Could It Be Time to Update WAR’s Positional Adjustments?

by
September 12, 2014

It’s been quite a week for the WAR stat. Since Jeff Passan dropped his highly controversial piece on the metric on Sunday night, the interwebs have been abuzz with arguments both for and against the all-encompassing value stat. One criticism in particular that caught my eye came from Mike Newman, who writes for ROTOscouting. Newman’s qualm had to do with a piece of WAR that’s often taken for granted: the positional adjustment. He made the argument that current WAR models underrate players who play premium defensive positions, pointing out that it would “laughable” for Jason Heyward to replace Andrelton Simmons at shortstop, but not at all hard to envision Simmons being an excellent right fielder.

This got me thinking about positional adjustments. Newman’s certainly right to question them, as they’re a pretty big piece of the WAR stat, and one most of us seem to take for granted. Plus, as far as I’m aware, none of the major baseball websites regularly update the amount they credit (or debit) a player for playing a certain position. They just keep the values constant over time. I’m sure that whoever created these adjustments took steps to ensure they accurately represented the value of a player’s position, but maybe they’ve since gone stale. It’s certainly not hard to imagine that the landscape of talent distribution by position may have changed over time. For example, perhaps the “true” replacement level for shortstops is much different than it was a decade or so ago when Alex Rodriguez Derek Jeter, Nomar Garciaparra, and Miguel Tejada were all in their primes.

I decided to try and figure out if something like this might be happening. If the current positional adjustments were in fact inaccurately misrepresenting replacement level at certain positions, we’d expect the number of players above replacement level to vary by position. For example, there might be something like 50 above-replacement third basemen, but only 35 shortstops. Luckily, the FanGraphs leaderboard gives you the ability to query player stats by position played, which proved especially useful for what I was trying to do. For each position, I counted the number of plate appearances accumulated by players with a positive WAR and then divided that number by the total plate appearances logged at that position. Here are the results broken out by position for all games since 2002.

Ch1

Based on this data, it seems like the opposite of Newman’s hypothesis may be true. A significantly higher portion positive WAR plate appearances have come from players at the tougher end of the defensive spectrum, which implies that teams don’t have too difficult of a time finding shortstops and center fielders who are capable of logging WARs above zero. Less than 13% of all SS and CF plate appearances have gone to sub-replacement players. But finding a replacement-level designated hitter seems to be slightly more difficult, as teams have filled their DH with sub-replacement-level players nearly 30% of the time. Either teams are really bad at finding DH types (or at putting them in the lineup), or the positional adjustments aren’t quite right. The disparities are even more pronounced when you look at what’s taken place from 2002 to 2014.

Ch2

The share of PAs logged by shortstops and center fielders hasn’t changed much over the years, but the numbers have plummeted for first basemen, corner outfielders, and DH’s. From Billy Butler and Eric Hosmer, to Jay Bruce and Domonic Brown, this year’s lineups have been riddled with sub-replacement hitters manning positions at the lower end of the defensive spectrum. Meanwhile, even low-end shortstops and center fielders, like Derek Jeter and Austin Jackson, have managed to clear the replacement level hurdle this season if we only count games at their primary positions.

The waning share of above-replacement PA’s coming from 1B, LF, RF, and DH has caused the overall share to drop as well, with a particularly big drop coming this year. Here’s a look at the overall trend.

 

Ch3

And here it is broken down by position…

 

Ch4

And just between this year and last…

 

ch5

 

Frankly I’m not sure what to make of all of this. I’m hesitant to call it evidence that the positional adjustments are broken. There could be some obvious flaw to my methodology that I’m not considering, but I find it extremely interesting that there’s been such a shift between this year and last. We’re talking an 8 percentage point jump in the number of PAs that have gone to sub-replacement-level players. Maybe its been spurred the rise of the shift or maybe year-round interleague play has something to do with it, but it seems to me that something’s going on here. And I’m interested to hear other people’s thoughts on these trends.





 

Chris works in economic development by day, but spends most of his nights thinking about baseball. He writes for Pinstripe Pundits, FanGraphs and The Hardball Times. He's also on the twitter machine: @_chris_mitchell None of the views expressed in his articles reflect those of his daytime employer.

43 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
haishan
9 years ago

This really isn’t much more than a just-so story, but it seems plausible to me that in 2002 — well within the steroid era, and pre-PITCHf/x — there was a relative surplus of players who could hit the crap out of the ball but couldn’t run/field/throw very well at all. In other words your prototypical 1B/DH. If this is the case there would be a good argument for tweaking those positional adjustments.

