Dan Szymborski FanGraphs Chat – 3/25/21
| 12:05 |
: Hello, been trying to figure out why it’s not on the front page!
|
| 12:05 |
: Oh wait, NOW it is!
|
| 12:05 |
: lol
|
| 12:05 |
: Legend you would most want added to MLB The Show 21?
|
| 12:06 |
: I don’t know who is in it yet!
|
| 12:06 |
: I want to see some 19th century mustache petes
|
| 12:06 |
: When do you get to first set foot inside an MLB stadium this season? Sooner rather than later, I hope?
|
| 12:07 |
: I’ll probably go sometime in May. I have my first shot tomorrow and so give it 3-4 weeks for second shot and 2 for effectiveness and I’m looking at early Mayish.
|
| 12:07 |
: The same credential quotas don’t apply this year, so me going won’t take away the chance of someone who *really* needs to be there
|
| 12:08 |
: Using my BBWAA card won’t *quite* be normal yet, but it’ll be getting there.
|
| 12:08 |
: I’m a Royals fan and want to know your opinion for the most likely current MLB player career comp for Bobby Witt Jr.
|
| 12:08 |
: That’s kind of tricky because of his lack of professional experience!
|
| 12:09 |
: Maybe Wil Cordero? I don’t mean as in he’ll be a bust
|
| 12:10 |
: But Cordero has a lot of phenom traits – he was able to play in the majors very early, there were concerns he’d grow out of the position, had a very strong arm and good power
|
| 12:10 |
: That Cordero didn’t stick at shortstop and his offense regressed a lot doesn’t mean anything ill for Witt
|
| 12:11 |
: Who should trade for Mike Tauchman?
|
| 12:11 |
: So when is the tauchman to the mariners trade going to happen?
|
| 12:11 |
: Honestly, he’s way more useful to the Yankees than Bruce is
|
| 12:11 |
: Are Robbie Ray and Steven Matz for real after both seem to find the strike zone at higher rate than before? If its not a mirage, are the Jays serious contenders for the AL East?
|
| 12:11 |
: It’s certainly a possibility. They were both good rebound candidates in any case
|
| 12:11 |
: Does it feel like there have been more arbiter rulings about option years than ever before? What do you think is behind it and do any of the recent rulings have actual ramifications for teams in how they construct their rosters?
|
| 12:11 |
: I think the messiness of the COVID year made everything a little tricky
|
| 12:11 |
: Best soft pretzels in Cincinnati…..for those with a jaw condition. Have a great chat Dan…we are here to help and serve you.
|
| 12:12 |
: I don’t have an actual jaw condition, I just ate too many pretzels lately!
|
| 12:12 |
: With Adam Haseley back and Odubel Herrera’s cooldown are Phillies fans safe from having to root for a team with Herrera on the major league roster?
|
| 12:12 |
: i DUNNO, i STILL THINK HE MAKES THE TEAM
|
| 12:12 |
: We are not already having the “Tatis Jr. is a fragile player” are we? He is SO SO talented and fun to watch… Hamstrings, back injury, hand injuries..(thank you head first slides) and the recurring shoulder issue now. Say it ain’t so Dan. I want to see him play 150 games fully healthy.
|
| 12:12 |
: I hope soo too.
|
| 12:13 |
: I always grumble a bit about ZiPS giving him low games because of his missed time
|
| 12:13 |
: Hey Dan, ZiPS is great and really appreciate your expertise in these chats and in your articles. A while back Ariel Cohen posted and article about auction results comparing the different projection systems and ZiPS seemed to come out the worst. Do you have thoughts about the article or its conclusions?
|
| 12:14 |
: Ariel’s methodology is very fantasy focused, so ZiPS can’t possibly look good given that it’s playing time estimates are not “ready made” for fantasy in this way.
|
| 12:14 |
: That’s not a knock on Ariel; he’s looking for something very specific!
|
| 12:15 |
: How depressed will Pirates fans be when Cole, Glasnow, Musgrove and Morton all receive Cy Young votes? And what the hell, give Taillon one too.
|
| 12:15 |
: I think Pirates fans are already depressed!
