Assuming league-average pitch shape, up and away on the corner at 92 vs. middle-middle FF at 97:
xRV/100: -1.67 vs. -2.27
The 97 MPH fastball thrown middle-middle would be ~0.6 runs better per 100 pitches than a 92 MPH fastball spotted on the corner.
Now you’re not comparing apples to apples. We’re not talking about a deadzone 97 mph middle vs a 92 mph elite sinker that’s being tunneled with other pitches. We’re talking two 4 seams with = metrics i.e. spin, extension, HB and IVB. 97+ middle is more challenging than 92 corner.
Grayson Rodriguez looks legit. All 5 of his pitches had Stuff+ marks above 110 and he had elite fastball command (117 Location+) last night, which is especially difficult in an MLB debut. If he can improve the breaking stuff command, he's a top-10 starting pitcher in baseball.
Yordan Alvarez mashes no matter what you throw him. This Ohtani sweeper had a 222 Stuff+, 133 Pitching+ and -3.35 xRV. That's the 7th highest Stuff+ of any pitch hit for a homer in 2023. Unsurprisingly, Alvarez is the best hitter in MLB against Stuff+ >= 120 by 85 points of wOBA.
Insane piece of hitting right here:
- 44% whiff prob (if swung at)
- 50% CSW prob
- 12% prob of being put in play
- 49% ground ball prob (if put in play)
- -5.77 xRun Value per 100 pitches
- 0.33% prob of being hit for a homer
Just gotta tip your cap🤷♂️
There are differing opinions on how much home runs are under a pitcher's control. FIP looks at pitchers as fully responsible for home runs, whereas xFIP attempts to address some variation by looking at fly balls instead. However, I would like to suggest a new method, p(HR):
🧵
With fantasy baseball drafts and spring training approaching, I wanted to create a resource to share my pitch quality metrics, so I put together a web app with the help of
#rstats
and {shiny}!
Stuff+ is great because it stabilizes super fast, especially for fastballs. Already, among fastballs that were thrown at least 250 times last year and 25 times this year, YOY R^2 = 0.65. This means big changes are meaningful, like Kris Bubic's, who added 75 points of Stuff+!!!
A look into how pitching staffs have developed, in terms of stuff quality, from 2020 to 2022. The
#Yankees
have led the league in Stuff+ the past two seasons, while the
#RedSox
saw the biggest jump in the league from '20 to '22, and the
#Dbacks
saw the biggest drop-off.
Good point here from Anthony, in terms of decrease in Stuff+, the bottom quartile of pitchers in tempo from 2022 have been hurt the most, very likely due to the new pitch clock. Interestingly, there is a non-negligible relationship between pitcher tempo and stuff quality.
I had a theory that not all velo differences are created equal, so I compared Stuff xRV/100 at different velo & velo diff combos (MLB average slider velo diff is ~10 MPH). I found that fast sliders are hurt more by a small velo difference than slow sliders.
I've continued to learn python and recently created a simple xRV model. Below is a leaderboard of pitch type xRV+ leaders and a few heatmaps of model predictions (code linked in next tweet):
I've been taking a look at python/pandas recently and have finally started to get a bit more comfortable with data manipulation and aggregation, here's a leaderboard of pitcher pitch type CSW% this season, created in pandas:
These are the best pitch quality adjusted hitters of 2023 so far. Adam Duvall of the
#RedSox
has put up elite production while facing high-quality pitching, making his success even more impressive.
#Brewers
rookie Brice Turang has also excelled against elite pitching.
I put together an app to share my similarity scores! You can check out both hitter comps and pitcher pitch type comps. Shoutout to
@DolphHauldhagen
and
@NotTheBobbyOrr
, their work on similarity scores was the inspiration for this project.
This awesome thread inspired me to create a new stat, Wiggle Room+: the amount of area in inches a certain pitcher pitch type can be thrown to and expect good results (a negative xRV), compared to the average for that pitch type. In other words, the blue area on a heat map.
🧵
While stuff dominates the game, if a pitcher cannot at least be somewhat competitive around the zone, then that stuff won't be very useful.
An example of the effect that location has on wExecution+ using Aaron Nola's CB - which shows the effectiveness of it as a chase pitch.
