Rates & Barrels - Saberconference Takeaways, Late-Season Pitching Strategy & Sweepermania

Episode Date: August 15, 2023

Eno and DVR discuss the disturbing allegations against Wander Franco that have prompted the Rays to place him on the restricted list. Plus, Eno shares a few takeaways from this weekend's Sabermetrics, Scouting, and the Science of Baseball Conference in Chicago, before digging into several mailbag questions about late-season pitching strategy and the effectiveness of sweepers as the league-wide usage of the pitch increases.  Rundown 1:10 Wander Franco Placed on the Restricted List 3:57 Eno's Trip to the Sabermetrics, Scouting, and the Science of Baseball Conference 19:24 Training Stuff+ On Previous Years & Players Making Adjustments 31:33 Checking in on the 2023 Sweeper Craze 40:04 Late-Season Starting Pitching Strategy 51:54 The Stone Age of Arm Health? 59:26 Using Defense in Fantasy? Follow Eno on Twitter: @enosarris Follow DVR on Twitter: @DerekVanRiper e-mail: [e-mail] Subscribe to The Athletic at $2/month for the first year: theathletic.com/ratesandbarrels Check out these offers from our ad partners... HelloFresh: Go to hellofresh.com/50rates and use code 50rates for 50% off plus free shipping! LinkedIn: Right now, you can try LinkedIn Sales Navigator and get a sixty-day free trial at LinkedIn.com/rates23 Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to Rates and Barrels. It's Monday, August 14th. Derek Van Ryper here with Eno Saris. Eno, fresh off a trip to Chicago, hence the late afternoon Monday record time. You'll be hearing this episode a little bit later than usual. Fresh is not the word I'd use. You know what? I was trying to round up and help you.

I think our listeners, our loyal listeners especially, know that travel is harsh on you. Especially with the... I'm an eight and a half hour night guy. You know. Between the beers and the 8am conference starts. Yeah. That's just how it goes. We'll talk about a few takeaways from that

conference. We'll talk about some mailbag questions including a good one about late season starting pitching strategy. We had a question about sweepers. One asking if we're still in the stone age of arm health. So a lot of pitching related things to get to. But the big news shocking the baseball world over the weekend, Wander Franco has now been placed on the restricted list. Major League Baseball is looking into social media posts that surfaced over the weekend.

The posts allege that Franco had an inappropriate relationship with a minor. So disgusting allegations at this point. And there's not a ton that we know. The latest that I saw as of Monday afternoon is that he'll be on the restricted list until at least August 22nd. And we know from similar instances, a lot of times the league's investigation takes several weeks to be completed so there's a lot of uncertainty just in general as far as when we might have more information that comes to light yeah and it's uh you know being a part of the the coverage and knowing now you know how the athletic works in times like these like it's a really difficult process right now uh you know reporters want to give as much information as possible but there are legal ramifications for saying the wrong thing

uh and uh in the case of a minor um you know there's uh even sharing the allegations is problematic like those are you know pictures of a minor and uh so right now it's just one of those really weird spots where you just have to kind of respect the process i think and i think there i understand the need for information and the desire for information especially when put up against the fact that now he is suspended it feels like you know we should have more information about why he's suspended um but because i think there's a minor involved in this you kind of have to be respectful and take the steps you're supposed to take and you know that includes asking the team for comment asking them the you know asking his agent for comment getting the comment from the

player and uh it's not really i mean it's it's not really good journalism to say much before you have those things right because you know you're then you're just kind of salaciously retweeting a picture of a you know a young girl kissing wanda franco it's just, that's not cool either. So, you know, I think we've just got to wait. And, you know, I hope there's some sort of middle ground of truth that, you know, makes some sense. Because right now it just seems really sad. Yeah, right there with you. But, you know, we'll share more information once we do have something to pass along on this situation.

As far as the trip goes, getting to Chicago, getting to listen to a lot of very bright folks talk about the game from a lot of different angles. It sounded like it was kind of a good late season opportunity for you just to get the the brain working in different ways like thinking about problems and in ways maybe you previously hadn't so i was curious all the different speakers you had a chance to listen to folks you got to talk to during the trip you know any any major takeaways either things you're thinking about working on or things you heard or learned that you thought were really interesting from that trip yeah i mean the the bulk of it for me was uh discussions about stuff plus um and there were three presentations that um included stuff plus in some way or another and and tried to improve on some aspect of it um or or get better results than it um and so i'll get to those in a second otherwise um you

know some of the cool uh pieces presentations i saw one of them uh by dr alan nathan was really interesting he found that um bats with higher frequencies like that's what i mean like like sound frequencies yeah bats with higher sound frequencies uh produced higher exit velocities and has something to do with how the bat vibrates and so uh he created a really simple test where you kind of you hold the bat kind of by its handle and then you take the ball and you you kind of tap it along it until you find the place where it vibrates the least that's kind of like harmonic center of the of the bat and then at that place you uh ping it so it's kind of like uh on the skinnier side of the barrel it's not all the way down on the barrel um and uh you

sort of just take the ball and ping it against the the baseball bat and if you have two different wooden baseball bats and you do that you'll hear two different sounds and the higher sound is the is the bat that's going to produce higher exit velocities so i thought that was i love how uh it also comes with like a pretty simple test you know that almost anybody can do and it and i kind of goes hand in glove with i swear to god i've heard players talk about because i did this piece um with keenan long at long ball labs and you know where we talked about this on the pod too about you know how different each bat can be um and i swear when we were talking to players they would

they would some of them would talk about like sort of flicking flicking it and listening to it and like and listening and the joke is like you know oh i hear more hits than this one but like literally you can that is pretty remarkable so even with that test it's not there's there's not some other piece of equipment that goes alongside the bat as you're doing the tap test? I mean, he did it with more robust equipment where that's capturing the precise MHZ, you know, like whatever the measure of frequency is, you know? That's megahertz? Yeah. the the measure of frequency is you know i think it's megahertz megahertz yeah so you so i guess you're he had something that was like giving precise megahertz represent uh uh precisioned