|
| 12:15 |
: Dan! I know you aren’t a fantasy writer but I’m on a deadline here so I’m hoping you will help. What is your outlook on Gavin Lux this year and moving forward relative to Jean Segura? I could keep Segura at 5.5 or trade him for Lux at 11.00 and have Lux into the future.
|
| 12:16 |
: I think I’d keep Segura if your team is strong. I love Lux, but Segura has some characteristics that are just weighted heavier in fantasy than real life
|
| 12:17 |
: Seeing Arenado’s career home/road split it is hard not to buy into those pundits projecting a rather severe drop. You are perhaps the most pessimistic of all. Is there any hope for a better outcome?
|
| 12:17 |
: There’s quite a lot of hope!
|
| 12:17 |
: (Incidentally, there’s something weird going on with the wRC+ in the projections)
|
| 12:18 |
: But there are risks involved with Arenado
|
| 12:18 |
: What are your altered expectations for Gallen vs ZiPS altered expectations? Seems more likely this was a gradual pitching injury exacerbated by hitting
|
| 12:18 |
: ZiPS doesn’t have much to say about a stress fracture
|
| 12:18 |
: I hope it wasn’t a gradual injury
|
| 12:19 |
: And I’m a bit worried – I remember Tony Saunders
|
| 12:19 |
: Berrios 2.7% chance to win the Cy. How does your opinion of Berrios compare to what Zips is predicting?
|
| 12:19 |
: I’m about even with ZiPS on this one
|
| 12:19 |
: Given projections and what you have seen in spring who do you think the Mariners should start in LF on opening day?
|
| 12:20 |
: I’d just stick out Taylor Trammell to keep getting him at bats
|
| 12:20 |
: A week or so ago you wrote the article “What Did Teams Pay Per Win in Free Agency”. You found a lot of dispersion — you focused on the nonlinearity of value per win. But another point you should be looking at: the assumption of a single market clearing price for wins. This is almost surely wrong in the free agent market, where you probably should be using something like an auction model. Happy to blab more about this if of interest and what you’d want to look for in the data.
|
| 12:21 |
: I know what you’re getting at in the game-theoretic models
|
| 12:21 |
: but I’m not convinced such an approach makes sense.
|
| 12:21 |
: First, the data is just awful – there are very few players, so I prefer a simple approach
|
| 12:25 |
: And second, it’s not *really* an auction that would fit into the generalized auction formats
|
| 12:26 |
: I’m just not sure that the payoff is worth the mess, honestly.
|
| 12:26 |
: Who makes more starts for the Jays, Tanner Roark or Anthony Kay?
|
| 12:27 |
: Roark
|
| 12:27 |
: I have panama-erie-suez in my canal ranks. Is that right?
|
| 12:28 |
: no love for volga?
|
| 12:28 |
: Does ZiPS weigh different injuries heavier than others for the player’s future? If so what are the “worst” common injuries a player (and pitchers) can suffer in general according to ZiPS.
|
| 12:28 |
: Generalized.
|
| 12:28 |
: Most relevant for pitchers
|
| 12:29 |
: Labrum/rotator cuff/capsule worst for pitchers
|
| 12:29 |
: then generalized shoulder
|
| 12:29 |
: then TJ
|
| 12:29 |
: This year’s positional series has featured multiple instances of analysts questioning the reliability of projections for certain players. As the source of half of the projections, how does that make you feel?
|
| 12:29 |
: It would be boring if we didn’t fight with the projections!
|
| 12:30 |
: Remember: all models are wrong, but some are useful
|
| 12:30 |
: If you were running a team would you rather have a guarantee your team wins the next World Series or a guarantee your next first round pick is a 1st ballot Hall of Famer?
|
| 12:30 |
: The latter
|
| 12:31 |
: If I have that latter guarantee and he doesn’t know that, I offer him a 20-year, $200 million contract in the low minors and try to use his skills to win several!
|
| 12:31 |
: Why didn’t I know ZiPS wasn’t made for fantasy the year BEFORE I used it?
|
| 12:32 |
: I mean, it’s useful, but it’s playing time agnostic, which is a huge deal for fantasy
|
| 12:32 |
: Wander would have gotten a *way* higher dollar figure than was warranted, because ZiPS didn’t care he wasn’t going to be the starter in 2020. Fantasy players DO care though!
