Then ZR and UZR came out and they were a lot closer to reality. They’re not perfect, but they’re a heck of a lot better than nothing. Just having a measuring stick for defense has changed the conversation: It’s now valued much more highly than it was when people only wanted to talk about what they could measure (pitching and hitting).

I have always thought doing the little things made a big difference in scoring. But I couldn’t prove it. Without a measuring system, those who poo-pooed the little things would always win the argument because you can’t prove a negative (or a counter factual).

I would say, “if he doesn’t hit the ball to the right side there, the runner stays at second, where a fly ball doesn’t score him.” They would say, “Maybe if he doesn’t hit to the right side, he hits a home run.” The trouble with their argument is, it’s a lot easier to hit a weak grounder to second than a homer, and if you try hitting a homer, you’re much more likely to strike out or hit a sharp grounder to third, keeping the runner at second.

Now for the first time, we have a measurement system. Hopefully that will put the little things into the conversation where it belongs.

"You're thinking too much. Just have fun." -- Bennie "The Jet" Rodriguez in Sandlot

by cmathewson on Feb 3, 2009 3:25 PM EST up reply actions 0 recs

WOW!

This is so far beyond what some of us old fahrts can understand. Looks like a ton of work, will be great seeing your posts about how the Twins are doing during the 2009 season. Who knows, maybe there is a reason that the Twins score more runs than all the “experts” expect?

by roger13 on Feb 1, 2009 9:07 AM EST reply actions 0 recs

I'm impressed,

thanks for the work on this. But after all this, what I still want to know is…you actually WRITE that software? I’m a little envious. And also, how much time did you spend compiling this data?

by Jesse on Feb 1, 2009 9:25 AM EST reply actions 0 recs

Yep. I wrote a good amount of software for this.

I started pretty much once the season ended, working part time during lunch or on business flights (too many of them lately). Software and math is pretty much my background, but since I’ve moved into management lately, this was a decent way to keep myself up to date.

The good news is that I focused most of the software on reading and manipulating the MLB Gameday data, so I can write small snippets to answer nagging questions like, did Carlos Gomez become more patient as the year went on (chase rate, pitches taken, month by month or week by week). Etc. As I work through these analyses, I will post results. In the meantime, if anyone has any suggestions for when I complete the defense (pitching, fielding) part of “TRA” (yes, I need to find a better name), I’m all ears.

by Adam Peterson on Feb 1, 2009 9:35 AM EST up reply actions 0 recs

Adam...

…did you send this information on to Gardy? Gotta believe that like most of us, he wouldn’t understand it….but, he sure as heck would love having someone confirm that he and the Twins are doing a lot of the “little things” right! Now, if they can get back to playing defense like they used to this team may be special this year.

by roger13 on Feb 1, 2009 9:37 AM EST up reply actions 0 recs

Considering the Twins have been relatively outspoken about their lack of a "quantitative analyst"

I’m not sure how much they’d care. Perhaps they’re crazy like a fox… It’s not like I have Gardy’s email address anyway.

by Adam Peterson on Feb 1, 2009 11:24 AM EST up reply actions 0 recs

A lot of Twins front office people have email addresses

including Bill Smith, which you can find here:

http://minnesota.twins.mlb.com/team/front_office.jsp?c_id=min

Gardy doesn’t seem to have one, but you could always try just sending an email to twins@twinsbaseball.com with Ron Gardenhire in the subject line as they suggest.

"I don't care about feelings." - Lou Piniella

by natetheskate on Feb 1, 2009 2:30 PM EST up reply actions 0 recs

I like TRA...

…Twins Run Away, or Twins Run Always!

by roger13 on Feb 1, 2009 9:41 AM EST up reply actions 0 recs

I must admit that I'm dubious about this

It seems beyond the realm of the possible that these “little things” could actually be worth that many runs.