uh measurements and that helped him of course and plot it all and and do you know good research but you know in terms of he ended it with a real subjective thing where he's like, hey, this one was lower and this one sounds higher. You can hear it. And the one that was higher produces better exit velocity. So that one blew my mind. Stuart Wallace did a pretty cool presentation about how neck mobility is an unexplored risk factor for elbow injury. And he was able to show that shoulder strength goes down for some pitchers depending on how far they can move their neck

towards their shoulders. So that was an interesting thing because I think the neck is not really something that people measure force on or think about that often and so he made a case that there should be more research there um sean o'rourke uh did a cool presentation about uh whether or not baseball data can be democratic. It was about the delicate thing when it comes to baseball data of measuring the needs of the many versus the needs of the individual. That's how he set it all up, was democracy is this interplay

between wanting to do quick action action wanting to do unified action uh but also wanting to let people have autonomy and also you know have have a vote right you know it's not just about voting for stuff it's that relationship between the many and the few um and so you can think about that in baseball data and be like, yeah, you know, teams and baseball writ large wants to, for example, let's do something that's altruistic. They want to keep their players healthier, you may want to ask every player for detailed biological, biomechanical measurements of their bodies that go beyond sort of reported data, which is like he had a double today in the game. Now we're talking about his vagus, his, you know, ability to rotate things,

his strength and mobility in different joints. Some of these things may be unchangeable. And therefore, you could be giving the team information about yourself that is health information that could be used against you. He does not have good hip rotation or hip mobility, and we found that guys... Neck mobility, I just gave you a piece. What if he finds that neck mobility is a risk factor,

and now teams measure everybody's neck mobility in spring, and you give them the opportunity to do that and let them do that, and now they know that your neck isn't mobile and you're going to get hurt, and so you're on the list of pitchers to trade before he gets hurt. opportunity to do that and let them do that and now they know that your neck isn't mobile and you're going to get hurt and so they're you're on the list of pictures to trade before he gets hurt yeah they could use it against you there they could use it against you in an arbitration case in some way i mean what was nice is that uh sean pointed out that they they're not allowed to use

an arbitration because there's a collectively bargained process in arbitration in which the union has power and has said nothing wearable is allowed nothing that comes from wearable tech is allowed in arbitration that's in the cba so his solution was if we do want it to be democratic then we have to basically organize and have to have a minor league union and have to have, uh, maybe even a college player union because these college players are being measured in the same way. And that, that allows you some control over the data and you get to, you know, maybe get a quid pro quo, get something back for giving them access to that, you know, something that, you know, better pensions so that the people who are hurt you know you know better medical maybe something like that but you you you if you organize you have

some power over it and you can maybe have a democratic process around that data because without that power you're just it's just you're just taking the data you know taking the data from the individual well good on the players association for being ahead of the curve on on that yeah yeah good good for them to think about that issue and it's it's definitely something that i talk to with players all the time trevor may for example is recording every uh little bit about his life in terms of you know resting heart rate in the morning sleep that he had how much hydration every little bit he's he's recording that and i asked him if he would give to team access to that and he said sure because i i want to know it i want to know

as much as i can from it i want to you know learn from it and i also just don't think as an older reliever that you know there's anything in here that that'll get me in trouble basically he's like you know they're measuring me based on stuff plus and and fastball velo like they don't care about how much i sleep at night it might just help me figure things out so um there was um those were some of the uh particular ones i liked oh sarah thompson from sports info solutions had an interesting piece that has no fantasy value really but it was talking about how um the uh what's what's the impact of directional momentum on on an infield play so if you've got a guy who's going towards second base to get a ball and he needs to throw it to

first you know versus a guy who's gathering the ball on the way to first um and she went and they i think they revised some of the stuff in defensive run saves to account for this because plays i think she had some play by mookie bats where she was like, on plays going away from first base of medium difficulty otherwise, he could be 90% if he's going towards first base to get the play, and more like 65% when he's going towards second base to make the play. So I thought that was pretty interesting as well. But yeah, so the last part was a lot of talking about Stuff Plus. We had a presentation from Scott Powers,

who used to run the Braves, used to be in the Dodgers R&D department, and now is a professor of sports analytics and um at rice and university and so scott powers and vincente iglesias worked together to do something that was very similar to stuff plus in output but um in terms of how they modeled uh it within it was different and one of the things that uh they considered in their model that i did not really consider much in ours was the amount of noise that can be had around a pitch trajectory so for example you know a pitcher throws a an 88 mile an hour slider with uh you

know two two vertical whatever right and what we found was if he throws that five times then you can you can say pretty well that that's that's what a slider looks like you know there's not actually that much noise around it but what they did was there is some noise and you know sometimes it's 86 with four vertical areas you know i mean like there is some like difference from pitch to pitch and so they threw that uncertainty into their model um but uh they weren't able to beat a regress stuff plus so you know i thought it was an interesting idea um from the standpoint of that. There was another Stuff Plus model done by Cal Aldred

who used to be in the analytics department for the Blue Jays. His was pretty close to the actual Stuff Plus, but he did some stuff where he kind of did some of his own

pitch classifications so he kind of did some clustering analysis to just be like you know these i'm going to do my own pitch classifications instead of taking from anywhere because there there is some we've talked about this on the show there is some like there are some pitfalls that can that can happen in stuff plus due to uh pitch classification and and then privately i was talking uh with a lot of different people about um about stuff plus and the fastball. Then people have two fastballs. What do you do in that situation?