|
| 12:32 |
: Ockam’s Razor.. when considering a complex problem..the simplest approach/solution is often the best..
|
| 12:32 |
: either that.. or you’re a simpleton that likes chili…with beans..
|
| 12:32 |
: Who ends up having a better year in the TO OF? Teoscar or Lourdes?
|
| 12:33 |
: I think Teoscar, but there’s some genetic thing that Gurriels have that always make me underrated them
|
| 12:33 |
: Did you think the extension offer of 6/$120M was a fair one to Correa? And do you expect him to increase his value this season or will he likely regret not taking the offer?
|
| 12:34 |
: Definitely on the low side. ZiPS suggest 6/165
|
| 12:34 |
: Well, take a look at the potential number of landing spots for a given player (based on which teams had “holes” in a given position) and whether those teams are trying to win (and maybe if those are high-budget teams) and see if it’s correlated with the price per win. This should give a good signal about whether this is important.
|
| 12:34 |
: OK, I sent the last response before I saw your full reply (sorry, I’m a jerk!). Anyway, the other thing I think you’d get here is some deeper insight into the performance of a front office. The good front office will manage future holes to line up with markets where the price per win is likely to be low.
|
| 12:34 |
: Wait, where are you being a jerk?
|
| 12:34 |
: The problem is that there are relatively few free agent decisions made
|
| 12:34 |
: We’re hampered by the sparseness of our reality here.
|
| 12:35 |
: If ZiPS got Weird Science’d what would its human form look like?
|
| 12:36 |
:
|
| 12:36 |
: Matt Harvey makes Orioles Roster: Discuss
|
| 12:37 |
: I don’t expect much, but his velocity was up and he was more effective after that first appearance
|
| 12:37 |
: Blink once for “yes,” twice for “no:” have you ever used ZiPS for betting purposes?
|
| 12:37 |
: Of course! I’ve explicitly done so!
|
| 12:39 |
: At this point in you and Zips’ career, do you enjoy changes in the game to see how responsive Zips can be, or would you prefer lengthy stability?
|
| 12:39 |
: I like stability!
|
| 12:39 |
: I’m less focused on mean projections than the other stuff
|
| 12:40 |
: On Correa, we know he and Mike Elias are somewhat close. Any chance he comes to Baltimore?
|
| 12:40 |
: I guess it’s possible, but I don’t think it’s likely at this point
|
| 12:40 |
: You said last week that cats shouldn’t have human names. Does Joey Bag of Donuts count?
|
| 12:40 |
: It’s OK to have a human name as PART of the name
|
| 12:40 |
: So a buddy of mine decided to make “pizza burritos”. Is that not just a Chipotle sized pizza roll? Or does the fact that he used a high fiber tortilla in fact qualify it as something else? If I made the same with pizza dough would it be a true, overgrown pizza roll or just a monstrosity?
|
| 12:40 |
: I think a pizza burrito is just a mediocre calzone
|
| 12:40 |
: Now I am considering making a chili filled, burrito sized pizza roll with proper pizza dough.
|
| 12:40 |
: Do your projections give greater weight to 2020 for metrics that stabilize more quickly? Would you give greater weight to K% that, say, ISO because K% stabilized more quickly (I assume) even though the two are correlated?
|
| 12:40 |
: Yes
|
| 12:40 |
: Though they already do so
|
| 12:41 |
: If you discovered some ultra secret version of ZiPS that was 100% accurate in predicting MLB win totals, would you keep it to yourself and just bet those teams and make a fortune? Because once you release it wouldn’t the public/casinos try to backward engineer it and then negate your advantage? Would you still release the “lesser” version of ZiPS for all us plebes?
|
| 12:41 |
: If I had that, I’d just bet myself, make a fortune, and retire from baseball writing
|
| 12:41 |
: If I had this kind of magical accuracy, baseball would be utterly boring for me to watch
|
| 12:41 |
: On a more serious note, how often do you/have you updated ZiPS with newfangled statcast/trackman/etc data such as EV, Barrel%, etc? Do you wait a few years to verify that these measurements are accurate and useful or do you trust some right from the get-go?
