Effectively Wild: A FanGraphs Baseball Podcast - Effectively Wild Episode 1795: Measured Tones

Episode Date: January 8, 2022

Ben Lindbergh and Meg Rowley continue their “Measuring the Unmeasurable” series about studying difficult-to-quantify aspects of the sport by bringing on Rob Mains of Baseball Prospectus to banter ...about ESPN’s new Sunday Night Baseball broadcasting plans and discuss Rob’s studies about competitive balance, team mobility and inequality, starter vs. reliever performance, the magnitude and history […]

Transcript
Discussion (0)
Starting point is 00:00:00 Hello and welcome to episode 1795 of Effectively Wild, a baseball podcast from Fangraphs presented by our Patreon supporters. I am Ben Lindberg of The Ringer, joined by Meg R baseball podcast from Fangraphs presented by our Patreon supporters. I am Ben Lindberg of The Ringer, joined by Meg Raleigh of Fangraphs. Hello, Meg. Hello. We are continuing the series that we've been doing all this week about measuring the unmeasurable and talking to authors of research of some difficult to quantify aspects of the sport. So later in the episode, we will be bringing on the author of the site Harib's Hangout to talk about various research subjects. But before that, we are joined by our
Starting point is 00:00:51 pal and past guest, Rob Maines of Baseball Perspectives. Hello, Rob. Hi, Ben. Hi, Meg. Happy New Year to both of you. And to you. I think this is probably the last day that people are actually still saying that to each other. So we just got in under the wire, maybe, or at least that was the essence of the work conversation that I had today about when it's acceptable to stop saying that or start saying that. You can say it whatever you want. You can say it in June if you want. We should all just be having a happy year if we can. But we brought you on because you are a researching fiend and you are just constantly researching things. And often those things are difficult to quantify. But just before we get brought you on because you are a researching fiend and you are just constantly researching
Starting point is 00:01:25 things and often those things are difficult to quantify. But just before we get to some of the studies that you've done over the past year, we have some news about Sunday Night Baseball, a broadcasting personnel change for, I guess, what you would call the flagship baseball broadcast of the week, the most prominent broadcasting crew. And we have a change now from the A-Rod and Matt Vaskirjan booth that we have had in recent years. Now Vaskirjan, he had already left. A-Rod is now, I don't know whether you would say being relegated to or happily embracing, sort of a Manning cast kind of alternative broadcast on ESPN2 that will be more of a hangout, just kind of conversation with the Yankees broadcaster, Michael Kay. And the lead crew on ESPN itself will now be Carl Ravitch with two analysts, a hitter, Eduardo Perez,
Starting point is 00:02:21 and a pitcher, David Cohn. So what do you all make of the new Sunday night baseball broadcast cruise? I mean, I don't want to like speak ill of people on their way out the door, but this feels like quite an upgrade in terms of its general quality. Like he's not out the door yet. He's behind a different door. Yeah. Relegated to a different part of the building, as it were. I think that this is perhaps set up to set everyone up to succeed, right? They're put in their best light. I could see A-Rod being, well, if anything, he'll just be a conversational weirdo. That possibility exists, but that'll be engaging in its own way, at least for a couple of broadcasts. And then I think that that
Starting point is 00:03:05 that main broadcast crew is that's a really good booth. Like those guys each do well, I think, in their in their own individual broadcasting roles and I think have the potential to work really well together. So I'm excited about the main ESPN Sunday Night Baseball broadcast. It's been a little while since I've been able to say that sentence and actually mean it. So that's pretty cool. Yeah. What do you think, Rob?
Starting point is 00:03:30 Yeah, I agree with Meg. I thought that it was literally unwatchable. I didn't watch Sunday Night Baseball last year. And now I will be able to watch it. The question, Meg, have you watched any, because I don't watch enough football to know this, have you watched any of the Manning productions? Yes.
Starting point is 00:03:47 Now, I've heard that they are very fun and pleasant. And I kind of wonder what the appeal is going to be of the Michael K., Alex Rodriguez alternative. Well, it's, yeah. I wonder how- Yankee haters unite. I wonder how direct a comp they will try to draw there because, oh boy, I want to think carefully about how I phrase a not nice sentence. I think that one of the great things about the Manning cast
Starting point is 00:04:17 is that they intersperse fun conversation with the two of them diagnosing what is happening on the field and sort of predicting what they expect you know the quarterbacks in particular are going to try to do to you know be coverages and whatnot and they seem to have a more peaceful relationship with football as it is currently constituted right like it does not seem to be a tense relationship for them and so when they are describing the action on the field you're you're hearing from two like really great quarterbacks and you're just able to appreciate sort of what they know about the game what they're able to diagnose from their own experience about what the the play on the field might unfold as
Starting point is 00:05:04 and then sort of opine on its efficacy. But they're doing that within the context of what works for the game rather than that worked and it's why football is dying. There's none of that. So I think that the conversation piece might end up being the easiest part for A-Rod. I imagine that they will be able to get some pretty impressive and interesting guests. But he doesn't seem to be able to resist the temptation to say that baseball is going to hell in a handbasket. So I don't know how that part's going to play.
Starting point is 00:05:36 Right. I mean, my concern, I guess, is that, well, the Mannings or Peyton, at least, I mean, they're pretty engaging and funny, right? And off the cuff entertaining and they're brothers, obviously. So they have a close relationship. And obviously, A-Rod and Kay go back years as well. And it was reported that A-Rod seemingly preferred to work with Kay. But A-Rod being A-Rod, you know, will the experience of just listening to him in casual conversation be as engaging? And, you know, I like A-Rod on a certain level or sort of feel for him. He just seems very desperate to be liked, which is an unusual quality in a extremely successful professional athlete. And I think that leads to what I read as pandering often when he talks about bunting and sacrificing and small ball and all of that, despite being one of the best power hitters in history. I don't know whether that will be better or worse in this context than on the main broadcast, because there was that brief A-Rod renaissance right when he was a studio guy and everyone thought he was great in that context. was great in that context. And so might this be a bit more like that when he was just having fun with other ex-players and just kind of hanging out and ribbing each other? Maybe he's a bit better in that environment than he would be on the actual broadcast. And I think one of the complaints about
Starting point is 00:06:57 the ESPN Sunday Night Show lately has just been that they're barely paying attention to the game, right? It's like listening to a podcast while a game is going on we like podcasts but those things don't always mix well together it seemed like the game was sort of the sideshow and they just wanted to interview people and divert themselves in other ways and that is sort of explicitly what this broadcast will be i guess and so maybe it's good to cordon those things off and have the very hangout conversational kind of broadcast on ESPN2 if you're into that sort of thing, and then have the more meat and potatoes analysis and play by play on the main network. Mike Petriello put out something early that suggested to me that this means Eduardo Perez is no longer going to do the nerd cast. Yeah, you know, I was thinking, what does this mean for the nerd cast?
Starting point is 00:07:48 I think the nerd cast might be done. It might be, right. I mean, I don't know anything. I haven't asked. But because that was kind of the alternative to the flagship show, and that was on ESPN2, and because Eduardo Perez was one of the main people on that broadcast, it seems like that doesn't bode well for the nerdcast, which is sort of sad. On the other hand, maybe the nerdcast was like a traditional point, right? Like maybe the nerdcast was a way to get from the old school broadcast to the new broadcast.
Starting point is 00:08:21 Because like if the main broadcast has Eduardouardo perez and has david cone and cone is just fantastic i mean he's at the top of everyone's fantasy broadcaster list right like everyone wanted cone on this broadcast he is engaging he's funny he brings the player experience he's familiar with advanced metrics and very receptive to new ideas so if you have cone and perez on that broadcast and i think the former producer of the StatCast broadcast was promoted or transferred over to the main Sunday Night broadcast too, well, maybe it's not full nerdcast. Maybe you don't have a whole bunch of graphics on the screen all the time, which can be a mixed bag as it is, but you might have that same sort of intelligent analysis. So, you know, I'd be sorry if it meant less Jason Benetti and less Mike, obviously, but maybe that was kind of the proof of concept that got us to boog shambi being the main play-by-play guy with the two player analysts but i think ravage could do a decent job and and it's good if you're gonna have the three-person booth and x player analysts then it's good to have the hitter and the pitcher i think who can give you both sides so well and i i think that like the most
Starting point is 00:09:42 optimistic read is that like the nerdcast walked so this broadcast could run, right? I think that the goal, it's nice to have art created just for you, right? To feel like your folks are well represented. having a broadcast that we can all participate in that manages to satisfy multiple constituencies by being open to and not sort of in tension with analytics but is still comprehensible and enjoyable for folks who don't view the game through that lens is good because baseball is the most fun from a viewing perspective when we're all watching the same game that's part of why the playoffs are so fun and that is has always been part of the promise and potential of Sunday night is that we're like all, you know, getting together to bemoan watching the Yankees and the Red Sox one more time. So if we can do that in a way that manages to incorporate more of the audience into sort of a cohesive whole, I think that that's really great. Even if it does mean hearing a little less from some of the folks
Starting point is 00:10:44 who we like best in the business because you know the nice thing is we'll probably get to hear from them in other contexts and maybe they can be brought on to this broadcast as like right nerd nerd in residence your resident phone a friend crash the k-rod alternate broadcast too yeah i have mixed feelings about k because I grew up in New York listening to him first on the radio and first on TV. And so there's always a certain fondness for the person that you grew up watching baseball and heard all the time. And also there was one time when I was still a fan of the Yankees and I went on a trip to
Starting point is 00:11:20 spring training to see them play. I was a teenager at this point and I was on the same flight coincidentally as Michael Kay in a different section. And on our way to the baggage claim, I summoned the nerve to go talk to him because I had recognized him. And all I remember is like talking to him about how excited I was that the Yankees had signed Paul Quantrill, which dates this anecdote a bit. And also Paul Quantrill wasn't all that exciting during his Yankees career. But I remember Kay being just incredibly gracious and nice about this kid coming up to him and bugging him about a reliever that the Yankees had signed, which probably happens to him all the time. So I'm sort of grateful to him for taking that so well.