And what we just do is we base it right now off the number one hard pitch thrown in a game. So it's almost defined by game, and you're just going through the foreseam the most. That's his primary fastball. We're just like well you threw the foreseam the most that's his primary fastball we're going to define everything off the foreseam right um the problem with that is that maybe against lefties they never throw the the sinker you know and they're forcing against lefties but then maybe against righties they throw a lot of sinkers and so you know one thing you could do is split every pitcher in two and just be like this is andrew haney against lefties and this

is andrew haney against righties and i'm going to define all the stuff plus based on the primary fastball in those situations and then i'm just going to slam it back together again it could it could work we'll we'll look at it for sure um you know it's something to think about another thing to think about would be and this i think is better for player development than it is for uh evaluation or fantasy which is you could create a stuff plus say call it stuff four or stuff two and basically have a whole set of stuff numbers based around their four seam and then a whole set of stuff numbers based on their two seam now that why i think this would be cool in player development is if you're a pitching coach you can be like whoa okay all right i see some pitch pairings here that we need to do and like oh this pitch is way

better if he pairs it basically with the two seam than the four seam and then maybe even you could come up with stuff we're like man all his numbers uh are better stuff to us you know we should just be spending more time improving his sinker his two seamer you know just because it seems like all the stuff is better on that side that's that's the direction we should go um and it might even give you hints as to oh man these pitches are better off of a two-plane pitch than off of a one-plane pitch you know that's another thing you could sort of think if you saw a bunch of stuff too that was better anyway that would be hard then to sort of remash together to get a one number. And for evaluation purposes, it seems like that's a player development answer.

But it does make you think about applications for Stuff Plus and how it's used and how it's best used. And that was a primary mode of conversation for me over the weekend. That was a primary mode of conversation for me over the weekend. So we had a couple of related questions, even though I don't think the listeners who sent these questions were at the conference, but they're very much on topic with some of these concepts that you're bringing to the table. One of those is a question from Brian, and it's a Stuff Plus training data question. Hearing Eno talk about the league's hitters adapting to sliders and how sinkers seem to be more effective this year got me thinking about how Stuff Plus changes over time. Have you looked into how much training data is optimal when fitting the Stuff Plus model? As an example, are sweeper slider stuff numbers inflated for 2023 because they used 2021 data for model training and hitters hadn't seen as many sweepers at that point

and hadn't had time to adjust their approaches. If so, are there any trends that would lead to a low-dump approach that goes against particular stuff numbers and bets on other profiles? It leads to a philosophical rabbit hole since it would imply a pitcher throwing the exact same pitches in two different years might have different stuff numbers, but maybe that makes sense because a pitcher's quality is dependent on the hitter they're facing's ability to hit it yeah i think there's a couple interesting things in here which is that

like this year sinkers are performing better relative to other pitches than they have in pitch in the pitch tracking era and this, sliders are performing worse relative to the other pitches than they have in the pitch tracking era. And so you could come away with a few conclusions. That one is that maybe we're approaching too many sliders. Maybe this is an hour throwing, or maybe the hitters are being trained to hit sliders because they're seeing so many sliders from pitchers. So basically the game is training them to be better hit sliders because they're seeing so many sliders uh from pitchers so they're

they're basically the game is training them to be better against sliders and then the opposite would be true for sinkers is they just don't see good sinkers anymore but also you they don't see bad sinkers anymore you know the number of sinkers thrown goes down down down who's left throwing sinkers people who have good sinkers um so there's so that's, that's one thing is to think maybe okay, all right, maybe like a slider heavy guy is not going to be as good this year. And maybe there's some there's definitely some examples out there of guys with sinkers that seem to be outperforming some of their projections and some of

their their modeling. So that's a simple kind of answer there. But there's a deeper kind of answer there but there's a deeper more philosophical question where i saw you know criticism of stuff plus is saying of course it's going to beat fip or sierra all these other things because you know it's born yesterday basically you know like it's the newest it it was trained on 21 and 20 2021 and 22 data so it knows what the run environment is it knows what the ball is it knows things about how the game is played right now um and so it's only going to get worse as we train it less going forward and i and i totally get that uh criticism it is definitely trained on more data like i know for example there's another stuff plus

out there where that person has trained it against 2023 data and is getting some better numbers than my stuff plus but it's like you know so i've been on the other side i've been like wow well you just trained it on 2033 data of course this is better you know uh we have made the choice to not uh update stuff plus during this season because we want it's the first season it's on fangraphs we just want to you know have it be the same all year not do anything not do anything under the hood gather some different things we want to do and then maybe do it in the off season so that's that's our plan but i want to push back a little bit on the idea that because it was trained on recent data it's it's lesser and in fact wonder if that makes it is that is that a is that a feature maybe and not a bug and the my understanding is this would you do you think you'd be more likely

to know something about baseball the way it's been played the last three years you know with the ball that's being played and how it is and how and how strategy is right now or to know something about baseball that's been true for 130 years yeah generally i think you want the more recent snapshot because the game changes and changes and then changes some more and you'd have so much noisy information that would be i think worthless based on how it's played currently yeah we would have no i i like if we were running stuff plus and it was still the stuff plus that was trained in like it was trained in 1986 and was born in 86 and and through you know 84 through 86 uh was the birth of stuff plus then i think it would still it would love the heck out of sinkers because people used

to throw sinkers and hitters weren't necessarily uh grooving their arm their their their path their bat path to like destroy sinkers but then in the late 80s early 90s the ball started flying further 87 is one of the first big ball flight years where the ball just changed over one all of a sudden in one year and so you start getting this better ball flight and if you have better ball flight then what are you going to do? You're going to build a swing that can take that low sinker and put it in the seats. And so you had this massive change in how hitters approach things. And now, no matter what kind of model you're looking at, it'll tell you, you cannot throw a sinker righty on lefty.