I have no such hypothesis for why so many PAs are going to sub-replacement level players in 2014, though. Maybe it’s just weird noise and things will bounce back next season, maybe there’s some underlying cause — it seems too early to tell for sure.

1
Chris Mitchellmember
9 years ago
Reply to  haishan

Yeah I had that thought as well on the 1B/DH types. If that is what’s going on, it might be partly driven by teams’ increased focus on defense. They might be more willing to give DH AB’s to a guy who can at least be serviceable defense, even if he’s not quite as good offensively.

0
Z
9 years ago

I think it somewhat has to do with the fact that the offensive positional adjustments are done in runs, but the impact of runs on winning depends on the run environment. Saving runs isn’t quite the same as earning runs, because well… you can’t save runs past 0. No amount of run-saving will help you if your team scores zero runs. Correspondingly, as the run environment gets more bleak for hitters, people who give significantly extra offense add more value.

Since we’ve lost about 10% of the runs scored (~4.75 to ~4.25), you’d expect that the runs that you get are worth about 10% more. However, the positional adjustment doesn’t float with the run environment. The hit to each position is very proportional to the dependence on runs vs. positional value. I bet you could plot it, even. Some of this is just the current crop of players though. We have one of the strongest groups of CF in recent memory. Nearly half of them have hit above average this year, with multiple stars.

4
Chris Mitchellmember
9 years ago
Reply to  Z

This is a really interesting explanation. Still wrapping my head around it, but I think you might be on to something.

1
tz
9 years ago

Two points come to mind:

1. The overall decline in offense in baseball over this period is driven a lot by a big increase in strikeouts, which probably tends to hurt DH/1B/LF types more than guys who play up the middle. This would bring the average batting runs contributed closer together across positions.

2. If you consider the long-term fluctuation of talent mix by position, then the only positional adjustment that looks like it might need a change is for DH. This might be a consequence of fewer teams having a full-time DH, but using it as a place to either platoon or give everyday players a chance for some rest. I do wonder if this trend would reverse as guys like Fielder, Pujols, Miggy etc. begin to reach the point where they’re just full-time DH’s, which could skew the average offense from DH’s back upward.

3
Chris Mitchellmember
9 years ago
Reply to  tz

Yeah good point about fewer full-time DH’s. Instead of employing David Ortiz types, more teams seem to use their DH spots as a revolving door for guys who sometimes play the field. This definitely gives teams more flexibility, but could also explain the weaker production from the DH position.

2
Mike Newman
9 years ago

Hey Chris,

Thanks for building off of my piece the other day. Any chance to learn and grow baseball knowledge is awesome.

After reading it, I have a couple of thoughts on the topic. First off, my argument comes down to scarcity and the idea 3 people on planet earth can play SS defense as well as Simmons, yet a great deal of SS, 2B and CF can transition to LF and leverage their athleticism into strong defensive numbers at an easier position. Even with the positional adjustment, it doesn’t seem like it’s enough.

Speaking directly to this piece, common sense would say premium athletes will avoid the disabled list more than plodders. The field of 1B, LF, RF and DH includes older, less athletic players who are more likely to become injured with age. Younger, more athletic players up the middle will stay healthier and heal faster than their counterparts.

Also, this piece still looks at players position-by-position. My idea is positions are are not independent of each other. Literally every player in MLB can be penciled into the DH position. Far fewer can play SS. The lack of replacement level players at corner positions known for offense speaks to bad hitting across the league.

I tried to address this a little more in a piece at ROTOscouting yesterday and on the podcast which was released this morning. It’s too much to spell out here, but the gist is if we can’t close our eyes and envision a Brandon Crawford/Mike Morse swap ever happening, then a 15, 25, 125, 1,025 run adjustment still isn’t enough to accurately gauge defensive value.

2
Chris Mitchellmember
9 years ago
Reply to  Mike Newman

Thanks Mike, I’m a big fan of your work and appreciate your taking the time to read mine.

Health is an interesting idea, as older DH types are almost certainly more likely to go down with injury. I’m not sure why this trend would have become more prevalent in recent years, though. Even for DH’s, the positional adjustments/replacement level seemed about right in 2002.

