An Intra-Divisional Trade; Debates on Wells and WAR Continue
It’s early February, and the M’s equipment truck has headed south towards Peoria. Pitchers and catchers report in about a week, and there’s plenty to discuss now that football’s finished.
1: Today the Astros and A’s completed a five-player trade, with Oakland receiving SS/DL Jed Lowrie and RP Fernando Rodriguez in exchange for 1B/DH Chris Carter, RHP Brad Peacock and C Max Stassi.
Lowrie gives the A’s a legitimate bat at any infield position, and since he’s a switch-hitter, he provides a great deal of flexibility for Oakland. While he’s been much better from the right side in his career, he credits Astros hitting coach Mike Barnett with a mechanical tweak that helped him post an .820 OPS from the left side in 2012. His career’s been hampered more by injury than platoon splits, however; he’s yet to crack 100 games played in his five year career. Still, regular rest could potentially keep his bat in the line-up, either as a partner with A’s starting SS Hiroyuki Nakajima or at 3B with Josh Donaldson. The pick-up gives the A’s another bat to pair with their pitching depth, and another solid pick-up for a team that won the division despite its SS and 3Bs combining to post a .275 OBP in 2012 (Brendan Ryan’s 2012 OBP was .277).
For their part, the Astros strengthened their 1B/DH position. The Astros haven’t had a DH of course, so picking up a slow slugger makes some more sense for them than most other clubs. Their depth chart currently lists Carlos Pena and Brett Wallace as the incumbent DH/1B, both of whom are lefties, and both of whom come with quite a few question marks. Pena posted a sub-.700 OPS last year, and Brett Wallace has bounced between AAA and MLB for three seasons. Houston gets a cost-controlled player who may have had a breakout season in 2012, and who hasn’t shown much in the way of platoon splits in either MLB or MiLB. Brad Peacock gives some much-needed depth to Houston’s rotation, and could allow the club to keep top pitching prospect Jarred Cosart in the minors. To be fair, Peacock was awful last year in the Pacific Coast League, but he had a cup of coffee with Washington in 2011 and could eat some innings in 2013 free from the pressures of a pennant race. Max Stassi’s a great catch-and-throw catcher who’s had his own problems with injuries. Shoulder tendinitis has plagued him since high school, and kept him out for over a year between 2011-12.
The A’s are strengthening a team that made an improbable run last season, and the Astros are smartly stashing prospects and depth for two or three years down the line. From an M’s point of view, the trade highlights two things, neither of which is terribly comforting. First, the M’s can’t simply assume the A’s are a fluke who lucked their way into the postseason. The team used platoons and cast-offs to mitigate some black holes in their line-up, but they’ve taken steps to improve in the offseason, first by picking up Japanese SS Nakajima and now with Lowrie. Their pitching can regress (and it probably will), but their offense could offset the loss a bit. Second, the Astros are lovably terrible right now, but they won’t be for long. This doesn’t appear to be a Royals/Pirates situation. Like the M’s, they’ve quietly restocked the farm (though they’re far behind the M’s in that department), and they’ve made enough minor moves that suggest they won’t be perennial 100-loss threats for long.
2: Dave’s great article on WAR over at Fangraphs is a great read, and it’s helped clarify my own thinking on the topic. Then, a great twitter conversation between Dave Studenmund and Colin Wyers helped me understand where so much of the heat in this (ultimately silly) argument comes from. Dave (er, Cameron) is clearly right that WAR is transparent about its intentions – it’s attempting to measure a player’s value, as comprehensively as possible. It’s answering a question that’s among the most-asked in baseball (“how good is this guy?”), and essentially no other statistic does a really good job of this. We can talk about offensive stats which shed considerable light on how good a hitter is. But we all understand, maybe subconsciously, that this isn’t the complete picture.
But there’s a problem: reducing value to a single number obscures *how* a player produced that value. We’re so focused on offensive statistics for position players that Brendan Ryan being an above-average player often feels wrong. Sure, sure, you know he can pick it, and that there’s tangible value in that, but…he put up SLG and OBP numbers under .300. WAR is counting that, of course, but I think people aren’t used to seeing defense portrayed on the same scale as batting, and many who ARE dispute the value of the defensive inputs to WAR. For what it’s worth, I’m somewhat sympathetic to the argument, as so many baseball discussions growing up revolved around specific skills – Edgar’s eye, Jim Rice’s power, Nolan Ryan’s strikeouts, Jack Morris’ dependability, etc. – and we wound up conflating those skills with value. People didn’t engage in a discussion about Rice’s overall value because they felt that WAR short-changed how terrifying that power was in the offensive environment of the 80s. It’s harder to talk about a transcendent skill when you’re lumping everything together. To be clear, I think the view that skills trump value is incorrect, but I understand it, and I’m guessing it’s part of what underlies this ongoing battle over WAR.