Do not do that. So the game changes. And so, I don't know. That's something to think about long term. And then there was one last thing that came up that might be relevant. It's okay. I'll have more for the the lodum angle of that question down the road because i have to think about what exactly like what holes i want to poke in

everything before there needs to be some some reasoning behind lodum before i like try to scale that concept out to something i would actually trust for team building purposes well let me let me give you some some some food for thought maybe because uh there was a an attendee uh that came up to me uh sam walsh and his twitter handle is samuel j walsh uh young student just graduated looking uh for a job in baseball and he had a very interesting critique of not only stuff plus but xera x XWOBA, some of the expected stats. And it goes like this. So if you're doing this work where you have expected, you're trying to create an expected metric,

most of the time, the first thing you do is try to remove context and you're trying to just look at the probability of hitting a single or hitting a double you're not trying to predict something like rbi you know in in this because what you're trying to do is predict just the likelihood of these different events and you want the cleanest events so it's first second you know double single triple maybe ground ball fly ball you know maybe predicting the launch angles and exit velocities like we are some of that and stuff plus so but you want it to be devoid of context because you you just want to be uh trying to create these these context lists because context is where team stuff comes in and context is where that's the noise of the everyday game is context was there was there a guy on first and second

when you hit your homer right or was there nobody on base when you hit your homer i mean that's that could be the difference in the game but is there is there any reasoning was he more likely to hit the homer when there was first and second or no was he more likely to hit the homer there's nobody on first and second you know so it's like you would try to take that context out and be like i just want to know the likelihood he's going to hit a homer you know what i mean so that's that's sort of the underlying thing. But here, here's an interesting thing that happens. If you do that, you don't maybe, and this is Sam Walsh's point, you don't maybe value double plays enough. Right? Because you're just like, what's the value of a ground ball in all situations?

You know? And so, you know, you're trying to take context out by just saying what's the value of a ground ball what's the intrinsic value of a ground ball and you don't necessarily you do you start talking about what's the likelihood of a single and you don't talk about what's the likelihood of a double play in the current structures of xera xwoba and stuff now here's a problem with that if you start picking putting your finger in the model right and you're going to say hey i don't think we value ground balls enough i'm just going to take my finger in here and be like more value on ground balls because of double plays right well somebody pointed out well you know what's going to happen is that your value for

sliders and four seams is going to go down you You know why? Because you use four seamers and sliders for whiffs. And if you're pushing the value around and saying, oh, I need more double plays, then the contextual stuff for four seamers and sliders starts to come in and be like, well, maybe I should put my finger over here and put some more value on four seams and sliders because they get you whiffs if the bases are loaded and it's three two three two count they're not gonna throw a sinker you know what i mean so then you then you become this person who's like putting a bunch of fingers everywhere just trying to be like put like trying to move things around so i uh i'm done to talk to sam some more and um there there may be something that we can do about it. And he thinks that would greatly improve the model on changeups and sinkers, which is still one of the weaknesses of the model.

So, you know, and I also wrote on Friday about, you know, where there are some pitcher groups, starting pitcher groups that seem to do better than Stuff Plus and Pitching Plus. groups starting pitcher groups that seem to do better than stuff plus and pitching plus and what i found was that you're slightly more likely to be better if you have more pitches it's the stuff we've talked about here more pitches uh and a change up so uh if you're looking for pitchers uh i i like emerson hancock i think is really interesting in this analysis plus change up that stuff plus doesn't love yet it's in a small sample so you don't know however the change up is also not getting whiffs in the game but could it be a great double playground ball machine you know type of pitch you know um so but he also doesn't throw like five pitches he throws three pitches but you're talking about a guy who throws three pitches has command has a good change up he may be a guy that you don't sort of hue only to stuff plus on there's so much context that i think is

valuable but how pitches interact with each other right you have multiple pitches but are they the right pitches like do they all sync up in a way where you get that ideal banana peel effect as we've talked about in the past where you get your four seamer with ride you get your cutter you got something that cuts into a right-handed you get like all the different possible movement profiles you got something with some drop that's sort of the ideal but you could just have a lot of pitches and they don't work well together because bryce miller you didn't build an arsenal that makes sense if you thought that was comfortable maybe you can see that each of those pitches coming out of the hand or maybe they don't look like each other you know yeah so

i mean i think i think that's important to to keep in mind too because if you just say i'm just targeting anyone who has five or more pitches that's my strategy it's like um okay that it's a strategy i don't know if it's going to lead you to a constant success and the other question we got that's sort of related to this i'll fold this in now too is from our loyal listener, OJ. OJ writes, I find this sweeper craze confusing. Let me back up. OJ's first line was,

hope you guys are dealing well with your separation, which we're just, you know, we didn't see each other as much as you'd think for as close as we live together because we're just busy, you know, all that stuff. Yeah. So I find this sweeper craze confusing. Tim McCarver and Ralph Kiner always used to say the worst pitch in baseball is the hanging slider as they explained the hanging aspect was a function of getting too far under the ball so you're almost throwing it

like an underhanded frisbee toss the ball then breaks horizontally instead of vertically as long as the batter's on the right swing path the ball will be there the sweeper seems almost like an intentional hanging slider is there some amount of break it has to have before it becomes effective can you ever throw it to an opposite side batter it seems like a lefty hitter would just see it forever coming from a righty and most of all is it working are the guys who are throwing it succeeding with it so a lot to get to here i think is it are we comfortable saying if you hang any breaking ball, it's not great. Poorly located breaking balls are going to get hit.

But it is harder to spot a hanging sweeper because they are thrown high in the zone more and they don't have as much drop as other breaking balls. Right. And I think you had that piece way back in April and you had one of the examples I want to say was the Pablo Lopez gif in that example where hitters swing under them. They can swing and miss right underneath the sweeper because it doesn't drop the way they expect it to. And that's sort of unique. But as far as the usage of the sweeper, this sort of connects with the idea of, well, what else do you throw? Usage of the sweeper. This sort of connects with the idea of.

Well what else do you throw? Does this fit the other pitches that you throw? I mean Shohei Otani. Throws a sweeper I think more than anything. This season. At least by the savant classifications. And he throws it to lefties and righties. So it's not just a pitch that he uses only against righties.

