Data, Development and the FUTURE of White Sox Pitching | The White Sox Podcast

[almagest]

2 hours ago, southsider2k5 said:

In fact, in the short term we have seen a massive injury spike, which we did not see previously, which is exactly what allowed our pitching to be so successful.  Our starters were out there almost every day.  That's not the case anymore, so while maybe we get more out of our starters, we are also more often pitching our #10 starter, or some dude off of the street, so I am not sure net/net that this is a positive to this point.  Again, we obviously need to see if this is the new normal or not for injuries, but early returns are not all positive.

The Sox did not prioritize maxing velocity the way other orgs did, and velocity is what correlates heavily with TJ surgery. I don't know if the Sox are prioritizing max velocity more now (I suspect they are, as everyone is), but I do know that every pitcher is told to prioritize velocity as soon as they show any ability to pitch at a high level, so I'm not surprised to see a rash of surgeries needed. Hopefully they don't all hit at once like they did for the Sox this year, but I would expect every viable prospect to need TJ at some point.

[southsider2k5]

8 minutes ago, almagest said:

The Sox did not prioritize maxing velocity the way other orgs did, and velocity is what correlates heavily with TJ surgery. I don't know if the Sox are prioritizing max velocity more now (I suspect they are, as everyone is), but I do know that every pitcher is told to prioritize velocity as soon as they show any ability to pitch at a high level, so I'm not surprised to see a rash of surgeries needed. Hopefully they don't all hit at once like they did for the Sox this year, but I would expect every viable prospect to need TJ at some point.

That's great.  If we aren't going to be the LAD Dodgers and sign a 60 player roster, we don't have the depth to pitch over a Dodgers level of injuries.

[almagest]

49 minutes ago, southsider2k5 said:

That's great.  If we aren't going to be the LAD Dodgers and sign a 60 player roster, we don't have the depth to pitch over a Dodgers level of injuries.

You're missing the point. This is a problem throughout all of baseball because of how pitchers are taught to maximize velocity before they're even drafted. There's not much the Sox can do about it. Having so many pitchers drop around the same time is likely a statistical anomaly. 

[southsider2k5]

21 minutes ago, almagest said:

You're missing the point. This is a problem throughout all of baseball because of how pitchers are taught to maximize velocity before they're even drafted. There's not much the Sox can do about it. Having so many pitchers drop around the same time is likely a statistical anomaly. 

And this goes back to my original point in that if we are following along in increasing our injury numbers and having more bad pitchers pitch a lot more bad games, this isn't moving us forward.  If we aren't going to maintain a roster like the Dodgers do where we are willing to sign a ton of guys, we might actually being losing ground by not having our pitchers going at a slightly lower rate, but making sure they are on the mound 33 starts a year like we used to.

[almagest]

9 minutes ago, southsider2k5 said:

And this goes back to my original point in that if we are following along in increasing our injury numbers and having more bad pitchers pitch a lot more bad games, this isn't moving us forward.  If we aren't going to maintain a roster like the Dodgers do where we are willing to sign a ton of guys, we might actually being losing ground by not having our pitchers going at a slightly lower rate, but making sure they are on the mound 33 starts a year like we used to.

And I'm saying I don't know if that's going to matter because a lot of the damage is already done, unless the Sox focus on drafting guys who have some biomechanical advantage that they can identify, or who either already had TJS or need it right away.

[Look at Ray Ray Run]

2 hours ago, Chicago White Sox said:

You are missing the forest from the trees here man.  “Copying what others are doing” is 100% without question what we need to be doing.  Will that make us a best in class organization?  f*** no.  Will it give us an edge over others?  Definitely not.  But it will quickly bring us out of the god damn Stone Age that we were stuck in under Hahn.  You have to start somewhere and getting your basic foundations in place is the first step towards righting the ship long-term.  Going from being completely incompetent to just below average would be a massive improvement.

As for your pitching analysis, using a 20 year sample to highlight our capabilities as of late is completely flawed.  20 years ago pitching labs and biomechanics were barely a thing and have obviously grown in prominence over time.  So much of the value that you are citing is likely driven by a single draft pick that was selected and developed by a different group of people using different processes (ones that were likely more up-to-date at the time).  Change your sample to players added in the past 10 years and tell me where we rank.  Hell, make it five and tell me where we rank.  We are well fucking past the Don Cooper & Herm Schneider glory days.