As for your Crawford/Morse example, I agree that Crawford would be better in the OF that Morse would be at SS,but barring an emergency in a single game, a swap like that would never happen. If Crawford goes on the DL, the Giants would just call up a SS from the minors to take his place — someone like Matt Duffy. They don’t have to worry about how Morse would do at SS.

2
sam
9 years ago

I think a possible explanation could be that, with the decline of offense, more of the overall available WAR in a given season is given to pitchers, resulting in a decline in the share of WAR being given to position players.

-1
Thanks, Comcast
9 years ago
Reply to  sam

The ratio actually stays the same every year; 57% of league WAR goes to position players and 43% goes to pitchers. There’s certainly a reasonable argument to be made that more of that should go to pitchers, though, given that the decline in offense is largely related to increased strikeout rates and a decrease in homeruns (both of which are independent of fielding).

More to the point, more strikeouts means fewer balls in play. That’s reflected by the UZR portion of WAR (as it’s cumulative), but not by the positional adjustments. It makes since to me that if the range of UZR decreases via fewer opportunities, you’d also adjust rein in the positional adjustments. In other words, 2014 David Ortiz would be less of a liability playing left field today than he would have been ten years ago because he’d now have fewer opportunities to bungle plays. The reverse also applies.

-1
Chris Mitchellmember
9 years ago
Reply to  Thanks, Comcast

You’re correct about the 57% piece. Whether or not that should change is another argument entirely.

That’s an interesting argument on the balls in play. Due to fewer missed opportunities, all poor defenders would look better. But a poor SS and CF would benefit more, making them more likely to clear the replacement-level threshold.

0
Thanks, Comcast
9 years ago

Do batted ball trends come into play at all? League wide GB/FB ratio is at its peak, and overall groundball% has steadily trended upward over the past decade. One could certainly envision that causing a shift (oh, and there’s that, too) in the relative value of infielders versus outfielders.

But I think Z’s above comment about the change in run environment is probably the best starting point for improvement.

1
Chris Mitchellmember
9 years ago
Reply to  Thanks, Comcast

Possibly, but more than anything, the trend seems to relate to come down to offensive positions versus defensive positions. CF still has the highest share, while LF and RF have some of the lowest.

0
Peter Jensen
9 years ago

Interesting research. Three possible explanations come to mind. Injuries, which Mike Newman suggested above. Easily testable using MLBs DL data base. Long term contracts. The front office has much more incentive to stick with a player with a long term guaranteed high dollar contract even when he falls below replacement level. Testable by looking at the percentage of below replacement level PAs at each position going to subs or going to players with multiyear contracts. And third, September callups. I’m hypothesizing that September callups may be given more PAs as DHs that at other positions. You could run your query again leaving out the month of September to see if you get a more even distribution.

0
Chris Mitchellmember
9 years ago
Reply to  Peter Jensen

I like your long-term contract idea — probably worth looking into. And I wonder if September call-ups (and the fact that the data excludes the final 3 weeks of the season) might partially explain why 2014’s such an outlier.

1
Cool WHIP
9 years ago

Forgive my ignorance, but why would a drop in offensive production necessitate a change in positional adjustment? Why would you alter a metric of defensive aptitude based on offense? Unless 1B or DH somehow become “more difficult” to field relative to other positions, I don’t understand why positional adjustments would explain a drop in PA shares of positive WAR players. Maybe the league’s DHs as a group are declining offensively, making it harder for them to crack a positive WAR.

Would anyone care to elaborate?

2
Cool WHIP
9 years ago
Reply to  Cool WHIP

Third sentence is meant to read: “Unless 1B or DH somehow become “more difficult” to field relative to other positions, I don’t understand why you might consider changing their positional adjustments.”

0
Chris Mitchellmember
9 years ago
Reply to  Cool WHIP

WAR is essentially made up of three components: offense, defense relative to position played, and the positional adjustment. The pos adjustments are meant to account for the fact that someone who can play a respectable SS is much more difficult to find — and therefore more valuable — than someone who can play a respectable 1B. If DH’s as a group decline offensively, while SS don’t, the offensive gap between the two positions is now smaller. But since the positional adjustments don’t change, a DH would still be penalized as if DH’s hadn’t declined. As a result, a higher share of them would check in below replacement level.

0
Cool WHIP
9 years ago
Reply to  Chris Mitchell

Hmm. I thought the positional adjustments were purely a way of weighing player value based on how difficult their respective positions were to field. At least that’s how I interpret Tom Tango http://www.insidethebook.com/ee/index.php/site/article/fielding_position_adjustments/

Even if DH offense declines, they are still just as incapable of manning a middle infield spot. I’m having a hard time shaking the intuition that they should still be “punished” with the positional adjustment. Is there something obvious I’m missing?

