JotCast | Dan Szymborski FanGraphs Chat 7/23/26

Return to

Dan Szymborski FanGraphs Chat 7/23/26

View Full Chat Next Page

This chat is available to members only. Please log in to participate.

Dan Szymborski

12:03

And awaaaaaaaay we go?

Larry Mahnken

12:03

But why male models?

Dan Szymborski

12:03

Larry! Option J?

furiousherbivore

12:03

Is Jim Jarvis sort of, kind of good?

Dan Szymborski

12:04

He might be!

He was one of the biggest ZiPS movers earlier this season

And he hasn't done anything to change that in a bad direction

12:06

Just checked, he's moved up another 7 points of 2027 projected OPS+ since then

For those who can't see what I'm looking at, which I think is everyone but me, he a a 68 OPS+ projection for 2027 in the preseaosn, 82 a month ago, 89 now

toodamntall

12:06

Say something nice, I am a bit under the weather

Do you have a shorthand rule for how good a player's defense needs to be if their offense is weak, or how much they can suck with the glove if they hit well? Obviously position makes a big difference

Dan Szymborski

12:07

You'll only fail to recover from being under the weather exactly one time in your life!

12:08

No shorthand, but if someone isn't somewhere around two WAR now, or has a good shot at two WAR later, then I consider them a stopgap

but degrees of excellence and suckulence aren't binary

MMM

12:08

I have been thinking about run differential lately, mostly because I follow the Royals, and when they lose, as they often do, they have a tendency to get walloped. And frequently, when they’re well behind, they put in Tyler Tolbert to throw lollipops and lose by an even wider margin. As of this morning, they are minus-96 for the season.

As teams essentially surrender to save their bullpen, is run differential becoming less useful for predicting how many games a team “should have” won? Is Bill James’s “Pythagorean expectation” null?

Dan Szymborski

12:08

I haven't found a change in recent accuracy

hofheinz

12:09

Have you looked into whether different profiles of players age better/worse? Seems the common wisdom is speed profiles age worse and plate discipline profiles age better. But is that actually borne out?

Dan Szymborski

12:09

Yes, it's why ZiPS performs better by comparing to large cohorts of roughly similar players rather than just by age!

I've tested this extensively; I can click a box and ZiPS will simply ignore all the cohorts and go by age

BuccoFan

12:11

Who's someone that you think will be traded at the deadline that nobody is talking about right now?

Dan Szymborski

12:11

BUT WOULDNT I BE TALKING ABOUT IT THEN

12:12

Shoudln't there be more Robbie Ray talk then there's been?

12:13

But let's go weird with it. They've SAID they aren't suspiciously often, so maybe Logan Webb?

12:15

It's a seller's market and Webb would draw a TON of interest I think. The Giants haven't been better than 81-81 since that weird 2021 season, and I'm just wondering if they may be on the brink of just blowing it up

12:16

Because a seller's market might be the best time to cut your losses on Devers

Or maybe not. But it's INTERESTING

lain

12:16

Hi Dan, in head to head points I can run Mis, Schlittler, McLean and Eury Perez in my four starter slots, or one of Davie Peterson, Sproat or Thornton on a two-step. With the deadline and so much fluidity, am I better off running with just the four starters? Also, Jackson Holliday makes me sad.

Dan Szymborski

12:16

I'd go with the four. And me as well

didace

12:17

In the comments section of the JJ Wetherholt article from last week there was a discussion on if ZiPS was too low on his HR projections. As part of that this was posted "Because of the methodology of models I have often wondered if Mike Trout ever performed below his 80th. or even 90th, percentile between 2012-2019. What is the value of a projection if there is a high probability that it is not going to reflect the outcome? Am I correct in thinking that projections have a difficult time handling outliers such as Trout?" Is this true? Specifically about Trout and in general about great players. I know that there have to be some players that consistently out play their projections, but if a certain type of player always did then couldn't you correct for that?

Dan Szymborski

12:17

I can give you an answer right now, while I wait

12:18

that sentence doesn't make sense

if YOU can wait

12:19

I'm just going to do wRC+ percnetiles as a shorthand, since I can't spent 20 minutes on this getting the playing time matching since people woudl murder me in their head

12:20

2012: 94th percentile

12:21

2013: 74th percentile wRC+

2014: 61st percentile

12:22

2015: 39th percentile

12:25

here are preseason WAR projections each year from 2012-2019

2013 7.1 2014 9.5 2015 9.6 2016 9.3 2017 9.0 2018 7.9 2019 8.3

that's 60.6 WAR

Connecting…