And by all accounts it works really well. In part because he's also got a four seamer. And a cutter. Those are his three most used pitches. Sweeper, four-seamer, cutter. So how much of this is just having the right complementary pitcher to make the sweeper as effective as it can possibly be? Yeah, so people are having more success against the sweeper than they ever had. And I had the piece of people having more success against the slider than they ever had.

This is the improvement. So last year, there were 21,000 sweepers thrown. This year, they've already been 22,000 sweepers thrown. So we're going to go up by at least 50% more sweepers thrown this year. by at least 50% more sweepers thrown this year. Last year, batters hit 194, 247, 330 against the sweeper. That's why everyone started throwing the sweeper. This year, there's been a massive improvement. This is the best anyone's ever done against a sweeper in this database as a league.

It's 203, 256, 354. That's a pretty big improvement. That's 24 points of slugging in one year on the pitch so what you have is this is such a big craze that i think more people are throwing it to lefties more people are throwing them in hitters counts more people are throwing sweepers the hitters are starting to see it and train against it and that means more bad sweepers so some people are throwing sweepers that shouldn't be maybe bryce miller and uh then you have uh the the the slider well the slider the just the slider is being hit better than ever by itself so even if you take the sweeper out uh this year it's being hit 224 279 384 so despite there being this big jump up on how hitters are hitting the sweeper that 354 slugging for the sweeper is still lower than the 384 slugging for the slider

so the sweeper is still a very powerful pitch i would expect there i expect there to be more and more of it because even with this regression, you found that hitters have not learned it completely, and they're not even hitting it as hard as regular sliders. And that.354 slugging, I would probably venture is the best of any pitch type, but I'll check change-ups real quick because they do kind of reduce slugging. But yeah, so sweeper's still a good pitch. It's not as good as it was before. It's way more popular. Change-ups this year have a.381 slugging.

So the sweeper is still the best pitch in baseball, even though it's worse than it was before. Yeah, the reasoning on that, that all makes sense. The dilution especially. More people are throwing it. More people are throwing it doesn't mean they're all good. That's just a thing. It depends why you're throwing it. It might be the desperation adjustment

for some guys that are on the bottom of the roster just trying to find something that keeps them in the league. Well, maybe it works and maybe it's just as bad as the other pitches that were steering them out of the league in the first place. Maybe they's just as bad as the other pitches that were kind of steering them out of the league in the first place right maybe they're just a bad pitcher right it's uh harsh but unfortunately probably very true one last bit though he was talking about the hanging aspect yeah i think this is why some people say well the sweeper's been around

forever it's a slur and that's exactly what the sweeper is not like a good sweeper and this is also probably why there's some good sweepers and some bad sweepers. A good sweeper uses seam shifted wake to drop less than you'd expect. And what's happening is the seams are gathering at the top. So you get a little bit of a spinning slider dot at the top where the seams gather there. They create a wake and that keeps the ball from dropping. It sort of pulls the ball up. And so, you know, that's what you were alluding to with the Pablo Lopez thing. They swing under it. they create a wake and that keeps the ball from dropping it sort of pulls the ball up and so you

know that's what you were alluding to with the Pablo Lopez thing they swing under it it stays up and swings under it so it's in fact less of a two-plane pitch and more of a one-plane pitch that looks like a two-plane pitch so that's why we've come up with the sweeper designation because I think to the batter it looks like a slurve and then it doesn't drop as much as they expect and that's a sweeper and that's kind of difficult if you kind of expect a ball to have when when you see a ball have that much uh horizontal movement I think your brain thinks this is a two-plane thing this is a curveball you know and it's basically like a curveball without any of the curveball drop you know it's just like if you want to call it a frisbee,

now I can get on board. It's a frisbee, but it's not, I don't think, a slur. Anyway. I was just about to add the screenshot I was looking at of Otani's pitch locations. I always find these to be pretty fascinating. And again, these are the three pitches he throws the most,

the sweeper, the four-seamer, and the cutter. I'm sorry if we're describing something on the screen. Again, if you're watching us on YouTube, this makes sense, but we'll try to make it make sense if you're listening to the podcast version. Part of the reason I think this is working so well for Otani is that he has a pretty consistent location strategy. So if those pitches all look reasonably similar coming out of his hand, they're going to cause some problems. reasonably similar coming out of his hand they're going to cause some problems just looking at where he goes it's kind of low middle in on the zone like not middle in but low like middle third part of the strike zone for all those pitches for each of these pitches the sweeper is going to take off uh towards the lefty batter box in terms of movement you know the four seam is going to jump

more and stay true and the cutter is going to be somewhere in between them in terms of movement, you know. The foreseam is going to jump more and stay true. And the cutter is going to be somewhere in between them in terms of movement. So, yeah, you're right. That's three pitches coming from the same trajectory that are going to do very different things, at least horizontally and also kind of vertically. So, you know, maybe there's some sort of, you know, quote-unquote tunneling happening here.

And you can also tell from the seat map that Otani is trying to sneak his sweeper in back door against lefties and i think that is probably the best way to do it because what happens with the sweeper is it's so horizontal that the lefty sees it forever and knows what pitch it is and i think usually takes so the only way you can kind of sneak a sleeper sweeper by a lefty i I think, is try to pepper the outside part of the plate and don't miss middle middle.

But that's true of just about every pitch, right? You can't miss middle middle with most things. So, anyway. Good question, OJ. Thanks for sending that in. Some other questions that came in for today's episode. Alan was curious about late season starting pitching strategy. In this particular instance, Alan's situation is a 16-team head-to-head categories league

with quality starts instead of wins, currently in first place in his division and overall. So it's playoff time for Alan. It's going to be a few weeks away. Here's the problem. John Gray, Nathan Evaldi, Freddie Peralta, Tanner Bybee, Gavin Williams, Bryce Miller, and Bobby Miller. Those last four especially coming up on innings caps. So Allen's worried he'll be left with the old guys and then having to troll the waiver wire just as the playoffs begin. So especially in deeper leagues with not much pitching out there. And I just saw some breaking news from our friend Katie Wu

at The Athletic that Steven Matz's season might be over the last strain. There's one more pitcher who was pitching well who might be on the shelf for the rest of the season. He just can't. It's brutal. What is the adjustment here? How much should you worry if you're relying on some combination of those young starters or others that we've talked about throughout this season? How do you get from here to there, especially in head to head leagues where your most important weeks of the season might be the last two weeks of the season.