Starting point is 00:12:01 But also, you know, he can be a bit of a talk radio blowhard from time to time, which is his job. I mean, he is a professional talk radio blowhard. He's had a show on ESPN Radio in New York for years. So he is used to the conversational side and to just filling air for hours a day. So I don't think he'll have a problem with that here. Plus, maybe part of the impulse from ESPN's perspective is, hey, if we're going to put the Yankees on Sunday Night Baseball every week, let's just put Yankees broadcasters on here too. So whichever broadcast you listen to, it'll be just like you're listening to the Yes Network. Kony and Kay are here. You'll feel right at home. Come right in, Yankees fans. I think that you want someone who is able to talk. Having a little bit of take is good,
Starting point is 00:12:44 right? It keeps it moving. A little bit of take is good, right? Like it keeps it moving. A little bit of take is good. Too much take is exhausting. But I think that broadcasting strikes me as anything else. You know, my capacity to do it would depend heavily on me being well caffeinated. And you want to seek some sort of balance and equilibrium because you're having to appeal to such a broad audience if you're doing
Starting point is 00:13:06 it right and you're not going to be able to sort of hit that for every person at every moment of the game but having a little bit of this and that for everybody and still being able to weave it together is i think admirable yeah i thought you were going to go with well lubricated instead of well caffeinated because i might need a drink or two if I were going to be on a broadcast that might help loosen me up a little bit. Maybe the results would not be ideal, but I don't know. It would probably help me be more talkative, I think. But generally, all I want with a broadcast, either you got the good vibes broadcast where it's just fun to hang out with these people or just like don't deliver misinformation to me. Don't tell me things that aren't true.
Starting point is 00:13:50 And A-Rod has been quite guilty of that in recent years. And Rob, he's given you a lot of material, right? And I guess the good news for you here is that A-Rod will still be talking on national broadcasts. I think the really good news is that Cohn will be introduced to a national audience more so than he is already. And people will learn why he is so good at this. But also, A-Rod will be talking a lot, possibly even more than he was talking before, and that probably means more articles in your future.
Starting point is 00:14:15 Yeah, it would require me to listen to him. That's true. You can just catch the summary on Twitter, right? Whatever terrible thing he says will quickly rise to the top of your feet. Oh, yeah. So speaking of articles, we don't maybe have to talk about the one you wrote about A-Rod claiming that pitching to Contact is so much better than not. That turned out not to be quite true according to your research. Shocker. Shockingly. Yes.
Starting point is 00:14:42 But let's talk about some of the deeper dives that you've done in the past year on several subjects. And maybe we can talk with one of the more recent series you did because it required a very deep dive and a lot of data wrangling. And it is relevant right now because the conversations between MLB and the MLBPA are going on or might be about to go on again at some point soon. And those conversations do revolve around competitive balance to some extent and the idea of revenue sharing and what do small market teams need to compete. And you took a pretty comprehensive or as comprehensive as it could be look at our small market teams actually at a disadvantage. Or is Rob Manfred making too much of this?
Starting point is 00:15:30 Yeah, I would kind of lean towards the latter camp with that. What I did was, because you hear it all the time that if you're in a Milwaukee or a Cincinnati or Kansas City, there's no way that you have the financial resources to compete with New York, L.A., Chicago. And, you know, sort of the obvious fallacy of that notion is if you just look at the two central divisions, each of them have a team in Chicago, which is, you know, one of the largest cities in the country, then a bunch of small towns. And yet you don't see the Cubs and the White Sox dominating every year the way, well, you
Starting point is 00:16:10 don't see them dominating every year. So what I did was I looked at the census data for both the cities in which the teams play and also their metropolitan areas, because you can draw obviously from more than just your city limits. And I just correlated those numbers by year, and I smoothed them out between the decennial censuses to winning percentage. And what I found was that there is a positive correlation that market size does impact winning favorably, but that effect is so small, it's almost negligible. The correlation coefficient for cities was 0.10, and for MSAs, it's 0.16.
Starting point is 00:16:57 And usually, if you're doing any kind of correlation, you want something at least one quarter, really pushing a half before you can really claim there's any significance. So the idea that you have to be in a big city in order to be successful, it's illustrated by those numbers not to be true. And then if you just think about the way that revenue gets shared on national TV contracts, on a lot of the licensing stuff. And then even with, in some cases, the large markets versus small markets in terms of their local media revenues, it makes sense that if you're in Milwaukee, or if you're in Minnesota, or if you're in St. Louis, you can compete with a big city.
Starting point is 00:17:40 And, you know, the top team in the American League this year was one that beat out the enormous New York, Boston, and Toronto markets in the AL East. a couple of alternatives that might account for something like that, right? Because like Tampa is in a small market and we might say that doesn't matter, but maybe they're just so good at other things that aren't captured by this, that it overcomes those considerations, right? They're so good at player development. They're so good at drafting that they're able to overcome a disadvantage that is otherwise present. So was there accounting for that, that you tried to do to maybe just head off another comment on your piece? Yeah, no, I think you're absolutely right, Meg. And I think that that is in fact how small markets can compete. It's, you know, it's pretty tough to quantify that, but we had, you know, Tampa's
Starting point is 00:18:43 having a good run now. We had a really long run with Cleveland that, you know, they've decided to pull the plug on that. Kansas City had their moment in the sun. The Twins have been really successful for a while. And all of those did that with limited financial resources, where even in a league where, you know, there's no stupid front offices that don't embrace analytics or anything like that anymore, it does illustrate that there is a competitive advantage that can be gleaned from non-financial resources. So I think you nailed it. That's why the money is not the sole or market size is not the sole determinant of success. And does that suggest anything to you in terms
Starting point is 00:19:23 of potential tweaks to revenue sharing or to the competitive balance tax or just anything else that, you know, players and owners are, in theory, supposedly sitting down sometime soon at the table? Should this be more or less of a priority for the players, let's say, than they are indicating that it is. Well, I thought that Joe Sheehan's newsletter, which we talked about on the podcast before, he raised a really good point in that one of the weaknesses of revenue sharing can be easily viewed by the pirates in that if you're in a small market, you can just take all the money and bank it and turn a profit while putting out a pretty rotten product on the field. And there's no downside to that. So that would be, I guess, the counter argument to revenue sharing. But I think that revenue sharing probably is something that helps competitiveness in
Starting point is 00:20:20 the league. And rather than tweak it per se, there might be some sort of scaling based on team success over a few seasons. So a lot of teams go through booms or busts, but the bust for the Pirates has been particularly protracted. So maybe they wouldn't be eligible for as many dollars for that. If you lose all the time, you shouldn't make money, that kind of idea. for that. If you lose all the time, you shouldn't make money, that kind of idea. But in general, I think that one of the reasons for the lack of market size stratification is the revenue sharing. Yeah. It strikes me that if the ownership group's intent is to try to maximize profit with very little consideration to the product on the field, even beyond sort of the usual indifference or sort of satisfaction
Starting point is 00:21:06 with sitting around a potential wildcard number, but not pushing for more that you don't need to design your system to assume that everyone will operate that way. You need to design a system that disincentivizes that particular behavior because that impulse will manifest in other ways, depending on what the system is, right? You should just like go at those people directly and say like, pirates you gotta knock this stuff off or you're not getting any more money from us right yeah so one thing that we talked about earlier this week and we'll address
Starting point is 00:21:33 again later in this episode is the times through the order effect and some of the unmeasurable or hard to measure part of that is just what causes it is Is it familiarity or fatigue? It's also just hard to measure how big the actual effect is. And there are various ways you can do that. And you did a whole series early last year looking into that, looking into how big that effect actually is. And historically speaking, did it used to be bigger? Did it used to be smaller? Because this has come to the fore fairly recently in sabermetric circles. And so I think there's some question of, well, did this always exist or is it perhaps partly a product of just the way pitchers are used now? You know, they're not trained to go deep into games. And
Starting point is 00:22:14 so that's why there's a bigger drop off later in the game. And so you looked into earlier eras of baseball history to see how it compares. So I guess that's two questions. If you could explain briefly why it's so difficult actually to figure out the magnitude of that effect, and then also what did you find about the historic patterns? Yeah, well, this is in no way to slight the data that we have available now on times through the order, because it's not an easy thing to calculate. If you look at baseball reference, you can see that the first time a starter faced the order in 2021, the opposing batters hit 708.
