Tangotiger Blog

And frankly, the repl level value is going to be so close for all the league-leaders anyway, that this “WAR” talk really was a side issue.  We didn’t really need to talk about replacement level, and all discussion could have been fairly limited to comparing to the average.

(Though, in comparing nonpitchers to pitchers, you would DEFINITELY need to be introducing replacement level.)

For those asking why is the positional adjustment added to the fielding part and not the offensive part:

Because when you discuss players, say a prospect, on a 20-80 scale, you don’t say “at SS, he’s a 60, but in LF, he’s a 70”.  You would simply say “he’s a 65 fielder” or something like that.

And you know that to play SS, you really need to be at least a 50 fielder.

Similarly, you wouldn’t say that he’s got “30 power for a 1B, but 50 power for a SS”.  You’d simply say “he’s got 40 power”.

To the extent that you want to compare players across positions, there are better reasons to include the positional adjustment as part of his overall defense than as part of his overall offense.

It’s nice to see this (especially the defense) out front and easy to find. I had thought it was odd that the best fielding leaderboard was under the “Value” tab in the batting section where they had “Fld + Pos.” Meanwhile, in the fielding leaderboard it was sorted by UZR with no positional adjustment to be found. It’s great to see they’ve made getting at these numbers a lot easier. Now if B-ref would just take the positional adjustment out of oWAR…

—-To the extent that you want to compare players across positions, there are better reasons to include the positional adjustment as part of his overall defense than as part of his overall offense.
***************

Giving it only a few seconds thought, I am not sure that it wouldn’t be better—more reflective of how players are assigned to positions in MLB—to split the positional adjustment between offense and defense in some proportion. Maybe 2/3 defense and 1/3 offense, or some such.

To me, the answer has always been more straight forward. 

Everything you do when your team is batting is offense.  Everything you do when the other team is batting is defense.  You play your position when the other team is batting, so a position is defense, not offense.

Agree with Dave 100%. Think of how players are assigned positions. It has nothing to do with their offensive ability and everything to do with their defensive ability.

Well, my point was that it does have something to do with their offensive ability.  They know what the typical hitting profile is for each position,  and they do use that, along with the stronger factor of fielding ability, to assign a player’s position. It’s a smaller factor, which is why I guessed 1/3 . Just an idea anyway, no problem with me if it doesn’t hold water….

Right.  I’m not a fan of changing Mike Trout’s hitting value based on whether he is playing LF or CF, and especially if he goes from one position to the other mid-game.

I think it makes things unnecessary complicated.

Tango when you engineered WAR, was it your intent that the combined Positional Adjustments for all positions (including DH, PH and Pitcher Hitting) add to 0. If OFF and DEF is WAA, wouldn’t this have to be the case.

Darren: right, that’s the way it should be.

There are extra complications, such that it’s much harder to hit as a PH than as a starting player.  Basically, we’re asking “what would an average player do in this environment?”  PH is an environment, DH is an environment, Catcher is an environment.  Being a mid-game replacement is an environment. As it turns out, the other fielding positions are not environments.

I made a decision to split pitchers from nonpitchers, and so, pitchers-as-hitters is separated from the other nonpitchers-as-hitters.  In effect, I make that an environment.

Really, trying to get away from all the many “fit square peg in round hole” that you see so much, and try to keep things as honest as possible, in a way that we can actually explain verbally.

Thanks. I guess I would rather see Positional Adjustments and Environment Adjustments seperated. To me Positional Adjustments should add to zero as they all relate to each other. Right now on Fangraphs “DEF” adds to +220, which suggests there is 220 runs of Environment Adjustments ‘muddled’ into the positional adjustments. Now if that is the case, shouldnt there be -220 runs in Environment Adjustments that should be included in WAR. To me this should be a negative adjustment to a position player for not having to play in the more difficult environment of a DH, PH, Catcher or hitting pitcher. Equal debits and credits.

I am likely over thinking about 22 wins across an entire league, but I really like things when the neatly add up. Its a fault of mine.

No, I agree.

The DH “fielding value” should be -2.25 wins, not -1.75 wins.

That’s because the hitting difficulty accounts for +0.50 wins required.  That is, Ortiz is actually a better hitter than we know, because he has a sort of “platoon disadvantage”, always hitting as a DH.