I think, I think he's right to do what he's doing. I think you give up a little bit of future value. If it's a, it's keeper HH, right? I think this is, it doesn't say it's a keeper league, plenty of people are in that spot yeah even if it's not keeper what you can do is give up a

little rid of risk for uh for you can trade some upside for surety you know what i mean like you could trade uh bryce miller and tanner bybee to somebody for uh what's a what's a like a Zach Gallant, right? Would that be impossible? Is that crazy? What's a little bit lower than Zach Gallant but somebody like him? Is Sonny Gray too far lower? Yeah, okay.

Or maybe you take one of those guys and trade him for Sonny Gray. But I'm saying sort of package the risk to get up higher, but maybe Bilo and Sandy there's a guy who gives you innings right try to do two for one on Sandy and who is somebody else

I believe in their innings up there maybe Eflin what about like you say Kikuchi he's pitching really well has the walk rate down to that career best mark so far. Still missing a good number of bats. I think I would use one of those guys.

Yeah. But yeah, that's not bad. So you trade in some risk for a veteran that doesn't have the same upside as what maybe a Bryce or Bobby can do the rest of the way. Let me see if I can come up with numbers for you here. I'm going to put the over-under on innings for Bryce Miller at 150, and he's at 110. He could probably, maybe he can go a little more than that.

He went to 132 last year, 133. Yeah. Okay, so 155. Let's put the over under on 155 uh the the the mariners are probably a little bit more let's get to the postseason and then worry about our postseason rotation when we get there more than you know let's shut somebody down that's how i feel you know they're kind of got back into the race a little bit. There's some pressure on them to do something. I would say they're going to use up all of his innings in the regular season and that means that if he's got another 50 pitches innings in him, I think he can make it to the end of the season.

Do you think Bobby Miller's projected total is a lot less than Bryce Miller's? Alright, so let's go to Bobby Miller. We've got a total for 2022 at 111. But he's also pitched fewer innings so far. So even if we give him 130, 135 as the over-under, he's only at 83. So he has a similar amount of innings left. But his team may screw around with his innings to keep him in the postseason

rotation. On the 3-0 show, I believe, we put Bobby Miller in as their number three in the postseason. I still think that makes a lot of sense because there's a very good chance that he is their third best starter when they've got everybody healthy. We talked about the Lance Lynn schedule. That's happening right now. He's dominating against that weak part of the schedule, so I want to see what happens when Lynn faces some better lineups in the next few weeks. But I think Bobby Miller being handled a little differently. We've seen the Dodgers do it with young pitchers before, too, where they pull back a little bit in September and they push them more aggressively

in October. But generally, there is enough room here where you don't have to panic about Bobby Miller or Bryce Miller at this point. Here's the guy to panic about. Yeah, who's your guy that you're worried about? I think it might be Tanner Bybee. And it's a little bit less about the innings because he has a fairly robust number of innings. Last year, he put together 133 innings, right?

That's pretty good. You want to put the over-under at 155, 160, I wouldn't argue with you. However, he's already at 123. So despite having a higher possible innings cap, Tanner Bybee is closer to it than the other guys. And my last piece of reasoning is the team context situation. I feel like if the Guardians fall out of the race, they don't pitch him.

And this isn't something that I think will happen September 1st. But when is your head-to-head competition? Like, when is your head-to-head finals? Because if your head-to-head finals is the last week of the season, I would be like, if they're six games out and there's a the last week of the season, I would be like, if they're six games out and there's a week to go in the season, Tanner Bybee's not pitching again. Yeah, that's a fair point. And I think for Williams,

looking at the same rotation, I think he is matched last season's total. If I'm doing the back of the napkin math correctly, 115 last season, 115 right now. It gives them 40 more before they start thinking about a shutdown, which Williams, we talking about here,

Gavin Williams. Oh yeah. Gavin. Same kind of problem. I love him though. I love his fastball. Um, yeah,

again, a nice innings total, uh, but same, same team context and he's already at it. So what would you give him as another, you know, 25,

30 innings so there they have a couple of pitchers that they're that are going to run up against their innings uh totals and be interesting to see what they do and then there's also the risk that because they have all you know a couple guys that are coming up on these numbers that they do something like a six man you know just to just to do just to do that. And that, that doesn't, that's not obviously as bad, but it does mean that you're going to get fewer starts from Gavin Williams. You know? So I think the number of starts for Gavin Williams might be left, might be five or six.

What do you think the best tactic is for acquiring the innings? You need a 16 team league. The waiver wire is pretty thin. I mean, if you have to churn a lot and the rules are first come, first serve, decent pitchers will come up. Is it going after players that maybe have shown us signs of turning around a disappointing season,

like where the surface numbers are still bad, but the more recent numbers are good? Thinking about Jamison Tyon in particular, where it really seems like he has settled into being over his last seven or eight starts the pitcher many of us thought he'd be going into the season i liked him as a mid-rotation filler around pick 200 i thought with the pitch mix i thought with the ballpark he was really safe leaving yankee stadium some of the problems tyon's had against lefties those would be less of a concern because he wouldn't get punished quite as much as he did when he's at yankee stadium and that didn't look