Starting point is 00:22:53 That was their OPS. Second time was 747. Third time it was 779. So clearly there's a penalty there. But there are two limitations to that sort of research. The first one is that you're taking every pitcher and throwing them into the appropriate 1 through 9, 10 through 18, and 19 through 27 bucket when you're doing that. So if you have a pitcher who faces six batters,
Starting point is 00:23:22 doesn't get any of them out, and gets lifted, he's going to be included in that first time through the order, which is going to pollute that first time through the order number pretty substantially. And it seems to me if you're going to look at third time through the order penalty, you want to only look at a data set of pitchers who actually made it long enough to last three times through the order. So that's one confounding factor. The second one, and this one I did not think of myself. This was a smart commenter on our site that brought this to my attention. If a pitcher gets pulled after facing four batters the third time through the order,
Starting point is 00:24:00 he in all likelihood had to face much tougher batters in that third time through the order, he in all likelihood had to face much tougher batters in that third time through the order, the top four of the lineup, than he did the previous two times where he got to face all nine of them, especially in the National League where one of those nine was likely a pitcher. And so that would tend to overstate the impact of the times through the order penalty just because you're going to face tougher batters if you don't make it all the way through a time through the order just because the way lineups are constructed. So what I did is I looked at only pitchers who made it three times through the order, and I looked at the analysis by batting order position, whether it's, you know, facing the
Starting point is 00:24:40 first guy through the order, the first two guys through the order, first three, etc. And, you know, on an individual pitcher basis, that's small numbers, but you start aggregating things, you get pretty large numbers. And what I found is that the numbers are, in fact, larger than what has been reported. If you look at just the 2012 to 2020 period, I haven't been able to do the numbers for 2021 yet. The pitchers who face the opposing lineup three times allowed an OPS 12 points higher the second time through the order, and then 65 points higher the third time through the order. Difference of 77 points of OPS, and that's pretty significant. And that's larger than what you get if you look at
Starting point is 00:25:26 the numbers that are on baseball reference to other sites. Right. And that's why everyone always says, always read the comments, right? Yeah. Excellent advice. There's some smart stuff in there. Yeah. I've become a better researcher because of our comments section. You're absolutely right. Yeah. Comments section is a subscription site, a bit better than at some other sites that I won't name. Yeah, we do have that going for us. So after I saw that, I was thinking, well, exactly what you're saying.
Starting point is 00:25:52 Why is the effect so large? Why have pitchers become so much worse the third time through the order? How much better were things back when pitchers faced the order typically not just two times, but three and often four times? And so I compared the data that I had to the 1969 to 1976 period, the first few years of the expansion era. And what I found was that the times through the order of penalty for the third time specifically was actually larger then than it is now. It was 77 points, as I said, in the last few years, and it was 98 points higher back when, you know,
Starting point is 00:26:36 Wilbur Wood and Mickey Lulich and all these other guys were throwing 300, 350 innings a year. So basically what I found is that the times through the order penalty has always been there. Pitchers have always done worse subsequent times through, you know, that they face a batter. And, you know, Cameron Grove's comments a couple episodes ago, I thought really echoed this when he said it's a matter of familiarity. It's not because these guys are trained to only go five innings and then they say bring in a reliever, it's because the batters just are really good at what they do. And they were 50, 60 years ago and they still are today. And so I think the main reason why we're seeing pitchers not facing the order a third time isn't because they're
Starting point is 00:27:19 ineffective the third time. They've always been effective the third time. It's because if you're managing a team, you have much better options now than you did in the 1970s because of the emergence of the eight-man bullpen, all of whom throw 90 to 105. In that respect, does it surprise you that we even see pitchers facing the order a third time through as often as we do now? I mean, we obviously see it so much less often than we used to. But when you have dramatic effects like that, sometimes, you know, you figure that a team knows exactly what we know, they might know it to an even greater degree because of the data they have access to. And it's like, why do we see them even as much as we do, given what we know? I think it's a combination of three things. First is that you do have to manage your
Starting point is 00:28:07 bullpen to some extent. Oh, there's that. Yeah, if you're going to pull everyone after five innings, you're going to get, which, you know, it's roughly about where you're going to get three times through the order, you're going to have some tired relievers after a while. The second is that there is a subset of starters who are, even the third time through, going to be better than the guy that you're going to bring in to pitch the sixth inning or the seventh inning. You know, if you got DeGrom on the mound, ride him. You know, if you got Robbie Ray and he's dealing, ride him. And the third reason, I think it's a little bit of a selection bias issue in that if you've got a pitcher who's facing the order a third time, it's a little circular because he didn't get pulled the first two times. He's
Starting point is 00:28:52 having a pretty good game. And so, you know, there's a, especially not so much in the post season, but especially in the regular season, I think there's a tendency that his pitch count isn't too high. Just ride him, you know, see keep it up. Every pitcher's dealing until he's not. Maybe you can get him to deal for another time through the order. So I think it's those three factors. And that's maybe a good segue into some other research you did recently on the decreasing gap between starters and relievers in terms of performance, which makes sense, of course, because starters are pitching less and relievers are pitching more. And so you would expect their stats to come closer into line. But as you found, it is not necessarily that all relievers are
Starting point is 00:29:34 suffering some decrease. It is maybe a certain subset of relievers. Yeah. And I'll tell you, the idea of doing this was an Effectively Wild episode, I think it was a couple years ago, when you were talking about how it was early in the season, but the reliever ERA was higher than the starter ERA, and that had not happened since, well, it was a long time. It had been many years. And in fact, the gap between starter ERA andver era and everything else fit whatever has been narrowing and what i did to research that was i went on fangraphs and fangraphs very helpfully has a stat called game leverage index that tells you the leverage index of a pitcher when he enters a game and so what I did is I put all relievers
Starting point is 00:30:25 into three buckets. I said high leverage relievers are ones who enter game when the enter game with the leverage is over 1.33, 1.00 is average. Low leverage are when it's below 0.67 and everybody else middle leverages in between those two. And what I found was that if you look at the performance of the pitchers in those buckets relative to starters, it really hasn't changed much. The high leverage guys do really well compared to starters. The medium leverage guys are kind of out of power. The low leverage guys are not as good. But then if you look at the innings, what I did is I looked at the 30-team era from 1998 to present. The high leverage pitchers are pitching in 2021, pitched 5% fewer innings than they did
Starting point is 00:31:15 in 1998. The medium leverage pitchers pitched 29% more, and the low leverage pitchers pitched 163% more, you know, two and a half times as many innings. And you think about it, okay, if you're saying the average starter is going to go from seven innings to five innings, somebody's got to pitch the sixth and seventh innings, and that is not going to be the best guy in your bullpen. So the gap is not that the good relievers are getting worse by any means. It's just that you're going to throw more relievers in there. Some are not as good as the other ones. But overall, getting back to your question, Meg, that pitcher who might not be doing well the third time through the order is being replaced by a reliever who maybe is just a
Starting point is 00:32:00 little bit better. It seems perfectly designed to anonymize the fan experience of pitchers right because we lose the anchor of the starter as sort of the narrative thrust of a game and then you're not seeing the best and highest leverage relievers throwing more from just to manage their workload and make sure that they're available as they need to be. There's obviously a lot of good thought to that strategy from a game deployment perspective. But those, you know, those guys who sop up the in-between, you know, that bit of bread you're using to get the sauce at the end. And it's just it seems perfectly geared to the fan not being able to really latch on to any of those guys because they're not the guy who's
Starting point is 00:32:46 out there for five or six innings and they're not the guy who's gonna you know smack his glove at the end of the game because ha my favorite team won so it seems like it could be kind of dicey from a labor perspective well and in fact i think that's a big reason why pitcher compensation has been going down another subject to research recently. You have closers who make really good money, but in between you've got this fungible group of guys who spend the year going between AAA and the majors making the minimum. And yeah, the guy, they're not with the team a whole year.