However, even accepting your conclusion, there is one thing you left out of your calculations that I suspect is rather important: how many runs did the Twins give up by “giving themselves up”“hitting the ball the other way.” You have calculated how many runs they gained by doing this and running the bases aggressively, but there is another side of the coin. Specifically, how many base-hits, doubles, home runs did they NOT get because they hit the ball the other way/bunted/whatever?

Imagine 10 PAs in which the Twins successfully went the other way to advance a runner, and 10 where another team did not. What you have done is calculated how many runs those 10 “successful” PAs are worth. But what you haven’t done is calculated what the 10 PAs by the other team are worth—you have, apparently, calculated them as 0. But realistically, that simply isn’t true. Instead, there would be base-hits/walks/what have you in there, which would have run value.

by Eric in Madison on Feb 1, 2009 10:49 AM EST reply actions 0 recs

It's actually accounted for...

since I’ve also calculated the “standard” batting expected runs first. This is where the batter is penalized for grounding out to second base, in my notation first being penalized for creating a “transition” from situation X2X/0 (runner on second, none out, 1.19 ER) to X2X/1 (one out, 0.72 ER). Thus, the batter is assigned -0.47 “standard” ER. This is where he is penalized for giving himself up. However, if he struck out, he would be penalized the same -0.47 runs. Since the batter grounded to the right side, he should get some (marginal) credit for advancing the runner to third. Since there is a 61% “advance rate” going to third on a GB out to 2B, I assign the batter 61% of the ER for the runner advancing to third, 61% of +0.33 = +0.20. In other words, in this situation, the batter ends up penalized -0.47 ER for the groundout, but rewarded +0.20 ER for potentially advancing the runner.

To your point about considering the “lost” base hits, doubles, etc for going the other way, well, we should see this in below average “standard” batting runs. To some extent, we see this, as our +22.81 standard batting runs come in 9th in the AL.

Finally, to your point about the 10 PAs…the 10 PAs where the Twins gave themselves up would overall end up as negative runs for the offense, similar to the example above. What I am capturing is the difference between making a “productive” out and an unproductive out like a strikeout. Strictly replacing 10 strikeouts in the above situation with “productive” outs would add a total of about 2 expected runs difference. Over a season, those runs add up.

by Adam Peterson on Feb 1, 2009 11:23 AM EST up reply actions 0 recs

One more item

from a strategy analysis standpoint, you have a very good point. It may be a better strategy to try to pull the ball and increase chances for 2B/HR than to hit the other way. I’m not really tackling that aspect yet. My methods do explain the large gap between our “expected” runs scored using wOBA, OBP, SLG, and actual runs scored. I also believe there is very good evidence that doing the “little things” may contribute at least as much to closing that gap as hitting with RISP.

by Adam Peterson on Feb 1, 2009 1:05 PM EST up reply actions 0 recs

Bravo and thank you

Twins Territory needed to see a post like this. I’ve been struggling to come up with the numbers to make a more convincing argument, but I think you’ve done a great job.

Predictably most Twins fans want a HR masher and not more small ball guys. But I think with the advent of more stringent drug testing in MLB we will be seeing more and more offenses making a strategic shift towards this type focus. Yes, there will always be some teams that focus on the long ball for a variety of reasons (most likely circumstance—they drafted a bunch of guys who end up being able to crush it, or location—they play in a very HR friendly stadium), but I bet even this year you’ll see another few teams playing this style of baseball and having offensive success.

Finally, no statistical analysis is perfect. If we believed that what the Gleeman’s of the world said was the final result or analysis, well, as much as I love reading Gleeman’s stuff I’m not sure I’d be a Twins fan. We’d never make it to the post-season. Stats are useful to me if they help describe something that I’m seeing in a way that aids in my understanding. And to that end I really appreciate your work and look forward to seeing more of it.

2008 Twins runs/HRs – 829/111
2008 ChiSox runs/HRs – 811/235

by biggity2bit on Feb 1, 2009 12:19 PM EST reply actions 0 recs

The awsomeness of blogs

Adam, we here at TwinkieTown are forever in your debt. I could not have even conceived of this analysis, let alone executed it. To my knowledge, it is entirely original, and deserves a place with the developers of xFIP and UZR. I echo others here: THANK YOU!

