Article: Mailbag: Buxton’s Leash, Likely Lineup, Ticket Sales

[Otto von Ballpark]

Yup, that's sorta my point. Despite the fact that the positional adjustment sorta makes it apples and oranges (or, at the very least, weakens the comparison considerably) - the stat is still widely used that way.

How does the positional adjustment "considerably weaken" cross-positional comparisons with WAR? Cross-positional comparison using WAR is impossible without the positional adjustment.

[TheLeviathan]

 

How does the positional adjustment "considerably weaken" cross-positional comparisons with WAR? Cross-positional comparison using WAR is impossible without the positional adjustment.

 

The fact that it's impossible without it in no way weakens my point.  The problem is the nature of the adjustment itself.

[Otto von Ballpark]

When I said the vast majority of people use it more forcefully than it is warranted, I'm speaking about just about every time it's used on Fangraphs.  MVP arguments are rarely made with the very nuances argued in that link.

 

For example, rarely when there is a large disparity (say, 6.0 vs. 3.0) is any mention made of how much the defensive component makes up each player's total. Certainly, here on this forum with Byron Buxton, this is true but you also see it on Fangraphs.  I appreciate how they try and caution the use of the stat and they do a nice job laying out the limits of it, I wish they followed those outlines more frequently to set a better example.

I think context matters. An MVP discussion isn't that important to require all sorts of caveats about using WAR. It's a simple fan ranking, the digital version of bleacher discussions since the game began. WAR has just finally allowed those discussions to gain a little complexity and nuance beyond the previously irrefutable "he looks good" or "he doesn't look good enough". I think using WAR in that context is perfectly appropriate, even if WAR isn't perfect (an impossible standard anyway).

 

In the context of recommending or judging transactions, usually a little more nuance is called for, and most often it is provided. Those discussions often use projections, which by definition smooth out some of the rough edges of WAR measurement. Like the Realmuto discussion here the other day -- we were in talking in terms of WAR, but no one was simply taking Realmuto's 2018 WAR as gospel or anything. It was just one data point feeding his 2019 projection, which in turn was balanced by other factors and even presented as a range of possibilities rather than anything certain. And again -- this is a tool for the fan's toolbox. We knew the front office isn't pouring over Realmuto's WAR totals -- there are a million other factors which they are looking at. But those factors are often either off-limits or indecipherable to fans, so we simply use the tools we can -- WAR, projections, comps -- to guide our discussions, often as a "sanity check". Some posters really want Realmuto, but wonder if the prospect cost is too steep, or they wonder if a SP might be a better investment, and they can use WAR and projections to get an idea of parameters to discuss those questions in a way that fans never could before.

 

Also, FWIW, it seems to me that Fangraphs isn't shy about noting when defense makes up an abnormally large part of someone's WAR. Maybe not in every list or ranking, but when they specifically discuss Buxton's 2017, or Kiermaier, or even Bryce Harper's 2018, the defensive component of WAR is usually highlighted. Sometimes they stand by their measurement, sometimes not -- but noting it at least allows the reader to factor that into their own opinion.

[USAFChief]

 

How does the positional adjustment "considerably weaken" cross-positional comparisons with WAR? Cross-positional comparison using WAR is impossible without the positional adjustment.

The positional adjustment asserts that a run scored, or prevented, by a center fielder is worth more than a run scored, or prevented, by a first baseman.

 

That's preposterous.

[Mike Sixel]

The positional adjustment asserts that a run scored, or prevented, by a center fielder is worth more than a run scored, or prevented, by a first baseman.

 

That's preposterous.

Given that first base, DH, types barely get paid, and up the middle players get more and aren't out of work as much, I'd hypothesize that teams place more value, and find those players harder too replace....

 

Which is the point. How hard are they to replace? It's much harder to replace up the middle defenders than first basemen..... Hence, they get an adjustment in the calculation.

[Otto von Ballpark]

The fact that it's impossible without it in no way weakens my point. The problem is the nature of the adjustment itself.

I discussed this upthread. What exactly is the problem that you see? The positional adjustment isn't perfect (again, an impossible standard), but how teams deploy resources across the 9 positions affects the value they receive from those resources. WAR and its related discussions are better off incorporating a calculated estimate of that value rather than ignoring it, or relying solely on an "eye test" to do it.

[TheLeviathan]

 

I discussed this upthread. What exactly is the problem that you see? The positional adjustment isn't perfect (again, an impossible standard), but how teams deploy resources across the 9 positions affects the value they receive from those resources. WAR and its related discussions are better off incorporating a calculated estimate of that value rather than ignoring it, or relying solely on an "eye test" to do it.

 

 Seems like you laid out the problems just fine.

 

Sure, it makes sense to incorporate it, but remember all those problems when you use the stat too.  Your convictions behind the stat should reflect the flaws in the ingredients.

 

If I bake muffins I don't get to frame them as healthy because they have blueberries as I conveniently forgot the tub of sugar i put in it too.

[Otto von Ballpark]

The positional adjustment asserts that a run scored, or prevented, by a center fielder is worth more than a run scored, or prevented, by a first baseman.

 

That's preposterous.

That's not quite what it says. True, a run is a run is a run on the scoreboard -- but over the long run, it's worth more to a *team* to score or prevent runs at positions where the talent to do so is more scarce. That is what WAR is estimating, to complement what we know from the daily scoreboard.