Starting point is 00:33:19 So it's tough for fans to get any familiarity with them anyway. If what they do is they work them hard for a week, and then they send them down to AAA so they can rehab down there and bring up some new arm from AAA. So there's no fan identification. They're getting paid the minimum, and they're not accruing a full year of service time. So you're absolutely right. It's kind of a perfect storm from a labor perspective. So injuries, that's one area where a lot of further research is required. It's difficult to do a lot of that research with public information, but you have written a lot about injuries this year, either solo or in tandem with Derek Rhodes, who's done some great injury
Starting point is 00:33:54 visualizations at BP this year. Did you learn anything about injuries other than that there are a lot of them? So far, I've learned that, unsurprisingly, pitchers get hurt more than batters do. And I think somewhat surprisingly, there are some large differences between teams that can't necessarily be explained by factors like average age, for instance. Cleveland had very few injuries this year, but Cleveland also has nobody who's over the age of 30, practically. They're the youngest team in the majors. The A's didn't have a whole lot of injuries, and they had one of the oldest teams in the majors. So something's going on. What I think to learn more about with that, we've probably got to look at more years of data. And Derek and I are working on that.
Starting point is 00:34:49 The other things that I want to look at for injuries is what's going on specifically with pitcher injuries, shoulders and elbows. Are there some teams that have cracked the code there? And then just look at some of the more common injuries. There's a lot of muscle pulls. And you wonder whether that's because some guys are just prone to that or some players that is are prone to that. Or maybe some teams are just doing something, I don't know, some of the pregame stretching or something, I don't know, that reduces the number of muscle pulls.
Starting point is 00:35:20 Just some common injuries I'm interested in doing more about. But yeah, I'd say the biggest takeaway that I got from this analysis is that at least in 2021, Oakland's staff did really well at keeping those old guys on the field. What are your expectations in terms of the direction of the trend line for next year? Do you think that a more normal... Well, I guess we don't know that we'll have a more normal ramp up, do we? No, we don't. But assuming that we start relatively on time, is your expectation that we might see injury rates return to something that approaches normal or that, you know, it seems as if these guys having gotten hurt, aren't going to be less inclined to get hurt in the future. So I wonder if they'll just be operating at a discount going forward.
Starting point is 00:36:08 Yeah, 2020 was so weird that you got to figure that some of what we saw in 2021 was spillover from 2020. And by the way, Neil said, when I looked at injuries, I'm talking about non-COVID IL placements. On the other hand, I think there is more of a inclination on the part of teams that if a guy just has, if their ankle's a little sore, better to put him on the 10-day rather than have him sit on the bench for four days in hopes that maybe by the fifth day he'll be better. Just because you do have somebody who does not have a sore ankle in AAA that you can
Starting point is 00:36:46 plug in for the 10 days that he'd be gone. And since it's 10 days, you're probably going to miss only nine games. So I would expect, I wouldn't be surprised if injury rates were down a little bit from 2021, but I don't know that 2022 is going to be as relatively injury-free as 2019 was, just because I think that there's, with the presence of a 10-day IL, that's not really that much time to make sure a guy's at 100%. One more area of interest I wanted to ask you about is the DH. You just did a multi-part series on the DH, and we've had the DH for 50 years in one league at this point. You wouldn't think there'd be that much left to learn, but there is, and you took a look at a few different things.
Starting point is 00:37:29 One was something we bantered about recently, how much the DH plays into the AL's recent advantage in interleague play, and then you also looked at how the universal DH functioned in 2020 and how you would expect it to function in 2022 if there is a universal DH again, which we still don't know. But what did you discover about the DH, those multiple lines of inquiry? Well, yeah, as you discussed on the podcast, the American League's three-year below 500 results in interleague play came to an end. They crashed to an end in 2021. They've won 56% of interleague games.
Starting point is 00:38:10 And so the question is, is it because they get the DH and NL teams don't really have a DH, and then when they go to an NL park, their pitchers are horrible hitters, but so are the National League ones, and losing the DH isn't that much of a problem unless you've got a guy, unless you have Nelson Cruz, a guy who absolutely can't play in the field. And so what I did to test that is I looked at the home versus road winning percentage in interleague play, and it is a little bit higher than, historically, than it's been
Starting point is 00:38:43 in other games. You know, the 54-46 breakdown works pretty well, both for non-interleague games and interleague games. But what's interesting is that in the last few years, the home advantage has been less than the average. It's been diminished somewhat. And to me, that sort of speaks to the idea that whatever advantage the AL has gotten from the DH, it's diminished over time. And in fact, if you look at the performance in interleague games of NLDHs versus ALDHs, NLDHs in most cases don't do as well, but they're getting pretty close. The OPS is within about 90% of what the ALDHs have gotten. And over the long run, it hasn't been that good. So I think it's a combination of just general familiarity, more comfort with the DH on National League rosters.
Starting point is 00:39:36 Maybe even to some degree, you've got a guy on your team who is well-suited for that position. And it's not really an excuse, I think, anymore that the NL can use for what's going on in interleague play. Yeah, I'm going to be very curious to see once if we get a universal DH in this next CBA, which I think is the expectation, how we see teams in both leagues kind of go about staffing that position. Because not only are there not a lot of players like Cruz who are truly unplayable in the field, but even the AL side like the number of teams that have a dedicated thumper and use that spot in that way rather than rotating guys who they're trying to give a day off to through seems to be diminishing over time and so I wonder with you know 30 clubs having to fill that spot
Starting point is 00:40:21 sort of what the approach is going to be to populating it. Do they look at it as an opportunity to really have one guy who occupies that role pretty consistently throughout the course of the season, or do they go on a rotation? I just don't know what the answer to that is going to be. Yeah, 2020 was an interesting natural experiment for that just because it was basically sprung on the NL clubs. They found out a few weeks before the season, so there's no time to adjust the roster, nor would they want to because they were going to, in all likelihood, and in fact, they did lose it in 2021. And so if you had a guy hanging out on your roster, if you had like Jesse Winker, a guy who's better suited to be a DH, both because he's not a great
Starting point is 00:41:00 fielder and he tends to hurt himself when he plays the field, then he can do really well. But a lot of teams had a mix and match. And in fact, the two teams that did just about the best were the Dodgers and the Padres in 2020. And that's because they had ridiculously deep offensive rosters. And you can't necessarily count on that. But I think that you're right that there's a limited number of full-time DHs. I think that if we do get a universal DH, we'll see more of them,
Starting point is 00:41:27 just because there are guys playing the outfield in National League parks that probably shouldn't be, but who can still hit pretty well. But yeah, I think you're right. The idea of everybody trying to get a guy who you can write in the lineup 140 games a year, your DH is going to be unattainable. There's not a bunch of Otanis lying around. Do you two think we're going to get a universal DH in 2021? Because I am very skeptical. I have thought that all along. I guess the closer we get to a potential season start without that seeming to come up.
Starting point is 00:42:07 I mean, there was a report that maybe they were discussing some side issues like that before getting to the core economic issues that they can't agree on. And so you would have thought that maybe we would have heard something about movement on the DH, and we haven't seemingly. And it is connected, obviously, to the economic issues. So it's not as if it's an entirely separate subject. I mean, the longer we go without hearing, yeah, the lower my percentage estimate for the probability would be. I guess I still think more likely than not, but not with a ton of confidence. It strikes me as an area where the sort of bright line between the desires of ownership and front office personnel is perhaps particularly stark, right?
Starting point is 00:42:49 Where I think that if you asked most people who are working in a front office, should the National League have the DH, like they would say, yes, please, yesterday. Why are we still doing this? This is ridiculous. but because it is likely to be one of the bargaining chips that gets exchanged for you know bigger issues i kind of agree with ben like the longer we go and the more contentious the debate becomes i wonder we'll end up having it although i guess that you know you just have to fill you have to sign the free agents that still are sitting out there one way or the other right we're gonna have a rush of transaction activity once the CBA is resolved regardless.
Starting point is 00:43:27 So I don't know, but it seems like a thing that players want and that people who work for teams want and that people who own teams are like, well, here's another card that we can play. So I don't know what that means for its future. Eventually, we will have it, right? Like teams hate that they don't have the DH and the NL. Like NL teams don't like this.
Starting point is 00:43:48 So you would think that it eventually has resolution, but yeah. Yeah, I've heard two theories. One is that, well, I've heard one theory, then I have my own. The one I've heard is that if we're still sitting here a month from now and nothing, we don't have a CBA, that if we get it, they might implement it for 2023, just because it's, you don't have enough time for NL teams to build a roster for it. If it's February and we're still twiddling our thumbs. And my own personal view is that if Tony Clark, if they're going out to subway and Tony Clark says he likes turkey, Rob Manfred would want to know what he's willing to give up in order to get turkey.
