I would like to coin a new phrase, combining two words that are not often used togeter – “baseball nerd.” We are all baseball nerds. This is great stuff.
I am new to viewing this site and it has lot of great information.
Just to understand, the UZR and the positional values are the defensive metrics. You can add these together to value a players defense.
Adding wRAA and replacement value (not yet discussed) are the offensive metrics. You can add these values together to value a players offense.
Adding them all together gives you a total picture of a players value. Are these assumptions right?
Also, off topic, do you plan on providing righty/lefty stats in the future. I see you provide this info in graph form. If not, could you atleast add the actual numerical stat to these graphs.
Lastly, do you plan on changing the current wOBA listed on your site to one that would adjust for park and league factors.
I wouldn’t necessarily phrase it that way. wRAA is an offensive metric, UZR is a defensive metric. The position adjustment, while based on historical defensive performance, is a neutral metric, as is replacement level.
For instance, you could add wRAA + Position + Replacement together and get our version of VORP, which is an offensive metric.
Adding all four components together gives you a player’s value. You could create other permutations of the components to measure other things – offensive or defenisve runs above replacement position, offensive or defensive runs above average, etc…
There’s another reason for using replacement level. In my mind, it’s much more important than finding a good baseline for salary. It has to do with playing time.
In essence, it’s easier to be ten runs above average in half a season than an entire season. The standard deviation of performance is greater when the sample size is smaller, so the full-season figure is more impressive than the half-season figure. However, if you just used WAA, you wouldn’t know the difference.
Replacement level fixes this problem by factoring in playing time.
Of course, you may have already said this in Part Four. I haven’t read that yet. Can’t keep up!
So if I add up all the position adjustments I get 0 with no DH and -17.5 with a DH. If 14/30 of the teams are AL and 14/30 games are with DH that weights to an average position adjustment of -8.2 per 150 games. That’s -8.8 per 162. If I go to the team page, I see the average team positional adjustment is -12.4. I was expecting those numbers to be equal.
Further, I was expecting each team’s positional adjustment to be equal to -17.5 * % Games in AL Stadiums * (162/150) since each team fields the same positionss, but that’s not the case. In fact, there’s quite a bit of variation among the team totals: AL ranges from -16 to -27 and NL ranges from -1 to -10. A quick glance suggests that these numbers are before any player movement (Teixeira listed with Braves and Angels).
Love your stuff here and on USSM. Just to clarify for my own sanity, using the positional adjustments you listed, would catching (+12.5 runs) be considered statistically the most difficult position or do catchers get the most credit for their defense? I’m sure I’m just over thinking this…
Hmm, I may have I figured it out: I think fielders get a full games positional adjustment whether they play the whole game or just an inning. So if the positional adjustments are meant to be per 150 full games rather than 150 average games, the positional adjustments are overstated.
More of the former. The way Tango discusses positional adjustments is called Runs Over Willie, for Willie Bloomquist, or other similar players who can play everywhere. If you put Willie at SS, then the average SS would save +7.5 more runs per 162 games. If you placed Willie at 1B, then the average 1B would cost his team -12.5 runs more per 162 games than Willie.
With regards to catching, this suggests that if you place a guy like Willie behind the plate, even an average fielding catcher will save +12.5 more runs than Willie, which is +1.25 wins. Catching is a specialized skill, so this basically says that even someone with average catching skills is over one win more valuable behind the plate than the regular utility guy you can throw back there.
“It is much harder to find a +5 SS than it is to find a +5 2B, and we need to represent that in the Win Value system.” This statement also applies to right fielders (SS) and left fielders (2B). The pool of candidates is smaller for right field; to play there you need to be able to throw. This is not reflected in the adjustments.
One thing that I don’t understand is why some players do not receive the full value for their positional adjustment. My guess is that it’s based off of playing time. Am I right?
Small point, but you write: “Clearly, shortstops are better defenders than second baseman, and we have to reflect this in their value.” I’d think that’s just as clearly not always the case (or, as someone at BBTF might say, “Derek Jeter says ‘hi.’”
So you might want to rephrase to say what you mean. (Or maybe I shouldn’t have read the explanation on a break from grading papers.)
mymrbig says:
December 30, 2008 at 2:30 pm
I would like to coin a new phrase, combining two words that are not often used togeter – “baseball nerd.” We are all baseball nerds. This is great stuff.
Columbo says:
December 30, 2008 at 2:34 pm
I am new to viewing this site and it has lot of great information.
Just to understand, the UZR and the positional values are the defensive metrics. You can add these together to value a players defense.
Adding wRAA and replacement value (not yet discussed) are the offensive metrics. You can add these values together to value a players offense.
Adding them all together gives you a total picture of a players value. Are these assumptions right?
Also, off topic, do you plan on providing righty/lefty stats in the future. I see you provide this info in graph form. If not, could you atleast add the actual numerical stat to these graphs.
Lastly, do you plan on changing the current wOBA listed on your site to one that would adjust for park and league factors.
Thank you.
Tony says:
December 30, 2008 at 2:53 pm
I understood all three parts, as I use something similar myself. I was just wondering how you come up with the replacement value.
Tony says:
December 30, 2008 at 3:38 pm
NVM my last comment. I didn’t read the last sentence. I’m a dummy.