 

The defensive spectrum has been known in some form or other since the game began. The skill of playing as an average defensive CF, SS, or especially catcher is much more rare than the skill of playing as an average defensive 1B (or god forbid, DH!). And with that rarer defensive skill, it also means that, on average, it is slightly harder to find an average MLB bat at those positions too. Historical data bears this out.

 

So a team isn't capturing as much value from Ernie Banks at 1B as compared to SS, or Joe Mauer at 1B compared to C. WAR is reflecting that. And baseball teams are bunched close enough together in talent that I guarantee front offices wouldn't shrug it off as "a run is a run is a run" if they can help it -- they are going to want to reallocate those resources to improve their team relative to their competition.

 

WAR isn't necessarily an absolute measure of a player's performance. It often can be used that way, because teams generally don't play guys out of position too much or for too long -- but it's important to keep in mind that it is really a measure of a player's value within a team, or their value as deployed by a team. I'll gladly include an asterisk if I have reason to discuss Darin Erstad's WAR at 1B, but that doesn't mean WAR is good for absolutely nothing (say it again).

[Mike Sixel]

Buxton is less valuable at 1B, than in CF......the defensive adjustment shows that. You can easily replace Cron/Austin's defense with the other (or Kepler, or Sano or lots of players), but how do you easily replace the defense Buxton supplies? Certainly not with Cron or Austin........

[Otto von Ballpark]

 

 Seems like you laid out the problems just fine.

 

Sure, it makes sense to incorporate it, but remember all those problems when you use the stat too.  Your convictions behind the stat should reflect the flaws in the ingredients.

 

If I bake muffins I don't get to frame them as healthy because they have blueberries as I conveniently forgot the tub of sugar i put in it too.

 

To that I'd say: how much are the proponents of WAR actually trying to frame their muffins as "healthy" -- versus how much are skeptics simply seeing that as an excuse to pick it apart? In this very thread, we've seen someone dismiss WAR because "there's no way Billy Hamilton can be equal to Mike Trout" -- nevermind that WAR doesn't say that, and no one said it. And other sentiments that suggest Daniel Palka might be unfairly underrated by WAR -- when his entire profile (traditional stats and scouting) seems to match his 2018 WAR exactly, as a one-dimensional, fringe MLB regular.

 

I am not trying to pick on anyone, or be combative -- I genuinely enjoy discussions on this site, I wouldn't be typing these long screeds if I didn't. And I think that WAR is a tool that can add depth to those discussions. So I really want to know if and when people perceive that WAR is being used inappropriately, so I can understand their concerns and try to improve my usage of WAR or even just my explanations and interpretations. Perhaps this thread has run its course to that end, but I hope you keep this in mind in future discussions.

[jmlease1]

 

That's not quite what it says. True, a run is a run is a run on the scoreboard -- but over the long run, it's worth more to a *team* to score or prevent runs at positions where the talent to do so is more scarce. That is what WAR is estimating, to complement what we know from the daily scoreboard.

The defensive spectrum has been known in some form or other since the game began. The skill of playing as an average defensive CF, SS, or especially catcher is much more rare than the skill of playing as an average defensive 1B (or god forbid, DH!). And with that rarer defensive skill, it also means that, on average, it is slightly harder to find an average MLB bat at those positions too. Historical data bears this out.

So a team isn't capturing as much value from Ernie Banks at 1B as compared to SS, or Joe Mauer at 1B compared to C. WAR is reflecting that. And baseball teams are bunched close enough together in talent that I guarantee front offices wouldn't shrug it off as "a run is a run is a run" if they can help it -- they are going to want to reallocate those resources to improve their team relative to their competition.

WAR isn't necessarily an absolute measure of a player's performance. It often can be used that way, because teams generally don't play guys out of position too much or for too long -- but it's important to keep in mind that it is really a measure of a player's value within a team, or their value as deployed by a team. I'll gladly include an asterisk if I have reason to discuss Darin Erstad's WAR at 1B, but that doesn't mean WAR is good for absolutely nothing (say it again).

 

This. but it's also noting that the defensive impact on run prevention at some positions is larger than others through sheer opportunity as well. there are simply fewer opportunities to make run-saving defensive plays at 1B than at CF. WAR is a useful way of summarizing a players total contributions to the team in a way that's fairly easy to understand, and in doing so it recognizes that certain positions are more valuable than others. That's all the positional adjustment really does.

 

(it's also important to note that fWAR and bWAR are very different estimators for pitching; fWAR tells you how good they think a pitcher should have been, attempting to null out luck factors. bWAR tells you what they actually accomplished for the team in that year. fWAR may be better for predicting future success. bWAR is more descriptive of what actually occurred.)

 

It's a useful tool, and it's nice to have a stat that tries to bring in everything about a player. It's not the end all be all, but it's a useful starting point for any discussion about a player's value. To my mind, if you have to argue that a player's WAR doesn't mean anything, then you have a weak argument.

[TheLeviathan]

I'm not saying you are doing anything wrong spycake, but I do question the way the stat is used broadly.  But best left to other threads.

[Halsey Hall]

Those familiar with Hammond Stadium know they put up photos up on the stadium above the ticket office.  So, some need to be adjusted.  This year Santana, Mauer and Buxton came down.  Sano was going up.  Just how short is the leash?