And let me be clear here, there is certainly a PR angle at play here.  I’m not a fool who believes everything Getz says and I know there is an intrinsic motivation for him to hype up all the behind the scenes stuff.  But that doesn’t change the fact that they are actually trying to modernize the org by bringing in external ideas, best practices, and people.  Again, that won’t solve all our problems, but it should materially bridge the gap between us and most other orgs outside of probably the Rockies.  And if we can be even be in the realm of acceptability, there is a chance even a mediocre GM can get lucky and put together a competitive team.  But until the basics are in place, the current nightmare we are enduring will never end.

The job of data is merely to represent, statistically, the observations coaches and PD departments have been monitoring for years. Is it more exact? Of course. Has it revolutionized pitching in baseball? Not for the reasons you're citing. In fact, it has limited sustained success, dramatically decreased average innings pitched and tenures, all the while increasing injury frequency.

Also, decreasing the sample would just increase statistical noise. Using a 20 year window allows a long enough window to assess careers and the general impact of process/programs.

It's fascinating how I spend much of my days now telling people that data has become overvalued and overrated. The amount of people I've witnessed in my career who leverage data for pure stupidity and think, that, because they're leveraging data it's valuable, is endless. In my personal life I feel as though I spend half my days driving down Machine Learning (predictive models) and LLM implementations as oversold tools that should be a resource but not a decision maker, and the other half of my days implementing and selling the value of those products to businesses. I guess I'm part of the problem!

The Sox have shown zero examples, in my opinion, of the viability of their data usage. The examples I was given the other day were people that have accomplished nothing and actually have gotten worse YoY (Smith/Schultz and etc).

Edited June 25 by Look at Ray Ray Run

[southsider2k5]

14 minutes ago, almagest said:

And I'm saying I don't know if that's going to matter because a lot of the damage is already done, unless the Sox focus on drafting guys who have some biomechanical advantage that they can identify, or who either already had TJS or need it right away.

I would hope this was already a focus honestly, otherwise why do analytics if you can't identify the guys to stay away from?

[almagest]

8 minutes ago, southsider2k5 said:

I would hope this was already a focus honestly, otherwise why do analytics if you can't identify the guys to stay away from?

The "had TJS or will have it" part is easy, I'm sure, but the biomechanical advantage part likely isn't (otherwise I'd bet the Dodgers would be doing it). Hopefully that starts to show up as they draft and acquire more pitching in the next few years.

[southsider2k5]

12 minutes ago, almagest said:

The "had TJS or will have it" part is easy, I'm sure, but the biomechanical advantage part likely isn't (otherwise I'd bet the Dodgers would be doing it). Hopefully that starts to show up as they draft and acquire more pitching in the next few years.

If we aren't looking for advantages, what are we using data and changing deliveries on the basis of?  There is obviously there is SOMETHING we are looking to change TO.  Shouldn't we be looking for the same sorts of things in others?

[Look at Ray Ray Run]

14 minutes ago, almagest said:

The "had TJS or will have it" part is easy, I'm sure, but the biomechanical advantage part likely isn't (otherwise I'd bet the Dodgers would be doing it). Hopefully that starts to show up as they draft and acquire more pitching in the next few years.

Mechanics, in general, are optimized first to maximize stuff and then everything else is considered. There's a reason pitchers haven't thrown pitches that move or etc in a certain way and it's because it's even less natural than throwing already is. By focusing on depth and movement, you try things that your body wouldn't otherwise do which i believe leads to injury in addition to max effort. 

I'm not saying there's not value to obtaining those abilities, but the problem is the value isn't there on the health side meaning the individual players get much less long term return on that investment. Meanwhile, the teams can just recycle you and max out another guy.

The goal should be to maximize stuff while maintaining natural range of motion, but the stuff maximization tends to overwhelm the other goals from my experience. 

Edited June 25 by Look at Ray Ray Run

[southsider2k5]

20 minutes ago, Look at Ray Ray Run said:

The job of data is merely to represent, statistically, the observations coaches and PD departments have been monitoring for years. Is it more exact? Of course. Has it revolutionized pitching in baseball? Not for the reasons you're citing. In fact, it has limited sustained success, dramatically decreased average innings pitched and tenures, all the while increasing injury frequency.

Also, decreasing the sample would just increase statistical noise. Using a 20 year window allows a long enough window to assess careers and the general impact of process/programs.

It's fascinating how I spend much of my days now telling people that data has become overvalued and overrated. The amount of people I've witnessed in my career who leverage data for pure stupidity and think, that, because they're leveraging data it's valuable, is endless. In my personal life I feel as though I spend half my days driving down Machine Learning (predictive models) and LLM implementations as oversold tools that should be a resource but not a decision maker, and the other half of my days implementing and selling the value of those products to businesses. I guess I'm part of the problem!

The Sox have shown zero examples, in my opinion, of the viability of their data usage. The examples I was given the other day were people that have accomplished nothing and actually have gotten worse YoY (Smith/Schultz and etc).

