You never know when your pants will come down.

[ashbury]

4 hours ago, Riverbrian said:

Agreed

Yet... these defensive metrics are folded into WAR. 

I use a stick of butter when I make my scrambled eggs. There are people who tell me that is too much butter. 

Make friends with Paula Deen.  She will never tell you you used too much butter.

My point was that dealing with fielding statistics is inherently more difficult than with batting/pitching statistics. It was not that fielding is unimportant.  If you are going to try to give an estimate of a player's contribution towards winning, defense absolutely has to be "folded in."

I just got done describing why fielding suffers from the curse of smaller samples.  However there is an opposite force that makes the stats we do get more meaningful.

Everybody bats.  If you have a good glove in the lineup, he still has to bat, same as the other 8 players in the lineup, unless you pinch hit for him.  So the best hitters and the worst get measured on a relatively level playing field all season.  You can see who is good and who is bad.

That's not the case for fielding.  It's not like in indoor volleyball where the whole team rotates.  Your best bat is not necessarily required to man the CF spot for an inning each game.  (Might be an interesting variation, mandating a team of more well-balanced players, but to my knowledge it's never been seriously tried in the history of competitive baseball.)

Imagine if only Jeff McNeil, Freddie Freeman, Paul Goldschmidt, Luis Arraez and Aaron Judge batted during the season.  You might never realize just how elite Luis's batting average was in 2022.  "Meh, he's just a little better than Judge."

It's kind of like that in CF.  We know that Byron Buxton outshines Celestino, Kepler, Cave, Gordon, Contreras at that position.  But it's even more of a big deal, because guys like Gio Urshela, Jose Miranda, Gary Sanchez aren't required to take their turns out there - you would really see just how many runs Buxton saves us with his legs and glove and arm.

We're lucky that Buxton has a solid bat, because without him a lineup composed only of our best bats might lead to a debacle in CF.  As for the others, Celestino et al, an adjustment for simply playing CF at a competent level to satisfy the manager belongs in WAR, to reflect the contribution he brings to that position even if just as a backup; any defensive stats on top of that, to try to judge how he compares to his peers in CF is nearly an afterthought, although in Buxton's case that differential is worth including in WAR too.

Imprecise, noisy, important data.  That's a dilemma for any analytics person, whether at a Wall Street firm making investments or at an airline doing their scheduling or at a baseball front office deciding who contributes to wins.

[Riverbrian]

33 minutes ago, ashbury said:

My point was that dealing with fielding statistics is inherently more difficult than with batting/pitching statistics. Not that fielding is unimportant.  If you are going to try to give an estimate of a player's contribution towards winning, defense absolutely has to be "folded in."

I just got done describing why fielding suffers from the curse of smaller samples.  However there is an opposite force that makes the stats we do get more meaningful.

Everybody bats.  If you have a good glove in the lineup, he still has to bat, same as the other 8 players in the lineup, unless you pinch hit for him.  So the best hitters and the worst get measured on a relatively level playing field all season.  You can see who is good and who is bad.

That's not the case for fielding.  It's not like in indoor volleyball where the whole team rotates.  Your best bat is not necessarily required to man the CF spot for an inning each game.  (Might be an interesting variation, mandating a team of more well-balanced players, but to my knowledge it's never been seriously tried in the history of competitive baseball.)

Imagine if only Jeff McNeil, Freddie Freeman, Paul Goldschmidt, Luis Arraez and Aaron Judge batted during the season.  You might never realize just how elite Luis's batting average was in 2022.  "Meh, he's just a little better than Judge."

It's kind of like that in CF.  We know that Byron Buxton outshines Celestino, Kepler, Cave, Gordon, Contreras at that position.  But it's even more of a big deal, because guys like Gio Urshela, Jose Miranda, Gary Sanchez aren't required to take their turns out there - you would really see just how many runs Buxton saves us with his legs and glove and arm.

We're lucky that Buxton has a solid bat, because without him a lineup composed only of our best bats might lead to a debacle in CF.  As for the others, an adjustment for simply playing CF at a competent level to satisfy the manager belongs in WAR, to reflect the contribution he brings to that position even if just as a backup; any defensive stats on top of that, to try to judge how he compares to his peers in CF is nearly an afterthought, although in Buxton's case that differential is worth including in WAR too.

Imprecise, noisy, important data.  That's a dilemma for any analytics person, whether at a Wall Street firm making investments or at a airline doing their scheduling or at a baseball front office deciding who contributes to wins.

Yes I agree and understand.

It's adding the defensive wind to the offensive air temperature. I've spent enough time in North Dakota to understand what wind chill is. 

My issue is that I lose faith in the wind chill reports when the anemometer readings are "inherently more difficult" from the "curse of smaller samples". I take issue when the majority of wind events are routine and easily influenced by a couple of 100 mph winds and then apparently over-weighted when mixed with the air temperature.

They can tell me that Max is 10 degrees with a -15 below wind chill but that flag isn't flapping like that.  

Besides... I'm more of a hot air guy.   ? 

[Parfigliano]

On 10/19/2022 at 5:44 PM, stringer bell said:

I don't know if that is true or not, but I think the defensive portion of WAR is suspect. Josh Donaldson comes to the Bronx and suddenly is saving dozens of runs resulting in about a 2.4 WAR despite having a 94 OPS+. Kiner-Falefa also has supposedly nice defensive grades despite his defense being criticized by Yankee fans and getting benched in the playoffs. Until it is proven to me, I'm going to believe that defensive metrics are subjective.

I have always believed defensive metrics are highly subjective but not as subjective as pitch framing which is nothing but a crock of s#!+.

[Richie the Rally Goat]

10 hours ago, Parfigliano said:

I have always believed defensive metrics are highly subjective but not as subjective as pitch framing which is nothing but a crock of s#!+.

Well…

pitch framing has a statistically significant sample size. Hundreds of pitches per game. The strike zone definition is pretty decent and the one that would be used for an automated zone. 
 

while it would be great if umpires were perfect and weren’t biased and all catchers had to do was catch and throw, but it’s the game. The measurement is as good or better than most all other advanced metrics, but the reason for the metric sucks.

[Riverbrian]

22 hours ago, Richie the Rally Goat said:

Well…

pitch framing has a statistically significant sample size. Hundreds of pitches per game. The strike zone definition is pretty decent and the one that would be used for an automated zone. 
 

while it would be great if umpires were perfect and weren’t biased and all catchers had to do was catch and throw, but it’s the game. The measurement is as good or better than most all other advanced metrics, but the reason for the metric sucks.

Pitch framing data does seem to replicate.