Rich says:
December 30, 2008 at 3:42 pm
Does Fangraphs have a section where league average stats are listed?
Dave Cameron says:
December 30, 2008 at 4:08 pm
I wouldn’t necessarily phrase it that way. wRAA is an offensive metric, UZR is a defensive metric. The position adjustment, while based on historical defensive performance, is a neutral metric, as is replacement level.
For instance, you could add wRAA + Position + Replacement together and get our version of VORP, which is an offensive metric.
Adding all four components together gives you a player’s value. You could create other permutations of the components to measure other things – offensive or defenisve runs above replacement position, offensive or defensive runs above average, etc…
studes says:
December 30, 2008 at 5:50 pm
There’s another reason for using replacement level. In my mind, it’s much more important than finding a good baseline for salary. It has to do with playing time.
In essence, it’s easier to be ten runs above average in half a season than an entire season. The standard deviation of performance is greater when the sample size is smaller, so the full-season figure is more impressive than the half-season figure. However, if you just used WAA, you wouldn’t know the difference.
Replacement level fixes this problem by factoring in playing time.
Of course, you may have already said this in Part Four. I haven’t read that yet. Can’t keep up!
Dave Cameron says:
December 30, 2008 at 6:47 pm
Yep, that’s exactly right, and covered in part four. I’ll try to slow down so you old guys can keek up…
MattyD says:
December 30, 2008 at 7:02 pm
So if I add up all the position adjustments I get 0 with no DH and -17.5 with a DH. If 14/30 of the teams are AL and 14/30 games are with DH that weights to an average position adjustment of -8.2 per 150 games. That’s -8.8 per 162. If I go to the team page, I see the average team positional adjustment is -12.4. I was expecting those numbers to be equal.
Further, I was expecting each team’s positional adjustment to be equal to -17.5 * % Games in AL Stadiums * (162/150) since each team fields the same positionss, but that’s not the case. In fact, there’s quite a bit of variation among the team totals: AL ranges from -16 to -27 and NL ranges from -1 to -10. A quick glance suggests that these numbers are before any player movement (Teixeira listed with Braves and Angels).
What am I missing here?
studes says:
December 30, 2008 at 8:33 pm
Huh?
studes says:
December 30, 2008 at 8:34 pm
(said in an old guy’s voice, one who’s hard of hearing. In case the joke wasn’t obvious.)
Chris says:
December 31, 2008 at 2:07 pm
Dave,
Love your stuff here and on USSM. Just to clarify for my own sanity, using the positional adjustments you listed, would catching (+12.5 runs) be considered statistically the most difficult position or do catchers get the most credit for their defense? I’m sure I’m just over thinking this…
MattyD says:
December 31, 2008 at 2:17 pm
Hmm, I may have I figured it out: I think fielders get a full games positional adjustment whether they play the whole game or just an inning. So if the positional adjustments are meant to be per 150 full games rather than 150 average games, the positional adjustments are overstated.
Eric Seidman says:
December 31, 2008 at 2:26 pm
Chris,
More of the former. The way Tango discusses positional adjustments is called Runs Over Willie, for Willie Bloomquist, or other similar players who can play everywhere. If you put Willie at SS, then the average SS would save +7.5 more runs per 162 games. If you placed Willie at 1B, then the average 1B would cost his team -12.5 runs more per 162 games than Willie.
With regards to catching, this suggests that if you place a guy like Willie behind the plate, even an average fielding catcher will save +12.5 more runs than Willie, which is +1.25 wins. Catching is a specialized skill, so this basically says that even someone with average catching skills is over one win more valuable behind the plate than the regular utility guy you can throw back there.
M. T. Head says:
January 2, 2009 at 5:30 am
“It is much harder to find a +5 SS than it is to find a +5 2B, and we need to represent that in the Win Value system.” This statement also applies to right fielders (SS) and left fielders (2B). The pool of candidates is smaller for right field; to play there you need to be able to throw. This is not reflected in the adjustments.
Jack T says:
April 16, 2009 at 9:17 pm
I don’t get why guys like Albert Pujols are in the negatives for positional adjustment but people like Hanley are positive?
alskor says:
April 16, 2009 at 9:33 pm
Pujols = 1B
Hanley = SS
Its a lot easier to find a 1B than a SS. Its a lot easier to play 1B than SS.
WENTWORTH says:
June 1, 2009 at 1:10 am
??, ?????? ??? ? ???????????????? – ??? ??? ? ????. P.S. ????, ??????, ? ??? ????????? ???????? :)
brian recca says:
August 25, 2009 at 4:22 pm
One thing that I don’t understand is why some players do not receive the full value for their positional adjustment. My guess is that it’s based off of playing time. Am I right?
Marcus says:
October 21, 2009 at 12:57 pm
Nice post you got here. It would be great to read more concerning this theme.
krm says:
December 15, 2009 at 1:00 pm
Small point, but you write: “Clearly, shortstops are better defenders than second baseman, and we have to reflect this in their value.” I’d think that’s just as clearly not always the case (or, as someone at BBTF might say, “Derek Jeter says ‘hi.’”
So you might want to rephrase to say what you mean. (Or maybe I shouldn’t have read the explanation on a break from grading papers.)
IvanGrushenko says:
July 19, 2011 at 6:11 pm
What is the positional adjustment for pitchers as hitters? +75 or something?