3
Chris Mitchellmember
9 years ago
Reply to  Cool WHIP

I appreciate your comments. I had never dug into this stuff so deeply before, and you’re really making me think here.

I guess more than anything I have an issue with the way the positional adjustments were calculated (if the method from that Tango article is in fact the standard). The idea is that it should never be “harder for them to crack a positive WAR.” By definition, the difficulty of a player’s cracking replacement level should never change. At least that’s how I see it right now. I’m open to other arguments.

1
Cool WHIP
9 years ago
Reply to  Cool WHIP

Yeah, I guess what I meant was that DHs, who derive their value exclusively through hitting and suffer accordingly via the positional adjustment, are going to struggle in scarcer run environments- like the one we have now. Almost all hitter WAR values will be suppressed, but heavier players like 1B and DH are going to suffer the most because of their weighted adjustment.

Ultimately if we are going to change positional adjustments, we need to do so across the board to preserve the relative distribution of positional weight. I think this is what “Z” was getting at above ^^^. (It just took me awhile to figure it out!)

If we shifted only the DH positional adjustment, it would diminish the value of the other position players

1
Paul Kasińskimember
9 years ago
Reply to  Chris Mitchell

The Positional Adjustments do change, every single year, relative to the ability of other players at that position…

0
Cool WHIP
9 years ago

Forgot to mention- thank you for the article. It has inspired quite a bit of thought already

2
Chris Mitchellmember
9 years ago
Reply to  Cool WHIP

In response to your latest comment. Ok yeah I see what you’re saying now. The crux of my argument is basically that — by the definition of replacement level — it shouldn’t be any harder for a team to find a replacement level DH than a replacement level SS. So in this new run environment, I that would mean raising the positional adjustment for DH and 1B and lowering it for SS and CF.

0
Word
9 years ago

This is really interesting, Chris. Mike makes an interesting point about the role of injuries as well — is it possible we see more replacement-level (and sub-replacement) ABs at positions like first base because more ABS are given to actual replacement players?

My only contribution would be to wonder if we need to expand the time frame, to make sure this isn’t skewed by the current batch of players. There’s the famous example of CFs in the 1950s. I doubt it’s the case, but it seems plausible that right now we just have more good players at the premium defensive positions than on the other side of the spectrum.

0
Chris Mitchellmember
9 years ago
Reply to  Word

Good idea on expanding the time frame. It’s on my to-do list. Unfortunately, downloading and combining CSV’s for every position every year can be a little time consuming.

0
Anon
9 years ago
Reply to  Chris Mitchell

If you are using Fangraphs, the split seasons functionality should ease the data collection. This option splits multiple years of data by year.

Also, some players are listed at multiple positions; I don’t know if that could skew the data.

0
Chris Mitchellmember
9 years ago
Reply to  Anon

Yes, but it doesn’t allow you to split seasons once you’re looking at splits (such as by position). Best I can tell, doing it one year and one position at a time is the only way.

1
Anon
9 years ago
Reply to  Anon

Ah, I was using positional eligibility rather than positional splits. (and using positional eligibility creates some data problems such as having the full season of 2012 Zobrist be included three times in the final data)

1
olethros
9 years ago

How much of that DH drop is due to NL teams playing in AL parks with a bench/utility guy as DH?

0
Chris Mitchellmember
9 years ago
Reply to  olethros

That could definitely be a part of it. When interleague games all happened at once, NL teams would have theoretically had more flexibility to call up a DH from the minors. I remember the Brewers doing this with Prince Fielder way back when.

0
DavidKB
9 years ago

First off, great article. Second off, I have a question. Are WAR values for people who play in multiple positions based on a weighted average of the positional adjustments, or based on the most valuable position they regularly play? To me it seems like it should be the latter, since a player’s WAR really shouldn’t depend on external factors (e.g. platooning, logjam, etc), which are a result of the team he plays on.

0
Chris Mitchellmember
9 years ago
Reply to  DavidKB

As far as I’m aware, they’re based on a weighted average of positional adjustments. I think this is how it should be. WAR is meant to measure how valuable a player was, not how valuable he could have been in an ideal scenario.