Further, there’s no doubt that some people use WAR without really understanding what they’re using or why one measure differs from another. To those used to focusing solely on unambiguous measures like counting stats (“player Y hit 500 home runs”) or batting average, the fact that there can be such dramatically different interpretations/frameworks looks like evidence of serious flaws. It’s not, and the fact that you can use a different defensive stat or replacement level is great, but I think sometimes people argue that WAR is an argument-ender without being able to explain how and what it’s doing. To some, that may make it look like stat-savvy fans aren’t thinking critically. I tend to think this is overblown a bit, and that WAR was used pretty effectively in the huge Trout vs. Cabrera MVP debate, but I definitely concede that there’s often a lot of heat and not enough light from both sides of the traditional/sabermetric fight.
3: Matthew Carruth had a great piece on Casper Wells at Lookout Landing that’s generated a lot of debate. Wells (and to a lesser extent Eric Thames) seems to represent a case where the M’s front office really has changed its approach – something else that M’s fans have been debating since Dave’s post on the front office changes.
I’d certainly quibble with the probative value of Wells’ 34-game stint as an everyday player* but let’s assume he’s not. The M’s acquired several skilled players with large observed platoon splits in the past few years – Jaso, Wells, Thames, etc. Following the trade of Jaso and the acquisition of Ibanez, Bay and Mike Morse, it’s pretty clear that Wells and Thames don’t really have a spot. The M’s acquired Jaso/Wells because they were focused on what they COULD do, and they’ve moved on because of an almost obsessive focus on what they CAN’T do. It’s no surprise given the Jaso trade that Oakland is perhaps the best example of a team that doesn’t worry excessively about the label of “full-time” or “part-time” player. The A’s use of platoons in LF, 1B and DH allowed them to get above-average production from a group of cheap/flawed players.
Let’s be clear: the M’s are allowed to change course. After finishing last three times in a row, they *should* re-examine some of their strategies. If they no longer want to target players like these, that’s fine (though picking up Paulino/Shoppach/Ibanez suggests they don’t mind it occasionally), but there’s clearly a cost to the team: it makes the moves they’ve made in recent years highly inefficient. They could still get value for someone like Wells (maybe from the A’s), and given the injury history of many of the new M’s, he could end up starting for them at some point in 2013. Still, the offensively-challenged M’s have dumped/frozen out players who could help the offense, *even if you assume* they could only do so in limited duty.
Comments
16 Responses to “An Intra-Divisional Trade; Debates on Wells and WAR Continue”
Leave a Reply
You must be logged in to post a comment.
Fun fact: there are 20 individual players set to earn more this year than the entire Astros roster
The Mariners have built their team around Eric Wedge.
Sad that they let him dictate players they acquire and retain.
I have really enjoyed this whole sequence of discussion about WAR, including Marc’s addition here. A thought I have is that two players of similar WAR but with different skill sets could have different applicable value to different teams, based on need. The most clear example would be a team planning to platoon and already having a lefty masher with a choice between another lefty masher and a righty masher of equal WAR value. If the team chose a second lefty masher, there would be an increase in totaling up individual WAR, but team WAR would not increase because of redundancy.
So say for the Mariners, if there was a choice between Ryan or Jeter at shortstop, first judging their total WAR value to be equal, would there be any discernible difference in team value choosing one over the other? Or another way of putting it, is team WAR slightly different than simply adding up individual player WAR?
Great read, Marc. You made a very good point that I hadn’t heard made before… how the apparent change in direction may also have set the team back a bit.
I do think that, most of the time anyway, the arguments coming from folks against WAR usually amount to them fundamentally not believing that defense is particularly important. So many times I saw people refuse to accept that Cabrera’s bad defense gives back a non-trivial amount of the substantial offensive value he provides, for example. It’s not even about advanced stats, really – they see Miggy and Prince stumble and bumble their way around the infield, and they figure it doesn’t really matter how bad they are with a glove on. They mash, and that’s all that matters.
In that WAR discussion, I see only one person mention that UZR’s three-season sample-size-requirement for stability undermines WAR’s accuracy. In effect, the addition of single-season dWAR to WAR seems to me to introduce too much noise to accept that WAR is really a counting stat (if it is in fact literally a counting stat, since using linear weights seems to make it not quite that either, not in the sense that runs, doubles, or RBIs would be). The question for me, though, is that of UZR added to WAR in a single season. Everyone acts like posing this question means the poser (heh) doesn’t like defense or understand the stat. Is it possible to state a non-snarky, knowledgeable justification for using single-season UZR? Besides saying “it’s all we’ve got.” Somebody using one-third of a season of PAs to calculate a single season OBP would not be taken seriously, so why is it unreasonable to think present WAR calculations are seriously flawed?
One of the counter-arguments in that thread, for example, pointed out that batting stats vary season-to-season for individual players just as much as defensive stats. But since those batting stats are based on events whose noise-distortions are understood to essentially disappear within a single season, it doesn’t address the question, since UZR is NOT.