Starting point is 00:44:29 You know, everything's transactional. And when Manfred made his initial statement where he said that they offered a bunch of things to the players, including a universal DH, it sounds to me like he's expecting something back for that. And the players want a universal DH, but not as much as they want a lot of other stuff. So I think that there could be, you know, okay, dude, if you think this is an ask, forget about it. We'll talk about it next CBA. So I don't know. Yeah. In addition to the other reasons why I want a universal DH, I think it would be appropriate if Zach Greinke had the last hit of the pitcher hitting era. So that's another reason to call it here, but probably doesn't carry
Starting point is 00:45:11 a lot of weight at the bargaining table. So the last thing I wanted to ask you about, and feel free to bring up anything else because you've been a busy man and don't want to give short shrift to any other research, but mobility versus equality. and this Is something that maybe I could have brought up earlier When we were talking about competitive Balance and small markets Etc but it seems like A lot of fans still seem to think that Baseball doesn't have good competitive balance and
Starting point is 00:45:35 We've run down the numbers relative to other Sports you know the number of teams that win titles Or make the playoffs or however you Want to classify these things baseball seems to Stack up quite well so doesn't Ne't necessarily need a salary cap for that reason. But people still seem to think that there's a problem there. And you have come up with this way to measure the mobility and equality of competitiveness. And I wonder if that is what people are picking up on and interpreting as a problem. So what's the theoretical framework there and what have you learned? Yeah. If you think of equality kind of like income inequality, and actually I used a measure of income inequality called the Gini index, but I applied it to one loss records instead of household income. If you look at any most recent seasons, there's a lot of
Starting point is 00:46:26 inequality. Some teams are really good. Some teams are really bad. There's not an enormous middle class. And so there's a wide berth between the best teams and the worst teams. In the divisional play area in 2021, the National League works out having the fourth highest inequality among all the leagues since 1969. So yeah, if you look at, take a snapshot of any season, there's a big gap between the best and the worst teams. However, that's not the same thing as mobility. There are some teams that are rotten, but if you're a fan of, you know, the Tigers a couple years ago, and you can see that, you know, they're building a pretty good farm system, you know, this year, they're starting to spend on some free agents, you can see things getting better. And so you can,
Starting point is 00:47:16 if you're a fan, you can deal with some bad years if you've got the promise of some good years ahead. So I also looked at mobility, which is the ability of teams to go up in the standings, and by the same token, the ability of some teams to go down the standings. I measured that by looking at five to seven year stretches for teams, what the average standard deviation was of their ordinal finish. That is first, second, third, fourth, fifth, etc. And what I found was, and anyone who's looked at baseball history wouldn't be surprised by this, the seasons with the least mobility were the 1950s and early 60s, when the Yankees were winning every year, when the Senators and the Phillies and most years the
Starting point is 00:47:56 Pirates and the Browns before they moved to Baltimore were always awful. You could pretty much take last year's finish, run them through a Xerox machine or whatever they had back then, and they'd look about the same. There was very little mobility among teams. What we've got now is actually, relatively speaking, pretty high mobility over both five-year and seven-year spans. If you look back to the early 1900s, we're in the roughly, I don't know, top 15, 20, 25 percent for mobility. And so it can seem that the Dodgers and Yankees always win and the Pirates and Orioles always lose.
Starting point is 00:48:40 But if you think about it, it wasn't that long ago that the Orioles were always good. The Pirates had a pretty good run for a while. You know, teams like the Cubs have gone from being awful to really good to not so good again. The Giants have ruined everybody's prediction systems this year. And the amount of mobility that we see in contemporary baseball is actually relative to the long arc of baseball history. It's pretty good.
Starting point is 00:49:04 So yeah, you can be watching a team that's really bad this year, but their chance of being good in a few years is actually higher than it's been in most of baseball history. That is good news. Especially for the Mariners. Well, they've never, well, I guess they have had some years being that bad, but they're good again. Good example. Yeah, they don't make the playoffs. They're not always terrible. but that's the silver lining. So thank you, Rob. It's always a pleasure
Starting point is 00:49:31 to read you and to talk to you as well. So thanks for all the valuable research that you have done and continue to do. This was a pleasure. Always good talking to both of you. Thanks for having me. All right. So let's take a quick break here and we'll be back in just a moment with the pseudonymous Thanks for having me. I'm not asking your patience I'm out of it myself And everything is surreal in the evening All the names are down All the names are down The communication of the breath All right, we are back and we are joined now by the author of a site called Harib's Hangout.
Starting point is 00:50:30 He goes by Harib on the site and will also on this podcast, although that's not his actual name. He is part of the tradition of sacramentric pseudonyms. But one way or another, welcome to the podcast. Hello. Hi, thanks for having me on. So you've written a lot of subjects on your site. I think your earliest post and your most recent post were both about Rocket League. You've written about Magic the Gathering as well. So what is your baseball background, if you don't mind saying? Were you a fan, a follower of the sport for a long time, or did you just come to it via the analysis?
Starting point is 00:51:04 And do you find baseball analysis more or less rewarding than other types of analysis? I started off, my father was a fan. I believe he'd pitched in high school for a couple of years and the Chattanooga Lookouts minor league team, we saw a lot of games there. Some people from his work would often go down there and he'd take me along and so we'd probably hit maybe 30 or 40 games a season and then after that i watched the braves and it was a good time to be a braves fan in the 90s yes early in the early 2000s so i watched a fair bit of that book and that was fun and then i sort of lapsed from that and most of sports. And then I actually got back into it in the late 2000s, actually came through to it through
Starting point is 00:51:54 sports betting and the amount of data that was available for baseball was much better than for other sports. And I was a math major, so it was a natural place to start with and baseball analysis is a very good way to learn practical statistics because every bit of data you can look at is screwed open one way or another it's selection bias everywhere you look every possible way you can look yep and you can mess up in ways that you've never even considered possible. Yep. I've been there.
Starting point is 00:52:31 Yep. And it's fun and it's infuriating. And sometimes you can learn things. And sometimes you just get better at not making mistakes. I don't want to take us too far afield from your baseball research, not making mistakes. I don't want to take us too far afield from your baseball research, but I'm curious how your experience of doing baseball analysis relative to some of the other arenas, including Magic the Gatherings Arena, how it might differ from some of the analysis you're doing there. How sort of well formed are the communities around that kind of analysis?
Starting point is 00:53:03 And what are some of the differences in the ways that folks are sort of approaching their analysis and the kinds of questions they're interested in answering? When it comes to Magic, that's pretty different. What I was doing is the new digital play program they have, Magic the Gathering Arena. It has a rating system, and the rating and the pairing algorithm are completely opaque. They gave slight hints, most of which were either partially or not correct. And so it was a matter of I noticed some strange behavior, and I was just trying to reverse engineer the thing because I couldn't get any word on how it actually worked. And so I'd done a deep dive into that and as far as i know there there isn't really much of a community about trying to reverse engineer a
Starting point is 00:53:51 rating system but there is a lot of i mean there's been a long history for over 20 years of people trying to understand the theory of the game itself and there's some statistical work in that you need to play this many of certain types of cards in your deck to be able to reliably see these in the first 15 cards and things like that. Just basic statistics, hypergeometric distributions, things like that. But my work there was really pretty unrelated it was it isn't really as advanced it was it was something that anybody could do if they just wanted to sit down and do it but nobody did it so i did it and when it comes to rocket league rocket league itself itself is
Starting point is 00:54:39 extremely complicated because the there it's basically soccer three on three with flying cars trying to hit the ball into the net it's pretty and it's pretty fast and it's pretty fast paced it's a great sport to watch it's fun to play even if you're not very good at it like me but as far as like trying to do value-added analysis that's incredibly difficult. But what I was doing was looking at the viewership and the whole online viewership of esports. And this is a problem that goes back hundreds of years. Advertisers want to know what your circulation is. And the people who are trying to sell the ads are always trying to inflate how many people they actually reach. And so I was doing a bit of analysis and data gathering for a couple of years now, trying to figure out how much of the viewership is real people and how much of it is purchased through basically just buying ads that show an embedded stream to count as a viewer
Starting point is 00:55:47 and basically figuring out how much of that was real and how much of it was companies trying to scam people and pretending their viewership was too high. This is a slight sidetrack too, but you mentioned getting back into baseball analysis because of sports betting. Were you a successful baseball sports better and did that change over time um yeah especially a while back when in-play betting first started becoming a thing it was definitely a good time to be alive and i'd done a lot of work like basically trying to predict how the rest of the game would go and there was a lot of some basic modeling and then some studying and knowing how teams would behave like when
Starting point is 00:56:33 who had good relievers when they when they would use them when teams basic basically teams that had good lineups and if you could get the good part of the lineup against the bad part of the other team's lineup. And the people that I was, you know, wagering against weren't at that point the sharpest knives in the drawer. Like, cause those aren't the kind, those are the kinds of things that are, some of them are public knowledge and others aren't super complicated, but they just didn't realize that lots of those edges existed. And then over time, you can only beat somebody the same way for so long before they figure it out. Right. And so lots of things that I did over the years, you can just slowly see them adjusting to what I did, adjusting to things other people did.
Starting point is 00:57:27 And now a lot of the people who do the in-play sports betting are the people who used to beat it before. Now they're the ones setting the lines, which is a natural progression, but it also means that things are much more difficult than they used to be. Alas. So we brought you on to talk about a few different studies that you've done at your site, really about different areas of baseball, but all kind of falling under the umbrella of measuring difficult to measure things. And I guess we can start with one that's been the subject of some debate in the sabermetric community. We brought it up on our first segment today and in a previous episode this week, but that's the times through the order penalty. And there has been a debate about, is it fatigue? Is it familiarity? Is it some combination of both?
