Are Linear Weights Outdated?

[Ted]

Okay, so I'm not the greatest quant, but here's my understanding. Linear weights assign an average value of an event to that outcome, assuming a predictable distribution of those outcomes that fall within a typical range.

Umm. Let's try again. Linear weights say a single is worth "X". But we know that's not all singles are the same. The Dusty Rhodes conversation actually got me thinking. A real life Dusty Rhodes would actually have less valuable singles on average, because he HAS to be blooping the ball all over the place. Guys are not scoring from second or going first to third less on a real life Dusty Rhodes single as often (this is NOT true in OOTP. It doesn't care about physics).

So, we now have better understanding of batted ball profiles than ever. Is there a case where player A below is more valuable than player B consistently, simply due to type of batted ball? Assume identical extra base hits. The only difference is that player A has "better" singles than player B.

Player A: .270/.290/.400
Player B: .275/.295/.405

Can this be done consistently? Watching baseball every day, I do feel like I see these players. There do seem to be players who have weak singles and more walks, and players who have better singles and fewer walks. Now, typically the player with better singles also has more extra base hits as he's making better contact overall. But I wonder if MLB front offices are looking at this.

Anyway, it's pretty interesting (to me anyway).

[Ted]

Also, if I missed any writing about this, please point me in the right direction. I don't presume to think I know more than Tom Tango et al, but I've just wondered about this more and more recently (as I've seen some odd managerial and front office decisions on playing time).

[RonCo]

It's definitely an interesting question.

Whether your situation is correct or not (some types of batted singles are worth more than other types) stats based on linear weights have the problem of not mattering in the moment. Just because an average single is worth X, that's not meaningful in a game where the score is 9-2 in the ninth inning. Of course, you can say the same thing about home runs. A home run by the losing team in a 10-1 game is arguably not worth anything. So does this lead one to focus on WPA instead of wOBA? I dunno.

It will be interesting to see how run producing is measured in the near future. People are already using things like launch angle and exit velocity to calculate expected wOBA then comparing that to the player's actual wOBA. This kind of study might reveal interesting things about players.

Good question...

Bottom line for OOTP (and any computer game) is that current analytics and measurement systems are moving too rapidly for them to keep up for the moment, and design decisions they made in the 20-aughts have them painted a bit into a corner for a little while, anyway.