When I started reading USSM and fangraphs a few years ago, I thought WAR was a great stat and a useful tool for understanding roster contruction and contract issues. Including UZR seems to me to have seriously reduced its utility.
Just like Dave said in his article, WAR is NOT perfect and should be used to START the conversation not end it. Because it incorporates one season of UZR does not make it useless…and I agree that’s probably WAR’s biggest problem. But, I think even with the inclusion of UZR, we can all agree it’s better than rattling off batting average, RBIs, win-loss record, and World Series rings as the measuring stick to “how good is this guy?”
A single season of any statistic introduces noise, not just UZR. The more data you have to work with, the better idea you have of what that player will be able to provide going forward.
UZR for a single season in WAR still tells you what value he provided. Whether that player is as good as his single season stat for UZR or any other stat is always in question.
Ummm…can we get a report on this Miami PED clinic having Montero’s name? And what the potential fallout could be?
“UZR for a single season still tells you what value he provided”
vs.
Tango: “Roughly speaking, 200 PA as a batter (say 50 games) tells you as much as 400 balls in play (BIP) as a fielder (say 100 games). So, if you can make a judgement on a hitter’s batting stats after one year, you can have that same level of uncertainty using UZR after 2 years… Don’t talk to me unless you are talking about at least two years of data.”
from:
http://www.insidethebook.com/ee/index.php/site/article/mike_silva_chronicles_part_1_uzr/
Again, isn’t using one year of UZR simply adding a level of unnecessary uncertainty to WAR?
I definitely think that I’ve observed WAR being used to end more discussions than start them, so maybe, given the tendency to claim things like “Mike Trout had the most valuable season of any rookie for twenty years,” we need to change how WAR includes defensive metrics?
My understanding, based on comments like tango’s above, is that one season of UZR simply does not tell us the value a player provides.
I certainly don’t neccessarily disagree with what you’re saying, naufrago, and definitely wouldn’t disagree with Mr. Tango. After reading the two articles, along with all the comments, I wouldn’t have a problem with using a 2 or 3 year average for UZR in the WAR formula. Of course for second and third year players, a “fudged” average UZR would have to be used. Again, it still wouldn’t be perfect, but we don’t have perfect defensive stats yet. And I still say it isn’t a useless stat, it’s the start of the conversation.
Agreed that “useless” is definitely going to far w/r/t to dWAR – it’s easy enough to believe that Trout was better than Cabrera, after all. The way that WAR produces a solid, precise number just seems to mask the uncertainty that including dWAR introduces, and I don’t think there was nearly as much uncertainty before the defensive stats were added (which could be me).
Some kind of rolling 2 or 3 year average does seem to be the way to go, to me…
One season of WAR is never a great indicator of what a player will do in future seasons because of swings in BABIP and defense. Defense may be less certain, but some indication of what a player can do is better than nothing.
WAR has never been a good predictor of future value and that was never the purpose of the stat to begin with.
I’m not talking about future value. As I understand it, we need three years of UZR data because of the way that decisions are made about whether a play was made and the way that data is collected, in addition to the way we want to hedge our bets about talent evalutation and future value like we would when looking at a year of batting stats.
If noise is half or more of a sample (one year of UZR), I don’t think that sample can reliably tell us what happened on the field.
There’s no magic point at year three at which UZR becomes reliable and trustworthy. You’re taking the suggestion of using larger sample sizes for defensive metrics to make assumptions about true talent level out of context.
UZR is basically like ERA – it includes noise, it includes teammates contributions, it includes subjective decisions made by official scorers, and it can be influenced by parks and opponents in ways that aren’t obvious. Do you demand three years of ERA data?
UZR being like ERA is a good analogy, with “error bars” that surround the stat due to those factors you mention – I’d think the most significant are teamate contributions and official scorers. But isn’t that itself a good reason to question the way it’s included in WAR, a stat that tries to eliminate those types of factors in producing a number for individual player evaluation?
Using one year of UZR seems like using ERA instead of FIP. “Magic point” is snark, not substance. It fails to address the question of whether dWAR is really better than nothing as it is presently applied, especially because WAR gets used most often as if its an extremely accurate measure of player value (perhaps we have a screwdriver that’s shaped like a hammer?). Three years would be better than one, and one year of defensive data still seems to me to be a bad choice that reduces the quality of WAR.
Just throwing a wild suggestion out there, maybe it would be interesting to have a single-number context-laden counting stat (RBIs, runs, ERA, etc,), as well as a DIPS/linear weights-type stat (WAR) for purposes of comparison? Well, maybe not so useful. But if we imagine creating such a stat, wouldn’t UZR more properly belong as a component of that, than as a component of a stat that tries to remove context (teamates, scorers) from its calculation of player value?