"You're thinking too much. Just have fun." -- Bennie "The Jet" Rodriguez in Sandlot

by cmathewson on Feb 1, 2009 12:36 PM EST reply actions 0 recs

+100

Like Roger, I am an old fart. Your post was a little beyond me, but I too am happy to hear statistically speaking that the Twins are doing the little things much better than the rest of baseball. Again, thank you for all your hard work. Great job…. Go Twins….

by Beerbear on Feb 1, 2009 3:54 PM EST up reply actions 0 recs

+1

For everyone who loves this post, btw, don’t forget to REC it.

by Jesse on Feb 1, 2009 8:09 PM EST up reply actions 0 recs

Agreed

Although I’ve not had the time to look this over in details, I can only restate cmathewson’s line: This is not only awesome, it is a significant development in baseball research.

by PhoenixV on Feb 1, 2009 11:25 PM EST up reply actions 0 recs

This is awesome.

Rec’d.

by matty_b on Feb 2, 2009 11:25 AM EST reply actions 0 recs

great work

I’ve been anticipating this post for a while now, and I have to say it’s even better than expected. I’m just curious to know when we’ll be able to see statistics for individual players. That would be really great to see.
It would also be nice to see stats from previous years, so we could see how repeatable “doing the little things” really is. Either way, thanks for your hard work!

by lookatthosetwins on Feb 2, 2009 8:41 PM EST reply actions 0 recs

I'm working on the player sheet now

Hope to have something posted by the weekend calling out individual players and their ER contributions due to baserunning, “little things” etc.

MLB Gameday exists for previous years, so I should be able to apply the same software and analyze repeatability. How quickly will depend on whether MLB changed formats of the data.

by Adam Peterson on Feb 3, 2009 12:11 PM EST up reply actions 0 recs

Important stuff

The presence of year-to-year repeatability, I think, is the real key to this analysis – it changes it from a “how did this happen?” after-the-fact calculation to a potentially useful projection tool.

I’d also be curious as to how much the team is responsible, as opposed to the individual player (something that could be analyzed by comparing year-to-year results of players who changed teams) – for example, it would be interesting to attempt to figure out what fraction of the Twins’ enormous TRA output was from directional hitting and team speed, and what fraction was from a team culture of aggressive baserunning and focus on situational hitting.

"There are only two things that are infinite, the universe and human stupidity, and I'm not sure about the former." - Albert Einstein

by BeefMaster on Feb 4, 2009 9:03 AM EST up reply actions 0 recs

Also

based on your data, about how many runs would you say were accumulated from our high RISP numbers? I’m sure much less than Dave Cameron and co. have posted, since a lot of the discrepency between wOBA and our run production has been accounted for in the “little things” and baserunning. Taking whatever that number is, and giving Mauer and Morneau the benefit of the doubt to keep coming up big, our regression might not be as bad as we thought.

by lookatthosetwins on Feb 2, 2009 8:44 PM EST reply actions 0 recs

Depends on how many runs

others are considering the Twins scored above an “expected” total, based on wOBA.

Here’s a quick and dirty “back of the envelope” analysis:

By my calculations (rough, based on OBP/SLG), the Twins’ wOBA was .335. Applying to the 2008 MLB average .333 over Minnesota’s 6331 PA, that comes out to an “expected” RAA = +11.0. Actual RAA was +76, so this leaves a total 65 runs “unaccounted for”, not explained by wOBA.

By my TRA totals, the Twins were +42.2 RAA doing the “little things”, and +5.6 RAA due to opponents fielding (I’ll have to look further to see what drives this, high contact rate and team speed / aggressiveness probably two large factors). These two components are not part of the above wOBA, so that leaves a total of 17 runs unaccounted for that could be considered due to our high RISP.

This is lower than you would get by considering the total number of “extra” hits with RISP times the average number of runners in scoring position, probably because a team gets multiple chances to score runs. A strikeout with RISP doesn’t automatically end the inning…

by Adam Peterson on Feb 3, 2009 12:36 PM EST up reply actions 0 recs

This is a treat!