Starting point is 00:58:11 And a while back, you weighed in on this too. And I know that you can kind of get into the weeds on the math and everything, but if you can sort of sum up either the approach you took or the conclusion you came to, that would be interesting because I've kind of been on the side of the familiarity camp myself, more so than the fatigue. But there's been a lot of research done that has supported one side or the other. Yeah, I was glad to see Cameron's research earlier this week that it came out where he found it very strongly in terms of familiarity. And that's just the kind of thing from playing sports, from playing games, from watching other people play games, is people adjust to things. The more they see something, the better they react to it. The faster they react to it, the more accurately they react to it. And it seemed
Starting point is 00:59:00 pretty much impossible to me that familiarity wasn't a significant indicator. Although, obviously, it's somewhat intuitive, too, that a pitcher could get tired over the course of a game. You know, it's like you were saying earlier. Yeah, it was a question of was it both. In my mind, it was much more a question of was it both or was it just familiarity. Right. And as you were saying, this is one of those cases where there are all sorts of confounding factors and there are so many minefields you can walk into here in doing this sort of analysis that will lead you one way or the other. Yeah. I hadn't, in a quick look, I hadn't
Starting point is 00:59:35 found much evidence that, especially when it came to fastball velocity, that there was a huge drop off later in the game. And I believe Cameron had found the same thing. Because, I mean, obviously, if you make a guy go out there and throw 150 pitches, it's going to start getting real ugly real fast. But also, I think MGL had done a study on pitchers on the number of pitches that pitchers had in their repertoire. Yes. And the ones with one pitch, one major pitch, did much worse the second time up.
Starting point is 01:00:10 And still did badly the third time up. And pitchers with a much wider repertoire were able to have a lower penalty. And that dovetails with what Cameron found with this stuff, that it's not just seeing the pitcher, it's also seeing the individual pitches. And what I've looked at was I had the bright idea and this ended up not working, but I thought it was a good idea at the time that I would try to measure fatigue by the difference between the number of pitches that the pitcher had thrown between when the batter came up the first time and the second time and the difference between okay he thrown 20 pitches the first time up and
Starting point is 01:00:50 he thrown 60 pitches by the time we got to him the second time so there was going to be a 40 pitch fatigue difference and it seemed like measuring difference in fatigue was much better than trying to measure fatigue because the effect should be relatively, I don't want to say linear, but you'd expect it to be something approaching that if it exists. And what instead I found was that throwing a small number of pitches in between the first and second plate appearance basically means that you're significantly more likely to have gotten on base in the first plate appearance because especially a one pitch plate appearance follows a successful on base event much more often than it follows somebody just making an out and so that was a great case of
Starting point is 01:01:38 confounding that i found i'd found the effect i'd found the effect in the direction that I'd wanted it to that throwing a lot of pitches leads to a bigger penalty because throwing few pitches means you got on base the first time right so you're not going to have if your second plate appearance is normal and your first appearance is weighted toward getting on base it's going to look like you did better throwing fewer pitches but it turns out that effect that I found was completely not real. It was completely explained. Like, I mean, down to.001, I think, of WOBA, of simply the first plate appearance being good. So talking about measuring the unmeasurable.
Starting point is 01:02:23 Right. There we go. It was like a great little idea and then just completely confounded. But that was kind of interesting that I couldn't find, once I controlled for the thing that was obviously not true, that throwing fewer pitches in the future didn't go back and affect the first plate appearance, that I couldn't find a fatigue effect there at all. You mentioned that part of this was based on sort of the intuition you have had of playing sports. I think it's always interesting for new researchers as they're thinking about starting out and trying to get their hands around baseball data to understand not only the sort of methodologies they need to be conversing in, but also how more established
Starting point is 01:03:05 researchers go about formulating their research questions. How often do you find that sort of real world experience and intuition informing the way that you approach a problem or even the kinds of questions that you're asking? I think it's actually, that's an interesting question. I think it's actually extremely important to have some idea of how the world works relative to what you're looking at before you just start putting numbers in a model, because you can get total nonsense back out and not realize it. If I can go into a Rocket League example for a minute, because somebody posted something a couple of months ago. They'd analyzed a bunch of one-on-one matches, and there are lots of things
Starting point is 01:03:51 you can do in Rocket League. You can make your car fly a certain way, you can drive faster, you can turn faster, you can have better control. All of these things that, if you watch good players play, they're all infinitely better at than bad players. I mean, you'd have to watch the game for all of 30 seconds to realize that one guy is dribbling around and shooting like Steph Curry, and the other one's like some guy at your local gym. The class difference is just gigantic. But then this person ran analysis on, measured how much, how well everybody did with all of these mechanics and how that correlated to winning the matches and found that there was almost no correlation. And like actually believed that that was true. It wasn't just a case of did something wrong and realized, oops, I can't do that. Like actually believed and wrote a
Starting point is 01:04:46 paper that this was true and what had happened is the paper his sample was only people playing people of other similar skill and so what had happened was if we know going in that the that they're similar skill they may be similar skill in slightly different ways but they're only playing people of similar skill so the end package is going to be all the same and it doesn't matter how you get there you're going to be you're going to have a relatively even matchup and it's the same thing that baseball talked about a while back with velocity. When you had the mediocre low-end replacement level as pitchers, velocity didn't have much correlation to success. Because you'd have the guys with stuff and low velocity, and you'd have the guys with who threw harder but didn't have much control. And it would all come out in a wash.
Starting point is 01:05:46 Right. And I know there have been cases of, say, people running a regression to try to calculate linear weights, you know, what is each event in a baseball game worth, and you can end up with really strange results like, you know, a sacrifice fly is worth more than a double or something like that just because it's correlated with or associated with run scoring events right and and if you don't know the sport and you just know the regression output you might think that that's true whereas if you understand the rules of baseball well you know of course that's not quite true so yeah i had a book on basketball that where somebody had tried to not the author of the book he was talking about somebody else having done this we're coming out where an assist was worth five points right and you're like no that's not possible and this just can't be worth five points but yeah you can come up with the with similar things but if you realize
Starting point is 01:06:38 if you understand what you're looking at like that's a mistake you can never make right or at least it's it's not it's a mistake you can never make. Or at least it's a mistake you can never publish. You can make any mistake as possible. It's a mistake you'll realize is a mistake long before you actually act on it. Right. So a couple of related studies you did, I want to ask you about. They're both on the subject of valuing relievers. One of them is about reliever sequencing. One of them is about valuing closers. I want to start with reliever sequencing, which is a subject that has intrigued me for a while.
Starting point is 01:07:10 It's basically the idea of, does it matter if a certain type of pitcher comes in after another type of pitcher? Is there some adjustment period? Is there a hangover effect? You know, people have looked into, is there a hangover effect after a knuckleball pitcher leaves the game and hitters have to adjust to a conventional pitcher? Or it could be handedness,
Starting point is 01:07:29 or it could be arm slot, or it could be velocity. You hear this a lot, that you want to give hitters a lot of different looks, but it's a difficult thing to study. And you made an attempt and you seemed to find something. So can you explain how you looked into this? Right. It was the same motivation, and sort of related to the familiarity research, is that if you see something that's similar to what you just saw, are you going to have an advantage over something completely different? Right. And I thought the simplest thing to look at was handiness,
Starting point is 01:07:58 because you couldn't possibly have a bigger difference than throwing from the different side of the plate. And after controlling for platoon effects and pitcher and batter skill and all the things that you have to control for whenever you're doing baseball, I didn't find anything at all when it came to handedness. I didn't look at arm slot. I didn't, I could have, I probably should have, but I didn't have, I didn't have that data classified, and I could have gone through release points and come up with something there, but that wasn't something I actually looked at. that's sort of intuitive with playing sports is when you're used to something and then all of a sudden something comes a lot faster, there's an adjustment period. Yeah. I mean, that's the whole idea of pitching, right? Upset and timing, as they say. So it's the question of, does it carry over
Starting point is 01:08:57 not just from pitch to pitch, but from pitcher to pitcher? Yeah. And then likewise, if you're used to something fast and then something comes a little softer, it's sort of like easy mode. If it comes a lot softer and you trick somebody, that's called a change up and the batter looks like an idiot. And, but what I found was basically that throwing harder immediately, throwing hard immediately after a soft toss tossing reliever was a benefit to the pitcher and that throwing and that assault and bringing in a soft tosser after a fast starter was not good for the reliever. That their appearances were significantly worse after following a hard thrower than they were after following a medium thrower or another soft thrower. Right, and it was not a tiny effect. I mean, by the standards of these things where you're looking for small edges oneower or another soft thrower. Right. And it was not a tiny effect.
