22 Responses to “39+ Games As DH With WPA >= .15”

1. fordham'13 Says:
   January 10th, 2011 at 5:13 pm

   It's strange to me that Edgar Martinez's career dWAR is 0.3.

2. Johnny Twisto Says:
   January 10th, 2011 at 5:20 pm

   That's because oWAR and dWAR are not split up correctly. The positional adjustment (which includes the "penalty" for being a DH) is included in oWAR. It should be with dWAR. Sean disagrees.

3. Kahuna Tuna Says:
   January 10th, 2011 at 5:24 pm

   Ortiz: 183 matching games out of 1,272 games played at DH. Martinez: 183 matching games out of 1,412 games played at DH.

4. Thomas Says:
   January 10th, 2011 at 6:00 pm

   wait... there's a WAR 'penalty' for playing DH?

5. Johnny Twisto Says:
   January 10th, 2011 at 6:05 pm

   It's not a penalty, but some describe it as such.

   It's a comparison of the inherent value in playing different positions. A DH has no defensive value. A catcher has a lot. WAR estimates the difference between them being worth about 25 runs over a full season.

6. Tweets that mention 39+ Games As DH With WPA >= .15 » Baseball-Reference Blog » Blog Archive -- Topsy.com Says:
   January 10th, 2011 at 8:26 pm

   [...] This post was mentioned on Twitter by kevin bryant, Baseball Reference. Baseball Reference said: B-R Blog: 39+ Games As DH With WPA >= .15 http://bit.ly/f2YsR6 [...]

7. DavidRF Says:
   January 10th, 2011 at 8:49 pm

   Its more of a "positional scarcity" as I see it. Its the "Rpos" column in the WAR tables. That column is negative for the "bat" positions (3-7-9-D), but Johnny is right, you could think of it as a positional adjustment to the offensive replacement level contribution ("Rrep"). I could be wrong, but I believe Rrep + Rpos is always positive.

   Its possible that defense is not involved at all in the calculation of "Rpos"... that's what the phrase "positional scarcity" implies. Its not that catchers provide more defense, its that replacement-level catchers are worse hitters than replacement-level DH's.

   During the season, there was talk of having some "Nuts and Bolts of WAR Calcuation" articles during the offseason. I don't know if there is still time for that, but those would be very informative.

8. Johnny Twisto Says:
   January 10th, 2011 at 10:39 pm

   I guess it's somewhat semantics, David. Yes, catchers are worse hitters than DHs (etc), but the reason is because catcher, SS, etc require more defense. Fewer people are capable of playing the harder positions (plus, to an extent, the positions are more wearing), so fewer big bats play them.

   One does not necessarily have to involve actual defensive performance in calculating the difference between positions, but I think B-R WAR does. I think it looks at the difference in performance between all the players who appear at multiple positions.

   To me, one's defensive value is dependent on the position they play (the "rPos") and how well they play it ("rField"). That's why I think those should both be included in dWAR. No matter how poor a fielder you think Derek Jeter has been, it's absurd to think Edgar Martinez was worth 14 more wins than him defensively. But that's what dWAR as currently presented says.

9. DavidRF Says:
   January 11th, 2011 at 11:13 am

   Yeah... the baseline for dWAR is "average" while the baseline for oWAR is "positionally-adjusted replacement". There's a descrepancy there. dWAR is really dWAA. Rpos and Rrep have to go somewhere, though. I suppose you could split it into three by splitting split Rpos + Rrep out. Then you would have oWAA, dWAA and some sort of positional durability number.

   Also you could make a case that Derek Jeter did cost indeed his teams more runs defensively over who they could have been playing at SS than Edgar Martinez cost his teams defensively over who they could be playing at DH-1B-3B. Its not saying that Edgar Martinez could jump in and play SS better than Derek Jeter, that's is indeed absurd. Its just saying that there is no shortage of average-fielding SS's in AAA and that Jeter's positional value is that he can hit.

   I think that's also why there's so many defensive leaderboards. It doesn't make sense to compare Total Zone Runs between different positions. That's exactly what's being done when you compare dWAR between two different positions. I believe dWAR is just TZR/RunsPerWin.

   You're argument of putting the positional adjustment in the defense side has merits too, I'm just making the case for the other side.

10. John Says:
    January 11th, 2011 at 11:50 am

    @Johnny, #2:

    The positional difference should be offensive. Take Mike Piazza for example.

    His value as a catcher had nothing to do with his defensive abilities, but because of the offensive value he provided given the scarcity of production from that position.

    He's going to be a first ballot HOFer. If he put up those same numbers as a 1B, he most likely wouldn't make it.

    Or, if you prefer, it's the difference between having Eric Karros and a AAA catcher in your lineup.

11. Johnny Twisto Says:
    January 11th, 2011 at 12:18 pm

    But "the offensive value he provided given the scarcity of production from that position" is combining offense and defense. Well, that's what WAR already does. If you are going to break WAR down into its components, offense and defense should be separated, not leaking into each other. Piazza's value as a hitter had nothing to do with his defensive abilities or his defensive position either. His value as a *player* did. He didn't bat as a catcher. If his batting numbers were those of a decent 1B, that's how he was pitched. He wasn't any more dangerous at the plate just because he was so much better than other catchers. His value as a catcher was in his ability to play catcher (whether or not he played it well). Carlos Delgado didn't have the ability to play catcher, so he was moved to 1B. Piazza did. Even if he couldn't throw anyone out, that's defensive value.