0
DavidKB
9 years ago
Reply to  Chris Mitchell

My response would be: why is WAR park adjusted then? I think the idea of WAR is to say: “How valuable would this guy’s performance be on *my* team?” That being true, you should try to filter out all aspects that depend on the team he’s on. Being played out of optimal position clearly falls into that category. For me, this is a significant shortcoming of the positional adjustment. The fact that DHs, 1Bs, and corner outfielders are “underperforming” recently could have a lot to do with increased platooning. If you counted players only in their optimal positions it would probably do a lot to put the numbers back in line with the recent past.

1
DUTCH4007
9 years ago

Thank you for doing this article! To me it makes complete sense that it would be harder to produce WAR as a DH then a position player. The replacement for a SS is the next best SS the replacement for a DH is the next best hitter independent of position. If DH’s were producing WAR at about the same rate as position players it would make me wonder.

1
Chris Mitchellmember
9 years ago
Reply to  DUTCH4007

In theory, a replacement-level DH should be just as easy to find as a replacement level SS. But the fact that teams are running out sub-replacement DH’s suggests that isn’t the case.

1
Ben Fenker
9 years ago

It may be worthwhile to break out the overall WAR into fielding/hitting WAR. This is a more reliable stat and would be a more accurate representation of the offensive quality of hitters at the different positions over time.

0
Billy
9 years ago
Reply to  Ben Fenker

I hope I’m not weighing in too late on your piece, Chris. This is a bit similar to a question I’ve had for a while about positional adjustments, though mine was more narrow in focus since I was concerned with only the outfield. I was concerned with how merely changing from CF to a corner OF spot (or vice versa) could make a defender look better or worse due to comparison to easier or tougher competition respectively. Theoretically, shouldn’t the two even out exactly overall and leave WAR unaltered when a player makes a switch? In my casual observation, that doesn’t seem to be happening.

Also, numbers like DRS and UZR are found by comparing the player’s defense to the average at the position. It does not give the TOTAL number of outs they made. I would imagine that (hitting adjustments aside) an out made in center is more or less as valuable as an out made in left or right. Maybe these totals could give a better idea of how valuable these positions are relative to each other.

Finally, there is a second possible issue with this relative measure. With only 30 or so primary players at each position, a change in the defensive quality of 3 or so of these players could dramatically affect the overall metrics since it may change the baseline. This would make the remaining players there look significantly better or worse.

This stuff fascinates me and I love this piece. It looks like we’re dealing with some moving targets here…

0
Chris Mitchellmember
9 years ago
Reply to  Billy

On your CF/COF question, a good defender would generally have a higher WAR in CF since he would get more defensive opportunities than he would in LF or RF. By that same token, a poor defender would have a higher WAR in a corner spot since he’d get fewer opportunities, and therefore have fewer chances to not convert a play into an out.

You’re very right on the sample size issue at each position, and that’s certainly a good argument for not changing the positional adjustments. Any and all year-to-year variation might just be random noise.

0
Billy
9 years ago
Reply to  Chris Mitchell

Heh, this is so incredibly complicated I don’t know where to start. And I know I’m going to leave plenty out, so sorry.

I’m going to start with the anecdotal evidence that hit me last year. Brett Gardner moved from Left to Center and saw his defensive numbers go from elite to average. Shane Victorino went from Center to right and saw his go from average to elite. For fun, let’s just throw in some of the numbers Gerrardo Parra put up in the corners.

Now, for the sake of simplicity, let’s make the obviously silly assumption that their abilities remained the same from one season to the other (a notion Gardner is showing is silly with a lackluster year defensively that shows perhaps decline was a factor).

What I’m getting at is that all these players had their WAR greatly bolstered by their defensive metrics while playing LF/RF, and it would subsequently dissolve while playing CF since the baseline for posting above average runs saved is so much higher there. I have my suspicions that this is because DRS and UZR are based on being compared to average at the position. What about total outs made per opportunity as a raw number like batting avg or obp as opposed to a scaled number like wRC+?

0
Chris Mitchellmember
9 years ago
Reply to  Ben Fenker

Yeah I agree. That’s definitely on my to-do list.

0
Paul Kasińskimember
9 years ago

Positional adjustments change every year based on the performance of players at those positions.

0
Chris Mitchellmember
9 years ago
Reply to  Paul Kasiński

Do they? Fangraphs’ glossary page on positional adjustments hasn’t been updated since 2008, and I don’t recall ever hearing about year-to-year changes. I’m not saying you’re wrong, but if they do, I’m not aware of it.

0