|
| 12:41 |
: Oops, sorry I keep answering so many of your Qs
|
| 12:42 |
: (to others)
|
| 12:42 |
: I try to be random, but there were a lot
|
| 12:42 |
: (don’t want anyone to feel left out)
|
| 12:42 |
: There’s always ZiPS versions that are a few years out. I don’t like including things until I have a good fix on the nature o their predictive value
|
| 12:42 |
: I agree sparse reality is hard. That’s why I’m suggesting just counting number of fits (could probably scrape off of MLBTR’s free agent overview of something similar) and the count of these teams trying to win and the coiunt not trying to win. So $/win = #fits + constant + error or perhaps dividing #fits by trying to win. It’s not “right” but it will help you understand if it’s important. And FYI the academic literature definitely does fit auction models with relatively few observations to identify things like underlying valuations and whether collusion is trying to take place. Thanks for your patience!
|
| 12:42 |
: I’m quite behind on the academic literature!
|
| 12:43 |
: Two questions: when using zips to bet do you make your bets before you make zips data public? (I’d assume yes) Do you make a bunch of smaller bets to try and take advantage of let’s call it diversifying your risk, since the law of large numbers would require too many bets to be practical for an individual, or a few larger ones?
|
| 12:43 |
: I don’t really have a pattern
|
| 12:43 |
: Generally, I bet when I think about it
|
| 12:44 |
: Curious if you had to rank these 4 recovering aces for 2021, how would you rank them, Sale Clevinger, Syndergaard, Severino? Should we consider them likely to be fully recovered by then?
|
| 12:45 |
: Generally, Syndergaard >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Sale/Severino >>>>>>>>>>>>>>>>>> Clevinger
|
| 12:45 |
: Remember, Clevinger’s already ruled out for 2021
|
| 12:45 |
: no?
|
| 12:45 |
: And Syndergaard seems to have had the smoothest reccovery
|
| 12:46 |
: Severino’s had essentially two lost years
|
| 12:46 |
: and Sale’s goal being to pitch in games “at some point in 2021” is not a particularly sunny one
|
| 12:46 |
: Doesn’t Severino have a domestic violence issue that also clouds his future in addition to the injury?
|
| 12:46 |
: Wait, does he? I’ve heard nothing of this
|
| 12:47 |
: If you take a large soft pretzel….place some meat on it and fold it over ..is it a sandwich?….a pretzel burrito?…..a pretzel hot dog? it has no implied hinge in the design like a hot dog bun and the open area of a pretzel prohibits condiments from being liberally applied so it doesn’t really qualify as bread does it? Have we created a new category of food item here?
|
| 12:47 |
: I think it’s a pretzel taco?
|
| 12:47 |
: Pick one with the caveat of being completely healthy for their entire career, Eric Davis or Tatis?
|
| 12:47 |
: Tatis
|
| 12:48 |
: Love Eric the Red, but Bebo is a shortstop
|
| 12:48 |
: Joey Bag of Donuts was his official name, but in everyday use we just called him Joe or Joseph. My daughter named him before he grew up to look like a bag of donuts. His brother is Little Anthony, or just Ant, which in South Philly is a human name.
|
| 12:48 |
: No that’s German.
|
| 12:48 |
: I knew about German obviously, but I was wondering if there was Severino TOO
|
| 12:48 |
: and I just hadn’t seen it
|
| 12:48 |
: What % of players do you think are totally fine with expanded playoffs in exchange for the DH? I disagree with MLBPA’s argument that teams will spend less in free agency because of it. Just do it!
|
| 12:48 |
: Too many for the MLBPA’s case
|
| 12:49 |
: and knowing front offices, teams damn well will spend in free agency
|
| 12:49 |
: One of the my standard reports that I’ve sold, division/playoff probability at every win number, isn’t one fo the more popular ones for a reason
|
| 12:50 |
: The more randomized success is, the less value individual players have.
|
| 12:51 |
: Are MLB owners the poorest rich people ever?
|
| 12:51 |
: |
| 12:51 |
: Work from home means I can grill hot dogs for lunch, it’s not such a bad thing.
|
| 12:51 |
: So more just for fun than a planned portion of your income?