    Tango recently wrote about this: http://www.insidethebook.com/ee/index.php/site/article/reader_mail_of_the_day/

12. topper009 Says:
    January 11th, 2011 at 4:29 pm

    I have the best solution to this entire problem: eliminate the DH

13. kds Says:
    January 11th, 2011 at 4:40 pm

    The position adjustment does not come from offense and sould not be measured by offense. For some years in '50's CF's (Willie, Mickey and the Duke), hit better than RF. But it would have been absurd to give a better Rpos to RF. Very few of the RF could have played acceptable CF. All of the CF's would have been good in right. Suppose you took 1000 players who could at least field well enough to play 1B. Maybe 100 of them could field well enough to play SS. The SS would be better fielders at first than the IBmen, and the IB's would be much worse at SS. We measure the defensive difference to get Rpos. Because there are 1000 choices in the 1B group and only 100 of those in the SS group, we expect the 1B group to hit better, at the top and on average. That is positional scarcity. In practice the difference in offense, measured in runs, should be about the same as the difference in defense, so it can be used as an approximation, but the difference in defense is what matters. One would expect them to be close, otherwise the market is inefficient and GM's are making lousy decisions.

    For various technical reasons we often measure offense from average, but we could measure it from zero. So that a sum of the individuals would approximate the actual team runs scored. We can say that a good batter adds 100 runs to his teams offense, an average one 80, and a poor hitter 60. We have no way of doing this for defense. (Aside from the question; is it pitching or defense?) A SS fields a lot more balls than a 1B, and most of what a 1B does, catch throws, any other player at that level could do as well. So a SS is clearly worth more on defense than a 1B. Even a very bad SS and a very good 1B. Rpos is our best estimate of the average difference in defensive worth among the positions. (Including DH, another headache, but this post is too long already!)

14. Bastaducci Says:
    January 11th, 2011 at 7:29 pm

    Defensive war really is useless. I seen a post on her where it had Doug Flynn as a bad defensive player. that is all I need to know about defensive war. I mean I do respect that they are trying to put a number on defense and hopefully somehow they can improve it and make it worth looking at. but as of now it has no relevance.

    Now offensive WAR I love although it too needs to be tinkered with. right now WAR has hitters a little high and Pitchers a little low. take the yankees last year..I love Cano but there is no way he was worth more to the Yanks than Sabathia.

15. DavidRF Says:
    January 11th, 2011 at 9:04 pm

    @13
    Good point on Rpos and bunching of talent at positions (e.g. 1950s CF).

    Spot-checking 1950s data, they don't appear to be making the big mistakes that you are worried about. For example, in the 1955 NL, the highest offense position was CF thanks to Mays & Snider (and Ashburn & Bell). But they still give a higher Rpos to CF than they do LF-RF-1B. I don't know if they do that by checking the replacement level (i.e. ignoring the stars) or doing multi-year averages or a sanity check for obvious defensive spectrum ordering or combinations of the above. Anyone know where to find algorithmic details?

16. kds Says:
    January 11th, 2011 at 10:28 pm

    If I remember correctly, Rpos is recomputed for each decade. With some players you can see this when their Rpos changes between year xyz9 and the next year. (remember that Rpos, like Rrep depends upon playing time.)

17. DavidRF Says:
    January 11th, 2011 at 11:15 pm

    Ah... thanks. I can see significant shifts in rPos for Eddie Collins (2B) between 1919 and 1920 and Stan Hack (3B) between 1939 and 1940. Those two positions were indeed trading places in the defensive spectrum during the few decades but its interesting to see the sharp breaks in the positional adjustment at the end of these particular decades. I don't have the raw data, but at some point they might consider smoothing those transitions.

18. Johnny Twisto Says:
    January 11th, 2011 at 11:17 pm

    David, you misunderstood Kds. I am pretty sure B-R WAR does use defensive performance to set the differences between positions, by looking at the ratings of everyone who plays multiple positions in a season.

19. mccombe35 Says:
    January 11th, 2011 at 11:34 pm

    Interesting list.

    Guys like Ortiz & Hafner are way up there because they hardly ever played the field.

    Less than 250 games in the field between them. Too bad Thomas & Baines (Over 2,000 games in the field combined) will/are lumped into that DH group when being looked at historically...

20. DavidRF Says:
    January 12th, 2011 at 7:30 am

    @18
    "David, you misunderstood Kds. I am pretty sure B-R WAR does use defensive performance to set the differences between positions, by looking at the ratings of everyone who plays multiple positions in a season."
    --------------------
    I guess I'm still confused then. I thought Rpos was an adjustment based the difference in offensive replacement level due to positional scarcity. As you say, its the extra fielding skill required by the defensive positions that creates that scarcity, but the number itself is obtained by measuring offense, not measuring defense.

    Pitchers have a high value of rPos. Per plate appearance, its enormous. That number is not because of their fielding. Its because they can't hit.

21. 39+ Games As DH With WPA >= .15 » Baseball-Reference Blog » Blog … | Baseball News and Events Says:
    January 12th, 2011 at 9:57 am

    [...] more from the original source: 39+ Games As DH With WPA >= .15 » Baseball-Reference Blog » Blog … Posted in General Tags: david, games, kevin-bryant, mentioned-on-twitter, reference, twitter, [...]

22. Johnny Twisto Says:
    January 12th, 2011 at 2:56 pm

    David, I guess I'm confused as well. I thought rPos was derived from measuring fielding differences between positions, but you're right, that would not explain why pitchers have such a high rPos (I had never noticed that before).