Starting point is 01:09:45 I mean, by the standards of these things where you're looking for small edges one way or another, it was sort of significant, it seemed like. Yeah. I've got to look up the exact value, but yeah, 0.2 to 0.3 runs per nine was the effect that was the size of the effect I got. I didn't want to say something totally off, but yeah, I mean, that's a, that's real. Like that's something that you can just, if you have the choice of who you're planning to bring in, you can just not do that.
Starting point is 01:10:15 And like, okay. I mean, like, I mean, sometimes you don't have a choice. Sometimes your bullpen is stuck, but it's like, okay, I'm not going to bring this guy in after that guy, or I'm definitely want to bring the hard thrower in after this one. Like it's the kind of thing you can plan for in advance and pick up a couple of runs over the course of a season. Well, Ben mentioned that you decided to wade into the Hall of Fame debate. You wrote this in 2020, but it'll be perennially relevant, I imagine. And you found yourself dissatisfied with the way that relievers were being treated in Hall of Fame voting. I know from editing Jay Jaffe that this is an area where he has found Jaws to be kind of wanting over the years, and he is,
Starting point is 01:10:57 you know, experimenting with some different ways to try to more precisely account for and calibrate closer and reliever more generally contributions as we're evaluating Hall of Fame cases. So what methodology did you land on here and what sort of results did it yield in terms of the guys who might be worthy of Cooperstown? The first thing I want to say is that like when it comes to valuation by wins above replacement, I think the stats mostly do a good job. Like Mariano Rivera is head and shoulders above the rest and pitched for so long and so much in the postseason, still pitching incredibly well that even if you're just looking at war, that he's an easy inclusion. But once you get beyond that, then Tango is kind of the opinion of, if these pitchers were any good, they'd just be starters,
Starting point is 01:11:45 and that mediocre, not mediocre, but good, but not Hall of Fame level starters are just more valuable and are just better players all around. And I think that's half true, and in most cases it's completely true, but I looked at this a couple of ways and the first way was I compared just to see who the actual good closers were was calculating basically how good closers were in the years in each year which accounted for differing usage because obviously they pitched more than one inning significantly more often 30 years ago than they do today or 40 years ago and use the, use the average closer as a baseline and then determined how many wins above
Starting point is 01:12:39 that base. And then it's like setting that as a replacement level and determined how many wins above that level that the closers were worth. And the list for that was like second place was Billy Wagner. Third place was Joe Nathan. Fourth place was Zach Britton. Fifth place was Kimbrell. And that's in career.
Starting point is 01:13:02 And this data, I didn't have full data for Hoyt Wilhelm in this. Right. In what I did. But looking back, Hoyt Wilhelm was really good. He was in the second tier level below Mariano. And then when you looked at peak level, then you had Goose Gossage. And he was before my time but if you go back and look at his 10-year peak from 75 to 85 he spent one year starting and put up like two wins but his relieving years
Starting point is 01:13:34 around that were Mariano level I think I wrote that he was Mariano for 10 years and that was the best imitation we've ever seen. And the question, like, now, when you look at peak like that, you had Eckersley as a reliever was really good. Wagner was still good. Gossage was good. Joe Nathan looks better as a peak because he had a couple of mediocre years around his career and then Gossage's career value got dragged down because he pitched for something like eight years after
Starting point is 01:14:12 his peak in the last six years I want to say he was worth just under replacement level so if you just look at his career stats he's not as good but if you look at him pitching at the normal age that people pitch at he was very impressive and so i think when you're comparing based on that then you have like some of the older inductions like raleigh fingers his career was 0.7 wins over the average closer and 1.9 wins in his like nine year peak you're like okay he was basically an above average slightly above average closer for 10 years okay that's not very exciting to me and you had bruce suiter who was zero uh wins above the average closer for his career and 4.3 in his peak that's not as exciting to me lee smith 2.5 in his career and 4.5 wins in his peak. You have Mariano was almost 17 wins.
Starting point is 01:15:11 Wagner was a little over seven wins. And you see that there's kind of a big, there's a big drop off between the second tier guys and some of the guys who'd already been in the Hall of Fame. And by this measure, which uses, it's just using runs and not WPA. Trevorffman was 3.8 wins over his career and 5.8 wins in his peak so to me like if you look at wpa hoffman looks reasonable if you look at runs he's really not super impressive and the other thing with hoffman which is a bit of a digression, is I was
Starting point is 01:15:46 never fully convinced that he wasn't getting a benefit from pitching very late in the evening in San Diego. Because listening to announcers and watching those games, the balls just kind of start dying out there when the Marine layer hits. And even more than just in the average san diego game yeah i thought he i thought like on top of everything else he may have had some benefit from that more than the average closer just you know pitching at night when it's cooler than when the game starts yeah i think one of the interesting takeaways from this study which kind of matched people's perceptions is just that a lot of closers are pretty interchangeable, which I suppose you could say about players at any position, but it's hard to
Starting point is 01:16:31 be a good closer for a long period. I mean, like generally you have to be decent to be a closer at all. Not always, obviously. Shout out to Joe Borowski and Sean Chacon or whoever. Carlos Marmol. Shout out to Joe Borowski and Sean Chacon or whoever. But there isn't a lot of track record of guys who were better than the average closer for a really long time. Like you have Mariano with 16 seasons where he had positive wins above average closer. No one else is even close to him. Yeah, it follows like Wagner at 12.
Starting point is 01:17:05 And I think Hoyt Wilhelm I looked looked at and i think it's 13 i won't swear to that because i'd have to go back and run the numbers again because his data was before fangraphs didn't have all the data that i was using for him but yeah it's very that's the other thing that when comparing a closer who's trying to make a Hall of Fame case with a starter is a closer actually has to be good for a long time and pitch generally at least into his late thirties. Well, to have a case and most starters don't do that. I mean,
Starting point is 01:17:35 you have the Randy, Randy Johnson, you have the Scherzer, but for the most part, like that's, you follow more of a Kershaw trajectory where,aw trajectory where you're really good and then you're okay. And then you're not considered one of the best in the game by the time you're getting near normal retirement age. And I think it's kind of overlooked in the same way that catchers don't get as many plate appearances just because of the wear and tear of their position so if you just look at how you just
Starting point is 01:18:08 look at their batting stats like they're always going to be lower and closers have the same issue of it's just you're physically having to do something that other players don't have to do to put up hall of fame numbers and i it's hard to put that into an exact number, but it's at least a thing to consider. There was another line of research that I'd taken that I didn't write up on here. It was basically along the lines of if closers are just bad starters
Starting point is 01:18:38 that found a niche that they're pretty good at, which I think is like, the question is a hypothetical is if you turn good at which i think is like it the question is a hypothetical is if you turn good starters into closers how would they do and tango had the rule of 17 for converting relievers to starters and so you can sort of look at starters and and see how they do and and basically if you just have a reasonable number one starter like maybe the 15th best pitcher in the league and you convert him to a reliever and he's putting up numbers on par or better than your reliever is like the reliever is just not special and that's true for
Starting point is 01:19:18 almost all of them there are very few who over a career can beat out what you would expect a good, but not necessarily elite starter to do. Now, if you convert prime Pedro Martinez or Roger Clemens or something to a reliever, that's not a fair benchmark because those are the best players ever. So if you have an inner circle hall of famer or the best pitcher in baseball at any given time, like when I was looking at it, Kershaw had put up a monster run of great seasons. It's like, okay, this guy's not as good as Kershaw, but you don't have to be as good as Kershaw
Starting point is 01:19:53 to wind up in the Hall of Fame. Right. And what I'd found was that there were, at that point, four to five relievers who had really met that standard. And Nathan, I think when I looked at it before, Nathan had just fallen short on innings. Like his peak was really good, but it wasn't a super high number of innings.