You can make the same argument for catchers, that maybe their fielding value should be +1.0 wins and their hitting difficulty should be another +0.50 wins or something.

It’s not that the catcher is actually hitting with his full padding on, but it’s LIKE that’s what’s really happening.  So, if you have two players, one a LF and the other a C, and they both have a .350 wOBA, the C actually hit “better”, simply because of the “platoon disadvantage”.

You can go crazy and do the same for pitchers, that you would need to then look at “strength of opponent”, and that maybe Randy Johnson faces tougher hitters, because he’s only facing 12% LHH while all the other LHP face 22% (or whatever).

Or maybe RJ faces EASIER hitters as the managers overdo it by putting in really really bad RHH who couldn’t possibly be better than decent LHH.

All to say: there’s alot of work to do, so, I wouldn’t get hung up that it doesn’t add up.  Yet.  But, it should.

Unfortunately, I can’t be part of the final solution.

By the way, I’d *love* for an aspiring saberist to tackle the RJ platoon issue.  RJ is a unique talent and unique body, so, it’s a wonderful thing to be able to focus on.

Darren,  we actually do zero out RAA / WAA before turning it into WAR, so there’s not an extra 20 some wins floating around each season.  It’s just not reflected in Defense, which I think I’m ok with. 

Offense isn’t going to equal 0 either because the pitcher positional adjustment is going to offset pitcher’s negative batting.  If you just do non pitchers it’s going to be much closer to 0, but not exactly for various reasons.

Thanks Dave. However, the pitcher hitting adjustment is listed under DEF. It has to be, as no pitcher (under Leaders) is showing a negative DEF. Kershaw for instance is +8.1 DEF, and -7.1 OFF. Since I dont think Fangraphs WAR includes pitcher defence, shouldnt Kershaw’s DEF be 0, and his OFF should be +1.0.

Right now total OFF is -740.0. I guess what your saying is that if you moved the Pitcher Adjustment into OFF, you are going to be very close to 0 (as will DEF).

Picky ask I know, but it would make it less confusing. Thanks.

I agree that pitchers are their own universe, and their hitting should be compared to the hitting of other pitchers.

I draw a clear distinct of pitchers as their group and nonpitchers as their own.  I don’t see P as distinct to CF as CF is distinct to LF.  While all of them are colloquially “positions”, I treat P and non-P as two position groups, and CF and LF are two “roles” within the non-P position group.

Much like SP and RP are two different roles within the pitcher group.

We simply have to acknowledge the fluidity of the three OF roles.

(You COULD try to make the argument that the three IF roles should be treated as a separate position group from the three OF roles.  In my view, doing so adds complexity and it’s not clear that doing so actually results in a net benefit.)

I don’t really have an issue moving the pitcher positional adjustment into batting and then leaving the pitcher positional adjustment = 0. Assuming that’s what’s being suggested.

I’m not sure how quick a change that is though, I will need to check whether it can be done in the next week, or if it’s an off-season project.

Right, the Def should either be zero, or his actual pitching value.

The Off for pitchers should be centered around 0.

Really though, it would be based on people use the leaderboards.  If you include both pitchers and nonpitchers in the same leaderboard, then we wouldn’t want to do what is being suggested (centering pitcher hitting to zero).

That would be like centering RB and QB running yards to their respective averages.

But, if there is no overlap between the pitchers and nonpitchers in the leaderboards, then we’d like to see the best hitting pitcher at the top of the list, and not be a negative.

For example:
http://www.fangraphs.com/leaders.aspx?pos=p&stats=bat&lg=all&qual=0&type=8&season=2013&month=0&season1=2013&ind=0&team=0&rost=0&age=0&filter;=&players=0&sort=19,d

#1 is Greinke, which is fine with his 69 PA, as he’d end up being #1 no matter what system you use.

But #2 is Zack Duke, who went 1-1 (single).

Other than Alvarez, the top players are all going to be guys who went 1-1, or 0-0, since having +0 runs above average (as the 0-0 is going to be) will be better than a pitcher with a .300 wOBA on 50 PA, since .300 is less than the league average.

So, the answer is completely dependent on whether you keep P as part of ALL players on the leaderboards, or keep them as their own universe.