If you don't have solid samples, you can't build predictive models which should be giving you ideas of what to change towards.  If you aren't, I am not sure what you are using the data for?

 

[almagest]

3 hours ago, Look at Ray Ray Run said:

The things cited around R&D, but most specifically data was frankly meaningless. The advantage of data adaptation in baseball is around proprietary models and robust predictive analytics, and saying you "cleaned up the data and moved it to the cloud" in no way implies advanced modeling.

This is true, but you still need a platform to build upon. This is one of the biggest problems in data science - so many people want to start from the end goal and don't understand that you need to put a ton of work into collecting, sanitizing and normalizing your data first. I wouldn't be surprised to hear that the White Sox accumulated a ton of player info during their Hahn-driven modernization, but did not understand how to properly leverage it to develop the insights needed to build good proprietary models and predictive analytics. I think these are reasonable assumptions, given what we've heard about lack of opposing player scouting, internal siloing and lack of quality internal data-driven initiatives in the org, plus what Bannister said during this podcast.

4 hours ago, Look at Ray Ray Run said:

When he was asked to define an edge-area for the Sox, he was able to provide the specific edge for the Giants and the Red Sox but frankly provided nothing of substance/meaning for the White Sox.

This bothered me too and was the worst part of the podcast. "We're having pitchers learn changeups" isn't much of a revolution. I wish Chuck pressed him more on it, though I could imagine a couple reasons why he was reluctant to share more - A) they don't want to give away their perceived advantage, or B) they're still working on developing these internal insights (because the data collection and cleanup was such a mess) and are currently relying more on old-school methods, like Coop teaching everyone a cutter.

4 hours ago, Look at Ray Ray Run said:

In fact, the thing they cited as their focal point within the offensive data (zone command, swing decisions etc) isn't some under-valued market gap and has resulted in the White Sox having the 2nd worst offense in MLB.

It's not an under-valued gap, that's just a talent issue.

[Look at Ray Ray Run]

10 minutes ago, southsider2k5 said:

If you don't have solid samples, you can't build predictive models which should be giving you ideas of what to change towards.  If you aren't, I am not sure what you are using the data for?

 

Models shouldn't be your primary idea generators, they're idea assessors and maximizers. I see far too many people who try to develop hypothesis from correlated components as opposed to using the data to validate or substantiate your observations/areas of focus. I call it directionless modeling which might as well just be throwing s%*# against the wall until something sticks.

[almagest]

13 minutes ago, Look at Ray Ray Run said:

Mechanics, in general, are optimized first to maximize stuff and then everything else is considered. There's a reason pitchers haven't thrown pitches that move or etc in a certain way and it's because it's even less natural than throwing already is. By focusing on depth and movement, you try things that your body wouldn't otherwise do which i believe leads to injury in addition to max effort. 

I'm not saying there's not value to obtaining those abilities, but the problem is the value isn't there on the health side meaning the individual players get much less long term return on that investment. Meanwhile, the teams can just recycle you and max out another guy.

The goal should be to maximize stuff while maintaining natural range of motion, but the stuff maximization tends to overwhelm the other goals from my experience. 

It's also an arms race with hitters. Hitting has been optimized so heavily in the last 10-15 years in ways that have the advantage of not leading to more injuries. Pitchers likely need to heavily focus on stuff maximization, otherwise we'd have an offensive environment worse than the late 90s and 2000s, when a 5 ERA was a decent 4th starter.

[southsider2k5]

12 minutes ago, almagest said:

It's also an arms race with hitters. Hitting has been optimized so heavily in the last 10-15 years in ways that have the advantage of not leading to more injuries. Pitchers likely need to heavily focus on stuff maximization, otherwise we'd have an offensive environment worse than the late 90s and 2000s, when a 5 ERA was a decent 4th starter.

The thing that I come back to is that outside of our overall talent level, it hasn't felt like pitching talent was significantly underproducing the equivalent talent vs what other teams were seeing  with similar guys.  It's not like any of those guys left here and unlocked something we didn't either.

[Look at Ray Ray Run]

11 minutes ago, almagest said:

It's also an arms race with hitters. Hitting has been optimized so heavily in the last 10-15 years in ways that have the advantage of not leading to more injuries. Pitchers likely need to heavily focus on stuff maximization, otherwise we'd have an offensive environment worse than the late 90s and 2000s, when a 5 ERA was a decent 4th starter.

Interesting, I actually think data has heavily skewed towards benefiting pitching. Hitting is so reactionary by default that it's much harder to optimize, which is why the run scoring environment has tanked.