true for april and may but from about june on he's actually started to look like that guy are those the types of guys you're looking for who also don't have any real young pitcher innings risk attached to them either maybe because i mean if you want to trade you know your young hot starting pitchers that may have innings risk on them for a pitcher who's established and has been good all year well good luck you're probably not going to find a lot of willing buyers in that case or you're going to have to pair them as you mentioned before but give two and get one and even then that can be tough you know i tend to you know i tend to gravitate towards uh veterans with established track records that maybe aren't

pitching quite up to their numbers uh because you just feel like you know even a non-competitive team may just decide uh to continue pitching them so they get right you know what i mean um so i kind of think of aaron nola a little bit as like man maybe you could you could sneak that cheese past the rat you know like maybe you could get a you know a prime guy that people thought was a top 12 pitcher that just hasn't really pitched up to his numbers um and uh and get somebody like that um because i'm just worried that like let's say you take you know somebody who feel like he has more innings like hunter brown or you know somebody who hasn't used their innings up hunter green you know he's coming back um i don't know i feel like you run into just as much as those guys where maybe the astros

you know start coasting at some point or they go to six man because they've got a bunch of guys or whatever take dylan cease you know because he's more the veteran that has innings left but maybe he's just gonna he i saw some analysis that he changed his release point and his extension uh in order to improve his changeup, and that has messed up. His slider and curveball have morphed together, so he's kind of a two-pitch guy now. It's not a guarantee that Dylan Cease turns it around, but Cease and Nola I would feel amazing by.

That's sort of where I would look for something like that. I think Tarek Skubal would also be on my target list. Because of the time he missed, no real concerns about him getting shut down everything looks really good in that arsenal good swing and miss stuff so few innings they just want him to build up innings even on a bad team yeah yeah i think those are i think those are all viable paths it's a common problem you're playing a head-to-head league you play in a roto league too you're just worried about guys getting shut down and not having enough innings and keeping those counting stats rolling in during those final weeks of the season.

But I think also just think of the psychology of the offer. So the psychology of the offer is you're offering two hot young things, right? I think it doesn't make as much sense to then try to get a hot young thing back. I think you want to get a boring old thing back. So think about that in the psychology where you're like, oh, hey, take these hot, awesome hurlers, throw in 98, and give me your decrepit 92. Give me your Joe Musgrove.

He's super boring. Don't worry. I don't know if you can pull that one in particular, but that's the sort of thing I'd be looking at. You get something boring back. Thanks a lot for that question, Alan. We've got another question that is connected to the others. I love when things all connect together.

Sometimes it's because we purposely cherry-pick questions that work together, but when the mailbag is full of things that are related, it makes our job a lot easier. This question is about how we measure workloads the mason miller segment of your most recent episode led to confusion on why we use innings instead of pitches thrown we're talking about how long it might take the a's to build mason miller up to a full starter's workload probably a week or so ago not all innings are created equal a 30 pitch inning where a pitcher gets roughed up isn't the same as a three-pitch inning.

If you wanted to go deeper into the weeds, you can tally your pitches thrown in off-day bullpens and warm-up throws. More into innings, DVR and Eno made it seem like it's a race against time to get Miller to his high number to create a solid benchmark for the following season. Why not use AFL to increase innings or have him throw in-house simulated games?

See, the thing about this is, and this was something I thought about with Forrest Whitley several years ago. We were trying to figure out Forrest Whitley's workloads when he was coming off all the injuries. We don't know about all of the side sessions. We don't really understand or have a way of tracking those innings, those pitches, that workload as part of the season workload. So I do think there's a lot here to dig into.

But generally, what do you think? Are we doing this collectively, not just you and I? Are we doing this wrong, measuring innings instead of pitches and looking at pitcher workloads the way that we do? Well, there's us on the outside trying to approximate the work that they're doing inside without any of their information. And so that's why we look at former that past innings totals and uh you know i also think about how hard they throw how hard

they throw their breaking ball and so you know that's it's sort of generally trying to assess uh injury risk from the outside without any of the information they have then you have your kind of uh your organizations that are behind uh that are doing something very similar to what we're doing. And just sort of like guesstimating based on innings, then I think you have your more middle class play development systems, pitching health systems that they do. They wear sleeves, sometimes and directly measure stress on the elbow in throwing sessions. I would say that almost every team is tracking pitches thrown

when getting warm, getting hot, bullpen sessions. So I think the large meat of baseball is tracking a lot of what our question and question and answer it wants them to you know is is tracking it more on a minute pitches base level now the very best teams um are have a such such a robust collection of data that they will have they can marry direct stress on the elbow gotten from arm sleeves to you know maybe modeled stress on the elbow based on on what they found with direct stress and fatigue and they have a fatigue model and you know like the cubs just signed mike sun who had fatigue units and fatigue units is how how

how long you take between pitches how hard you you throw, how many pitches you threw, how much rest you got. And now that he works for a team, he gets to put five, six, seven, eight more features into his model to model fatigue better. So you know the Cubs at some point are going to be out in the forefront of exactly what we're talking about here because they've got one of the guys who is doing it in the public space. I know Casey Mulholland at Connect Pro has the sleeves. because they've got one of the guys who was doing it in the public space you know i know uh casey

mulholland and connect pro has the sleeves he has his own kind of sleeve where he's doing direct measurements he has his own uh sort of fatigue modeling he was close with my son and believed a lot of things he believed he told me for example in my running that when i do a peak run uh it's 72 hours recovery um and uh and so you know i've been thinking about you know i might do my peak runs on wednesday that means i'm running again on friday and i'm always fatigued and i can feel it um and so i'm not sure i can make my schedule work the way that you know casey mo holland wants my schedule to work because i have kids and work responsibilities and stuff, but that's something that pitchers will think about. Maybe if I throw 100 innings, I should not do anything for the next three days