|
| 12:52 |
: Between my main job, investments, and the occasional data sale, I have little desire to gamble for anything but fun
|
| 12:52 |
: Eking out a profit in gambling is SERIOUS work, and I’m against that notion it practicable!
|
| 12:52 |
: my aunt once called her cat Eldridge Cleaver…thought that was cool.
|
| 12:53 |
: Sorry, I screwed up the question. Meant for 2022!
|
| 12:53 |
: Oh, I won’t make any mental rankings of healthy pitchers a year from now while still recovering in 2021
|
| 12:53 |
: Dan, your revelation that ATC is fantasy oriented and ZIPS is “agnostic” is positively tsunamic! What about Steamer, BAT, BATX which do they favor fantasy or reality?
|
| 12:53 |
: Is this a stunning revelation?
|
| 12:55 |
: I mean, the ZIPS projections for the Rays have them throwing like 4000 innings with 491 starts.
|
| 12:56 |
: Playing time is a big deal for fantasy, which is why there’s a depth chart version this time around
|
| 12:56 |
: Is *any* filled and folded bread product a taco? Is there some threshold of implied southwestern-ness that needs to be met? Like if I take my leftover corned beef and cabbage and wrap it in a tortilla, is that a taco?
|
| 12:56 |
: I dunno, we have to have a summit to lay down all the rules
|
| 12:56 |
: This doesn’t seem like a terrible idea at all: https://theathletic.com/2474483/2021/03/24/mlb-will-use-statcast-lab-t…
|
| 12:57 |
: I dont’ think it’s a good idea to use analytics as a punishment device
|
| 12:57 |
: *sarcasm
|
| 12:57 |
: ah
|
| 12:57 |
: If it’s not in a corn tortilla it’s not a taco
|
| 12:57 |
: Now we’re getting into structure vs. ingredient issues
|
| 12:57 |
:
|
| 12:58 |
: I’m in the top corner
|
| 12:58 |
:
|
| 12:58 |
: Same here
|
| 12:59 |
: Corn tortillas are useless if not doubled up. If you don’t have to double up a corn tortilla, your filling is lacking.
|
| 12:59 |
: I’m ingredient chaotic for tacos. Spicy peanut chicken?
|
| 12:59 |
: Dan, are you bombing Italy or France?
|
| 12:59 |
: Nobody ever says italy
|
| 12:59 |
: Disagree on the corn tortilla requirement. I’ve had very authentic Mexican food truck tacos with flour tortillas.
|
| 12:59 |
: sorry did not mean to be condescending — this has been a fascination for me back since when dave cameron used to write about marginal cost per win. always dreamed of writing the article, realized I just don’t have time and figured I might as well pass the idea off to folks who might use it
|
| 1:00 |
: It might be something interesting to explore
|
| 1:00 |
: but not really just for valuation
|
| 1:01 |
: I think an actual econometrician who specializes in this branch of game theory would be able to come up with something better than I am.
|
| 1:01 |
: I was a math/econ undergrad guy. I’m not a professor, I have no doctorate, and by nature of my job, I’m a generalist.
|
| 1:01 |
: Nor am I particularly smart.
|
| 1:02 |
: I’m basically just a dumbass writer who has a moderately useful secondary skill set for his job
|
| 1:02 |
: Since you do sometimes address roto questions, are the West Coast pitchers’ parks relatively OK for the hitters actually early in the season, when it’s cold back East and Midwest? Or maybe Santa Anna wind patterns make them worse? No one ever addresses this, and it’d sure be useful to know.
|
| 1:02 |
: I don’t think it’s a big enough effect to change opinions in most cases
|
| 1:02 |
: Are you going to to do your “Most Indispensable” piece soon?
|
| 1:02 |
: Yeah probably early in the season and the first big injuyr!
|
| 1:03 |
: What do my cat friends do?
|
| 1:03 |
: Lazy day for cats
|
| 1:04 |
: Mercutio was roughhousing with Justinian before I came down for chat but Justinian’s sleeping on the office couch right now
|
| 1:04 |
: Constantine was chirping at some birds outside the kitchen window
|
| 1:04 |
: and Cassie was asleep at the top of the cat tree
|
| 1:04 |
: In your mind, at what point do non-legume ingredients disqualify a dish as chili? Is corn acceptable? Mushrooms? Seaweed? How about if it’s served over rice?
