Updates to ZiPS Three-Year and ZiPS Depth Chart Projections!
There have been a few updates in our ZiPS projections section.
First, we have updated the ZiPS projections for the roughly 300 players in the minor leagues who are in different organizations now than they were when the first run of ZiPS projections went live about six weeks ago. Free agents are still projected as if they were on their most recent team.
Second, the ZiPS three-year projections are now live for everyone in baseball. As usual, these are playing-time agnostic projections; I’m far more interested in a complex set of algorithms estimating a player’s performance than I am in team usage. A projection for a career minor-leaguer of .000/.000/.000 conveys little useful information. I find knowing how a player may perform for a team to have greater utility than a model predicting who front offices and managers will like. ZiPS assumes offensive levels similar to those of the last four years with the most recent years weighted more heavily, and is using the same estimated level of offense for 2022 and 2023 that it does for 2021. It’d be nice to know what effects changes in the ball will have, but as that’s mostly guesswork, ZiPS takes a neutral stance.
The full-season ZiPS comes in two flavors. New this year arethe ZiPS DC projections, which are ZiPS projections already pro-rated for the playing time used in our Depth Charts. Many of you were already manually adjusting ZiPS in this manner, but now it will be done automatically, which should aid in using ZiPS for fantasy purposes.
There are more ZiPS developments in the works. Let us know in the comments what things you’d like to see that I haven’t talked about yet! Our readers frequently have ideas I never thought to do myself and ZiPS has a great deal of flexibility in what it can be manipulated into doing.
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.
In other words, does this mean ZiPS weighs the 2020 “season” most heavily?
A few surprises to me. Tatis is merely projected for 5.5 war per 150 games in 2022, that’s still good of course but it lags behind Acuna and Soto (I’d take the under on Soto projection for sure but we’ve known for some time that zips is crazy high on him). Mainly it appears Zips sees 2020 defensive numbers as a fluke, I’ll take the over on that. Zips is insanely high on Yordan Alvarez, having him at 5.8 war per 150 in 2022, I believe a lot in the bat, but am confused by the 0 defensive projection, are we just ignoring positional adjustments for DHs? I’m a bit surprised by the Cronenworth, Nola, and Grisham projections too but will admit to being a homer.
Dan, 1 suggestion I do have for this is for pitchers that are projected for both starts and in the pen to run 2 separate projections for them as starters only and pen only and then the 1 mixed you seen.
Acuna and Soto have about twice as many career MLB plate appearances so it makes sense that projections would have more certainty about their performance level.
Tatis is great, and probably will continue to be great, but between his injury-abbreviated 2019 and the shortened 2020 season, he’d got less than 1 full season’s PAs at the MLB level. Acuna and Soto, who are similar talents, have the benefit of having full 2019s plus most of 2018.
Acuna has over 1400 MLB PAs. Soto has 1350. (both of those do not include playoffs, where both players add a substantial number, to boot, both having participated in a deep playoff run)
Tatis has 629.
It makes sense that the error bars on Tatis would be wider. Projections can be more confident about Soto and Acuna’s historic greatness, with 2 seasons of performance data instead of 1 (in terms of amount of playing time, not chronological time)
Acuna one I get, Soto not so much given the lack of defensive/base running value he brings. Unless you think he’s like a 180 wRC+ (aka a top 10 hitter of all time) he’s unlikely to be better than Tatis, let alone be a lot better
Tatis’ rookie year was propped up by a crazy unsustainable BABiP that is also still propping up his career totals (though to a lesser degree). There is of course more room for regression with Soto’s offense because it is better, there is also more reason to believe in it because it is over twice the PAs.
Tatis actually underperformed his expected stats in 2020. It’s true that his rookie year end line was fluky but he was still very good. I buy Soto as a 160 wRC+ bat and Tatis at 140 wRC+. People can spin this past year didn’t count or whatever but Tatis made big gains in K%, BB %, and exit velo. As I’m saying above I agree that Soto is the better hitter by roughly 20%, but Tatis’s BSR and defense +positional value will have him earn more war most seasons. And certainly not 7.7 war for Soto so young and with non bat questions
Yeah, that 7.7 is awfully aggressive. Some weird stuff in the 3-year ZiPS. I’m worried it is glitchy. Why is Freeman projection up in 2022 over 2021 then dips heavily in 2023?
But there are just a lot more “cautionary tales” for Tatis than there are for Soto, and I say this as a Braves fan who would love for Soto to become a cautionary tale, but don’t believe that is likely at all.
I think you underestimate the difference between an OBP in the .360s (which is still good!) and one that’s up at .430-.440.
The one piece I’d love to see would be a three-year surplus value based on ZiPS projections, projected salary (there are some excellent models already in the wild that predict arbitration results), and projected $/WAR. I have code partly completed to do something like this with a crappy homebrewed projection system but I would love to see ZiPS flex its computational might and take a stab at it.
Yes please.
And from there, it’s just a hop, skip, and a jump to produce a ZiPS trade value ranking to compare against the annual fangraphs exercise.
I would love to see the 10th and 90th percentile values for the projections for all players, even if it is just the single value of the 90th percentile WAR minus the 10th percentile WAR.
Yasss! Percentile values!
Yes! Some sense of the ranges would be great!
2023 projected leaderboard = reminder not to forget about Kyle Tucker post-prospect hype….
Is there an obligatory followup “3 years ZiPS readouts I was surprised by wrt pitchers/hitters?”
FYI: The Projections drop-down on the top menu still says 2021 and 2022.
Is there something funny going on with Matt Olson’s projections? His WAR leaps from 3.5 in 2021 to 4.4 in 2022 and 4.3 in 2023, even though his offensive stats all stay pretty stable (as you’d expect given his age). Does ZiPS assume he’s spending a lot of 2021 as a DH, or going full Keith Hernandez with his glove in 2022?
That’s kinda weird – I’ll ask David about that. ZiPS own WAR calculation has them pretty close. Perhaps there’s something weird in the FG calculation for 2022 and 2023?
OK, this should be updated now.
Yes, this looks a lot more reasonable now, just from the “eyeballing” perspective.
Is there something funny going on with the projections for the Julio Rodriguezes the catcher from the cardinals has the exact same projections as the OF from the mariners? I feel like the reported scout skill level is very different for those two.