Starting point is 01:20:13 And the four that did that were Mariano Rivera, obviously, Hoyt Wilhelm, Goose Gossage, and Billy Wagner by run prevention. And those were the only four pitchers that were significantly clear of the field of good, but not Hall of Fame level starting pitchers being converted to relievers, what they would hypothetically have done. And I thought that, I thought that that followed up, that matched pretty well with the intuition or not intuition with the opinion of people that
Starting point is 01:20:47 starters were generally just better and more valuable than closers good starters but it also matched up with mine is that some of these guys are actually doing something that's truly special not many of them but those four i thought definitely were simply in a... Mariano was in a class of his own, and then Widener Gossage and Wilhelm were in a tier two to me that I thought deserved some level of recognition. Yeah. Because they're actually doing something that almost nobody else could have done. Right. It does make sense, I suppose, that there would be a lot of closers who just washed
Starting point is 01:21:21 out of starting because they didn't have the repertoire or whatever, but also that there would be a lot of closers who just washed out of starting because they didn't have the repertoire or whatever, but also that there might be some guys who are just so well-suited to pitching in short bursts like that, whether it's the quote-unquote closer mentality of thriving in those situations, or whether it's just that they have an incredibly nasty unhittable pitch, like Rivera's cutter, for instance. Yeah, Rivera's cutter. I mean, now, like, it's interesting. Like, he was almost the only person throwing that with any frequency and now you see cutters everywhere like you have Kenley Jansen was throwing a cutter you have starters throwing cutters it's not it's not even a special pitch anymore but he definitely had that advantage on top of like absolutely ridiculous
Starting point is 01:21:59 pitch placement yeah where most people have their pitch their pitch placement graph is you know centered somewhere near the middle of the plate no i mean not middle middle but somewhere over the plate and then it's just a smooth distribution away from that and rivera has two rivera has two peaks the inside corner and the outside corner right and a big drop off in the middle of the plate and that's just kind of incredible that certainly some of the reason why he was so good is that his insane level of command like that. of the StatCast people, Tontango or Mike Petriell, have maybe touched on in the past that all opportunities are not created equal, even given where a ball is located. You know, you might have some slice, you might have some fade, and certain outfielders might face those more often than others, depending on what position they're playing or where they're standing. But you wrote that
Starting point is 01:23:02 everything else about the opportunities being equal, corner outfielders have a harder time catching pulled balls than they do catching balls hit to the opposite field. And I think maybe there are multiple potential explanations for the effect that you found, but I thought yours was interesting. So tell us about this one. Tell us about this one. Yeah, that was something. It was actually part of a bigger project that I'll touch on or what I was trying to measure until I realized that I just wasn't going to be able to with the data that I had. But in general, the StatCast model is how far does the outfielder have to go to get to the ball, basically to beat it to its landing point and how long is the ball in the air and it has an adjustment for whether you have to turn around and run backwards or the balls against the wall because those are obviously harder than just a ball that's a little bit off in front of you or a little bit off to the side
Starting point is 01:23:58 and that seems like it that seems like a model that's pretty much tracks with how people catch balls. I mean, you see where it's going, and you try to beat it there. Then I found, like, somewhat by, I wasn't really trying to find it, but I had by accident that pulled balls were significantly, like, for a competitive, like competitive like obviously a 99 ball is going to be a 99 ball whichever direction it's hit but for competitive plays like in the 20 to 80 percent range i think i'd found that pulled balls were caught approximately 10 percent less than sliced sliced balls and for even given the same hang time or correcting for hang time and distance to the ball and it wasn't clear to me i'm still not like really sure what's going on
Starting point is 01:24:58 there my guess is it has two factors and that there's one is visibility and that outfielder can see the entire stance. Like the right fielder can see the entire stance of a left-handed batter and see that the pitch is going outside and see that the batter is starting to reach for the ball. Because, I mean, the swing and the contact point are just different for a ball that's going to go oppo than a ball that's going to get pulled and can see that and get a better jump than a left fielder trying to look at a right-handed batter and see exactly what's you don't see what's happening until later in the swing and I think I'd found that the right-handed batters pulling the ball to the right side, the left field line side of the left fielder, and the opposite were much harder to catch.
Starting point is 01:25:52 Then a right-handed batter slicing a ball down the left field line. And it seemed, I think visibility and the quality of the jump is a significant factor there. And the quality of the jump is a significant factor there. And also, it may just be that just the arc of the ball getting down, the topspin on pulled balls, especially on heavily pulled balls, getting down faster may throw people off or make outfielders a bit less aggressive. But in the week, you'd had the series on like, what data do you wish people had? And I would absolutely love to have play-by-play jump data, how fast the initial step was,
Starting point is 01:26:33 the jump, the route, and all of that. Because now what we have reported is the seasonal average jump and the seasonal average sprint speed. We have no idea how quickly, or well, I have no idea. Teams may have the data, but I don't believe it's accessible to the public unless that changed very recently on a play-by-play level of how fast the reaction time was and how good the jump was. And I think if we had that, that effect would be much easier to explain with a lot more confidence. Well, you anticipated my closing question, but I'm going to put another one to you. First, I'd love to know, and this is, I guess, these are related questions, sort of what other data, apart from what you just mentioned for this
Starting point is 01:27:15 study, you're kind of keen to get your hands on, and what some of the questions you're going to try to answer are that are coming up next? What's next for you on the research front? you're going to try to answer are that are coming up next. What's next for you on the research front? As far as what data, one thing, there's been a lot of talk about the ball. And I actually like bought and stored a couple of balls in different humidity environments, trying to figure out things that could happen after basically if short-term storage could be affecting ball performance. And it turns out that the post-season we were looking at just was using old balls, which is kind of not the most interesting explanation, but it's certainly a good one that they just started using balls in the previous season
Starting point is 01:27:56 and the postseason and they behaved differently because it was an entirely different set of balls. But one thing I believe in 2018, But one thing, I believe in 2018, they attempted to mandate storage conditions across all ballparks where they were supposed to be kept in a reasonably confined temperature and humidity environment. But then with the kerfuffle last year about possibly sending two balls and sending balls out right before home stands that I thought was interesting was balls don't come to equilibrium very quickly. It can take two weeks approximately for a ball to equilibrate with its surroundings. And if balls are being put in play or if balls are being used in games and they haven't been stored on site in a controlled environment for very long, then you can get all kinds of effects. Like you could have balls that are heavy.
Starting point is 01:28:52 You could have balls that are light, balls that are dried out and bouncier. And even though they're in that room for a day to a couple of days, that's not nearly enough time. And I would, I wish I knew the supply chain and how all this worked because that would rule in and rule out various explanations and obviously i wish i just had you know we had better ball data because i mean it's just all the analysis of drag coefficients all of this or they're assuming that all the balls are the same size and all the balls are the same weight and they're not they're just not like they're just straight up manufactured differently i mean the differences
Starting point is 01:29:28 aren't huge but you can get a percent or two here and there and i'd love if we actually knew what was being put in play but that's probably fantasy land to expect that level of level of data when you're just hand a bag of balls to an umpire and he just picks one and throws one out but i wish we had that but that's just not i don't think that's realistic in the near future as far as research questions in the in the future that's i've been doing a little looking at with the pandemic of just doing a little modeling of how infections peak and decline. And I think this was a case of where some people had ideas that were significantly more realistic than a lot of the major epidemiological models. And the idea there, I don't want to go off on too much of a tangent, was that some people are very easy to infect. They're the people who go out and
Starting point is 01:30:26 socialize and the people who don't have a choice and they're working in crowded environments. Those people are always going to get nailed first. And then you have the people who are relatively hard to infect, like the ones who are distancing and not seeing many people. And a model that treats everybody the same is going to have a much longer duration of infection, while a model that realizes that there's a huge divergence will see the rapid rise and relatively quick fall that is what we keep seeing in practice. And I thought that was interesting. I looked at that. As far as the future, I mean, honestly, I don't know. I like finding problems. I like anything that catches my eye. Anything that's interesting,
Starting point is 01:31:13 I just have a tendency to pick up and look at. And if I can do anything with it, I try to. Some things, the data is just hopeless. But if it's interesting, I'll try to take a look at it. I don't have a specific topic for what's next, but I never have. And I always keep finding things. So I'm sure who knows what my next post on the site will be about. All right. Well, we look forward to whatever you come across and decide to pursue next. And we will link on our show page to all the research that we discussed today. But thanks very much for coming on and talking to us. Thanks for having me. All right. Thanks to our guests from today and this week. By the way, we are aware that Harib's nom de plume is a play on words, not in the best taste,
Starting point is 01:31:55 perhaps. He said when he started his accounts, he didn't expect them to be serious. Then they turned into actual research that people paid attention to. But he does prefer not to use his real name. And he had some interesting things to say. As did all of our guests this week, I hope you've enjoyed this series on measuring the unmeasurable. We may do more of this. And of course, this measuring the unmeasurable rubric that we're using here is not unlike what we do at other times. This is not the only time when we talk to authors of interesting research pieces. We talk all the time about defense or the shift or clubhouse chemistry or injuries or measuring the effect of shadows on the field. But hey, during an offseason lockout, you got to get content where you can.
Starting point is 01:32:33 And you can get content here at Effectively Wild because of our Effectively Wild Patreon supporters. become one of them by going to patreon.com slash effectively wild and signing up to pledge some monthly or yearly amount to help keep the podcast going and help us stay ad free and get yourself access to some perks chris hegg jamie felix toll stephen gonzalez reggie deal and steve smeaton thanks to all of you and of course if you are a patreon supporter you can get access to patreon only perks like special bonus episodes every month and the Effectively Wild Patreon Discord group. You can contact me and Meg via email at podcastatfangraphs.com or by messaging us through the Patreon site if you are a supporter, among other enticing incentives. You can join our Facebook group at facebook.com slash group slash Effectively Wild. You can rate, review, and subscribe to Effectively Wild on Twitter at EWPod, and you can join the Effectively Wild subreddit at r slash Effectively Wild.
Starting point is 01:33:35 Thanks to Dylan Higgins for his editing and production assistance. We hope you have a wonderful weekend, and we will be back to talk to you early next week. And now what do you say? Maybe me and mommy, mommy, she just want to go away. I've covered the world. Where did I go?

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.