Thanks for all the time you put into this.
Fun stuff.

"You got to be careful if you don't know where you're going, because you might not get there." - Yogi Berra

by Marv_MN on Feb 2, 2009 8:49 PM EST reply actions 0 recs

awesome!

Thank you! This is brilliant!

What an incredibly huge endeavor — but what a useful thing to try to quantify!

It’s so encouraging to hear the Twins actually, quantifiably do the little things to advance runners. That gives me hope that there really is something to the Twins’ mystique, and hope for their continued success at exceeding expectations.

Just curious: Do you think this is related at all to their high RISP numbers? If you are good at consistently hitting the ball to the correct part of the field, wouldn’t that help you, say, hit a ground ball through the hole when a fielder is holding on a runner? And doesn’t good baserunning cut down on force outs? Both of these things increase batting average with runners in scoring position.

Thanks again! I am really excited to see what else you pull out of all this analysis.

by by jiminy on Feb 3, 2009 12:07 PM EST reply actions 0 recs

Could be

It would be an interesting analysis to assess team speed relative to the probability of a force out. TRA would assign advancing the extra base (avoiding the force out) partially to the batter and partially to the runner if the play was not scored as a hit. If scored as a hit, it’s assumed the runner advances.

Some analyses that would be useful in understanding the Twins and RISP:

- Compare GB, FB, LD% in RISP versus non-RISP situations.
- Compare plate discipline (SO/BB)
- Are the Twins more likely to hit to RF in RISP situations?

by Adam Peterson on Feb 3, 2009 12:44 PM EST up reply actions 0 recs

Agreed

Especially when it comes to Mauer and Morneau, it would be interesting to get a little deeper in the numbers. Mauer especially seems to be a guy that can hit it to a certain field at any time, and that may help him hit a little better with RISP. Also his patience in those situations really helps, as many pitchers are probably trying to be a little too perfect with their pitches. Most of these things would be hard to quantify, but comparing some of the things you talked about would definately be a start.

by lookatthosetwins on Feb 3, 2009 2:57 PM EST up reply actions 0 recs

Looking forward to comparisons against 2003, 2004 and 2006

It is easy to point to a high batting avg with RISP as the one major reason we scored so many runs in 2008. What is interesting is that our team batting avg. with RISP was .296 in 2006 where we scored 801 runs…but in 2003 our team batting avg. with RISP was .268. …..nearly 30 points lower….yet we scored the same number of runs 801 runs scored both years. We had the exact same team OPS both years .772 So there is much that can be analyzed to explain the performance differentials, but your work here would be the first to dive into the “little things” to fill in gaps in any other form of analysis.

In 2004 the Twins scored 49 fewer runs than they did in 2008. Yet, in 2004 the Twins stole more bases than they did last year…they also had a much higher slugging percentage, higher team OPS. The Twins hit into fewer double plays in 2004 (130 versus 142 last year) In a nutshell, in 2004, the Twins were on base about the same percentage as 2008, they had way more power, stole more bases, hit into fewer double plays. Yet the offense produced substantially fewer runs in 2004 than in 2008. Why? Apparently, we did not do the little things that well… that…and our team batting avg with RISP was .277 in 2004. So a much lower batting avg. with RISP in 2004 may explain some of this differential, but how much? I look forward to your analysis.

In 2006 we grounded into a lot of double plays (163….the most in any Gardy year) but we also had a team OPS of .772 versus .748 last year. We stole about the same number of bases and scored similar number of runs (801 in 2006 versus 829 last year)…so why is it that we scored more in 2008 than in 2006 when we slugged so much less, had a much lower OPS…did not steal any more bases….was it the little things…we hit pretty well with RISP in 2006…I don’t know …but it will be interesting to see how your results look year by year

by NorthDakotaTwinsFan on Feb 4, 2009 3:56 PM EST reply actions 0 recs