[almagest]

10 minutes ago, Look at Ray Ray Run said:

Interesting, I actually think data has heavily skewed towards benefiting pitching. Hitting is so reactionary by default that it's much harder to optimize, which is why the run scoring environment has tanked.

It definitely is in game, but it seems like hitters are pretty well maximized to do as much damage as their pitch recognition and talent levels will allow. If we went back to the times when pitching low in the zone with high 80s and low 90s velo was the norm and tunneling wasn’t well known, we’d see some crazy damage.

[almagest]

17 minutes ago, southsider2k5 said:

The thing that I come back to is that outside of our overall talent level, it hasn't felt like pitching talent was significantly underproducing the equivalent talent vs what other teams were seeing  with similar guys.  It's not like any of those guys left here and unlocked something we didn't either.

I think have just been so atrociously poor at developing hitting that by comparison pitching looks good. It’s not like the Sox lead the league in pitching statistics every year. They’ve just hit big on three left handed starters.

[southsider2k5]

9 minutes ago, almagest said:

I think have just been so atrociously poor at developing hitting that by comparison pitching looks good. It’s not like the Sox lead the league in pitching statistics every year. They’ve just hit big on three left handed starters.

I honestly wonder if keeping good guys healthy is being extremely undervalued here.  Is what the White Sox got out of Chris Sale better than the Red Sox who saw his stuff tick up, but lost him for long stretches?  Then the Red Sox had to drop to the next option to fill in for Sale, meaning an even bigger drop-off.  If we have to sign six Mike Clevingers off of the street to cover our injuries is that worth getting an extra tick or two in a shortened season from guys like Davis Martin?

[WestEddy]

4 hours ago, Look at Ray Ray Run said:

The problem is moving data to a cloud doesn't actually improve outcomes in anyway. It's merely a way to provide access to the information, which in theory can provide you with more opportunities but as someone who has worked through many-of-these types of moves, it's not worthy of celebrating or citing as a large advancement. Properly defining the data also has value, but only if you can derive meaningful outputs from the data.

Yes it is. 

The example given in a previous interview was that a player would get promoted, and it would take days for his data to catch up to him, so the coaching staff would have literal days of starting from zero with this new guy until they had access to his data. Now the access is instant. One is clearly better than the other. 

I wouldn't think that in a "fluff" interview for general consumption the team will be revealing their proprietary models, or competitive philosophies. 

[fathom]

26 minutes ago, almagest said:

I think have just been so atrociously poor at developing hitting that by comparison pitching looks good. It’s not like the Sox lead the league in pitching statistics every year. They’ve just hit big on three left handed starters.

I think playing games in Charlotte hides a lot of faults that the hitters develop 

[almagest]

20 minutes ago, southsider2k5 said:

I honestly wonder if keeping good guys healthy is being extremely undervalued here.  Is what the White Sox got out of Chris Sale better than the Red Sox who saw his stuff tick up, but lost him for long stretches?  Then the Red Sox had to drop to the next option to fill in for Sale, meaning an even bigger drop-off.  If we have to sign six Mike Clevingers off of the street to cover our injuries is that worth getting an extra tick or two in a shortened season from guys like Davis Martin?

The Red Sox probably still win that World Series without him (108 wins and 11-3 in the postseason) so I’d say yes. I think you have to do that with guys like Davis Martin, though, otherwise they wouldn’t make it to the bigs, much less be effective there. You can probably take it easier with the guys with 90th percentile+ stuff. 

[southsider2k5]

8 minutes ago, almagest said:

The Red Sox probably still win that World Series without him (108 wins and 11-3 in the postseason) so I’d say yes. I think you have to do that with guys like Davis Martin, though, otherwise they wouldn’t make it to the bigs, much less be effective there. You can probably take it easier with the guys with 90th percentile+ stuff. 

So if the Sox got say Tyler Glascow out of Chris Sale because he was incredible, but hurt half of the year, that was better?  

[almagest]

12 minutes ago, fathom said:

I think playing games in Charlotte hides a lot of faults that the hitters develop 

I’ve been a big Charlotte fan but I’m starting to think the Sox should look for a better affiliate that doesn’t allow for 87 mph exit velo 42 degree launch angle homers

[almagest]

9 minutes ago, southsider2k5 said:

So if the Sox got say Tyler Glascow out of Chris Sale because he was incredible, but hurt half of the year, that was better?  

No, I’m saying the Sox got the better of the deal for sure, but Sale has overwhelming stuff. He’s hall of fame caliber and can pitch at 85% and still be great. Guys like Davis Martin need every advantage they can get, consequences be damned.

Edited June 25 by almagest