|
| 1:04 |
: NONE
|
| 1:05 |
: I think you’re too down on yourself, I think your stuff is pretty f-ing cool! (Also, you get paid to write about baseball, while I don’t which is why I don’t make the time to write about it.)
|
| 1:05 |
: Everyone should know who they are and who they are not.
|
| 1:05 |
: I have strong points, but I’m also perfectly at piece with my weak ones.
|
| 1:05 |
: In essence, this is the basic challenge of life, is it not?
|
| 1:06 |
: One of our final life goals is answering the question “Is it OK to have been me?”
|
| 1:06 |
: I was pretty surprised by the ZIPS (and other systems’) projections of Pache’s defensive value. Any idea what’s going on there?
|
| 1:06 |
: In what way?
|
| 1:07 |
: ZiPS had him a +6 based on minor league data
|
| 1:07 |
: I use gameday tracking to make a probability-based model of every chat
|
| 1:07 |
: Baseball chat with a side order of existential angst…
|
| 1:07 |
: I’m not actually very angsty
|
| 1:08 |
: One of my greatest hopes is that my final words are me joking and/or making light of my imminent demise!
|
| 1:08 |
: Does Zips get into if a player will stay at a position? For that 6/$165 on Correa, does Zips see him as a SS all 6 years?
|
| 1:08 |
: Generally speaking, yes
|
| 1:08 |
: Though I’ve played with a more dynamic model
|
| 1:08 |
: But I try to keep the interpretations from getting too confusing
|
| 1:09 |
: Getting into corn/flour tortillas involves talking about the history of colonialism and imperialism which seems beyond the scope of this chat.
|
| 1:09 |
: (flour tortillas are colonialism in action)
|
| 1:09 |
: How long does it take to process/update a new version of zips? Has the time measurably decreased since you switched to your 5900x?
|
| 1:09 |
: ZiPS is about 60% faster with a 5900x
|
| 1:10 |
: Turns out I’m an anarchist. I also put beans in chili.
|
| 1:10 |
: Would you bet your 3080 that Luzardo has a better year than Donaldson?
|
| 1:11 |
: No. Why are you making me make high-stakes bets with my things?
|
| 1:11 |
: Looking at the CF projections, I saw that Pache’s defensive value is pretty low (for someone described as an 80 grade defender). Digging deeper, it looks like Steamer is actually quite pessimistic about his defense compared to ZiPS. Do you know what might account for the difference of opinion there?
|
| 1:11 |
: ZiPS is based on the minor league defensive data and it’s conservative; the data is less accurate than a UZR/DRS approach in the majors
|
| 1:12 |
: ZiPS had Luis Robert a +5 in CF last year coming off like a +21 season in the minors (lower before 2019)
|
| 1:12 |
: I don’t think Steamer uses any minor league defensive data
|
| 1:12 |
: but I could be wrong
|
| 1:12 |
: I have enough to worry about explaining ZiPS stuff without Dansplaining someone else’s work!
|
| 1:13 |
: just remember that “the wallpaper and i are in a duel to the death, either it goes or i do” is already taken, last-words-wise
|
| 1:13 |
: What do you think is the next major thing to unlock for projection systems? Is it taking the jump into more advanced ML like decision trees or neural nets? Or incorporating more granular data like spin, etc?
|
| 1:13 |
: We’re still figuring out the best approaches to use some of the advanced data available
|
| 1:13 |
: there’s so much to sift through and we’re still figuring out how predictive it is, multicollinearity, and so on
|
| 1:14 |
: If you had a Doctorate you would need the tweed coat with elbow patches and old timey pipe..as I said last week.
|
| 1:14 |
: On that note, time fo rme to head out for another week! Thanks for coming everyone and I’ll hopefully see you in our Opening Day chat next week!
|
Dan Szymborski is a senior writer for FanGraphs and the developer of the ZiPS projection system. He was a writer for ESPN.com from 2010-2018, a regular guest on a number of radio shows and podcasts, and a voting BBWAA member. He also maintains a terrible Twitter account at @DSzymborski.
Dan Szymborski