because I have the 72-hour cycle. Maybe you'll see teams, when they push their relievers pretty hard, if it gets past a certain number of pitches, 25, 30, 35 35 pitches they will get two straight days off um and so i think the better teams are doing most of what he's talking about but it is hard for us on the outside to approximate all that without the same information yeah no i think that's uh it's well stated and i think it's still is it fair to say that the wearable tech like the sleeves it's still too early to know like how much of an impact that might be having in injury prevention like it's input so it seems to me like it's still in the let's just gather a lot of information and see what we can figure out are we further than that or is that where we're really at i mean i i think we're in the in in a

phase right now of figuring out some stuff about injury because you know jimmy buffy uh gave a not buffett uh jimmy buffy gave a presentation at uh the uh at the conference and he's works at reboot motion and what he does is he can give you biomechanical um advice consult consultation if you just upload video into his thing and he kind of runs it through his system gets certain angles he's done he has a model that's like modeled all of these things and related them to stress on the elbow and velocity and stuff like that so he can tell you oh you're you know your lead leg flexion is not where other people's are that throw harder you know like you're this or that or

you know what you do here with your body so that kind of analysis the fact that he's presenting at the saber conference means two things it means that some teams are doing more than that already you know enough that like you know this guy is presenting you know what i mean and he's comfortable presenting it like he's not he's comfortable presenting it because he's it's not something that nobody knows and he wants he wants people to know that he knows it that also means it's not something that everybody knows because he's at the conference trying to tell teams hey you can just send all your video to me and i'll do your biomechanical work for you you know that's why you know it's it that's how conferences work you kind of present you show a little bit of leg and then the teams are, well, that's interesting.

And then you, you show them a little more leg and then you, you tell them you'll do all their, their videos. So, um, we're in a spot right now, I think where there is separation between teams there.

And, but all the teams are interested in this space and are all trying to get better at fatigue modeling and biomechanical modeling so that they can both, it's always two things, keep their players healthy and have them throw harder. Right.

Which are diametrically opposed. That's tug of war right there, just pulling opposite directions. You want both of those things, but you may only be able to have one. And which one do you think a team will target if they have a choice? I gotta guess that

throwing harder is probably going to win. Probably throwing harder. Staying healthy is going to get thrown in the mud in that game of tug of war. Especially if you're in Tampa. Thanks a lot for that question, Joseph. I got a question here from Ted about using defense in fantasy. Why do so few fantasy

baseball leagues factor in fielding? Is it because scoring around the league is so inconsistent? Would it make a difference? If we had to drop a stat to factor it in, which stat would you recommend runs RBIs? Basically out of the typical roto stats, what would you throw out if you were going to add defense? I think the reason people don't do it is because for a long time, at least back when fantasy baseball started, there were very few useful ways to measure defensive value. I think we've, in the last 10 years, seen a lot kind of come to light that actually helps in that regard. And we've got more public facing things now too.

So if you wanted to use defensive runs saved or outs above average, you could do that. You could reasonably make that a category. Whether or not you should um i'm more on the no i don't really want defense to be part of fantasy because i think it's already built in as far as playing time goes i think understanding a player's defensive limitations or actual skills that actually gives you a better sense for projecting playing time. It's still in fantasy because it's still important for us to know because of playing time for sure. Yeah, I think it's present enough. I know it's a bigger part of games like score sheet.

Some of the sims out there, as far as your Roto Leagues go, I think we're better off not having it than having it. If I was going to cut one of the Roto hitting stats to add a defensive stat, jeez, which stat would I get rid of first? Runs? I guess I'd get rid of, I don't know. It'd be interesting because you would actually probably, there'd be similar players that you would hurt and help in that situation. There's a lot of more top of the order guys that are playing in the most prestigious defensive positions.

So if you kind of switch it over to defense, you get a lot of top of the order guys still who will be losing some value on runs. Another thing I think is just one thing that is impractical about it is just the nature of defensive statistics. So most of our stats that we use are counting stats. So most of our stats that we use are counting stats. And even though it says defensive run saves, it's not really a counting stat in that way.

You can have minuses. You can lose it. And also, if you did defensive run save, it'd be really weird because you'd play a whole year and you'd get a five would be an okay number. So what would the spread be there uh you have to think about the spread so then if you start doing something that is a counting stat and i have seen stuff like outs like outs made defensively um the weirdest stuff happens you need to look at who leads the league and out so it's like a lot of times it's first baseman. Right, yeah.

Like there isn't really a great defensive counting stat, I would say. Is it worth trying to make one? No. No. I don't think it really works that way, man. I think you can hurt your team on defense and you can help your team on defense it doesn't just add up out i was just going to see if i could pull it up on on baseball reference while we're talking here outs made no that's not what i want i want the defensive outs outs made pitching leaderboard

fielding leaderboards put outs put outs single season yeah so i have put outs and it's jt real muto nate lowe matt olsen spencer torkelson frederick freddie freeman jose abreu thai france like okay yeah the first baseman because they catch those throws yeah that's not really what we want to add what's this is that a sis yes this thing okay so now then you get the shortstop. So if you did a sis, Wanda Franco, Francis Boulendor, Ozzy Albies, Ezekiel Tovar, Andres Jimenez. So a sis could be your number. And you have, you know, a league leader will have like 400 and a bad, what's a bad shortstop? Ahmed Rosario has 209.

So you'd have a spread there. Again, that'd still be kind of weird. It'd still be kind of weird because you would have a category that has a much different spread than most places. Like, what other stat do people have 300 of right now? Yeah, it's not good. It's not a stat that I would want to use for that purpose. But I like the spirit

of the question. I just don't think you need to add a counting stat in this case because you have it. It's factored in. It's playing time. That's where your defensive value gets baked into fantasy. Thanks a lot for that question, Ted. We are going to go on our way out

the door. A reminder, you can get a subscription to The Athletic for just $2 a month for the first year at theathletic.com slash ratesandbarrels. Find Eno on Twitter at Eno Saris. You can find me at Derek Van Ryper. That's going to do it for this episode of Rates and Barrels. We're back with you on Tuesday. Thanks for listening. Thank you.