🧵
With my pitch quality models, I took a look into the quality of pitches a hitter sees through an at-bat. Although pitches earlier in an at-bat tend to have worse stuff, they also tend to be located better, so a hitter is better off waiting for a mistake or trying to work a walk.
Only 9 hitters had above-average contact ability, swing decisions, launch angle value, and exit velo value. Some names are expected, like Alvarez, Freeman, and Machado, some are young rising stars like Pasquantino and Murphy, and some are surprises, like Rivera and Arcia.
Chase% OE (over expected) factors in difficulty in holding back a swing, and although similar to the actual Chase% leaderboard, it better credits hitters like Brice Turang, who would have been expected to chase ~38% of OOZ pitches but has offered at only 14%.
Shane Baz = Jacob deGrom?? For both the FF and SL, deGrom's offerings are in the top 10 most similar pitches. He's set to miss 2023 with TJ surgery, but he'll be a big part of the
#Rays
future.
I've recently been working on rewriting my pitch quality model training script(s) in python and I wanted to share the function I wrote to process statcast data and add all the features I use in modeling (I'm also working on a polars translation):
I've seen multiple people who work in python wishing for mlbplotR-like capabilities, so I adapted some
@CFB_Data
code to at least plot team logos (such as in the plot below), hopefully someone can use this as a starting point for further development (code in next post)
It's hard to draw many conclusions early in the season, but one of my favorite things to look at because of its quickly-stabilizing nature is fastball Stuff+. Here's a look at the pitchers who have improved their fastball the most in 2024:
Within the first week of the season, you have a really good idea of how good a pitcher's fastball is. First week Stuff+ has an R^2 of 0.70 with rest of season Stuff+.
The
#Dbacks
got an absolute weapon in Miguel Castro. He’s got three pitches with elite stuff (his
#1
SL comp is Matt Brash🙃) and a deceptive delivery with a 5 ft release point from a 6’7” frame.
I figured out how to pull info on weather from the mlbstats api so I used that data to see which ballparks are the best to go take in a game at. Lots of California teams at the top.
*Note: The way I defined good weather doesn't include domes/roofs, which penalizes TB/MIA/HOU/etc.
Breaking balls should be swung at almost everywhere in the zone, whereas hitters should lay off fastballs at the top of the zone and swing more at low fastballs, and vice versa for offspeed pitches. *Swing% OS means Swing% Over what hitters Should.
@Drew_Haugen
From a previous newsletter, this chart showed where hitters (as an MLB aggregate) should swing for a few common (MLB aggregate) pitch types. What I'd love to see is a heat map of the difference between "should swing" and "do swing". (Drew, do you have that?)
I've been taking a look at python/pandas recently and have finally started to get a bit more comfortable with data manipulation and aggregation, here's a leaderboard of pitcher pitch type CSW% this season, created in pandas:
Interesting idea here, prompted me to look into the correlation between pitch height and launch angle at the hitter level. League wide, the correlation between the two is 0.217. Top and bottom 3 hitter plots along with leaderboards below, full leaderboard in next tweet:
I feel like there is an obvious way to pitch to Juan Soto.
Consider his median launch angle on high pitches vs. low pitches.
I call it the "hit it where it's pitched" syndrome.
Josh Hader's sinker might be my favorite pitch in baseball. It has a 230.1 Stuff+ (my models, highest of any pitch type in MLB) and 122 Pitching+ despite a 93.8 Location+. His stuff is elite and gives him a much higher margin for error than a league average sinker.
This question prompted me to investigate how pitcher handedness affects success while controlling for stuff quality. When hitters have the platoon advantage (RHB vs LHP & vice versa), LHP usually outperform RHP with the same quality of stuff.
Stuff+ question -- is a 100 Stuff+ pitch from a righty just as effective as one from a lefty? Or maybe, does a lefty perhaps outperform a righty with similar Stuff+ ratings? (Due to differences in platoon rates and lower frequency of LHPs overall?)
Stuff grades are great because they rate just pitch specs, which vary far less than location, but there is a slight relationship between a pitch's stuff and location value. Pitches thrown in optimal locations tend to have better stuff, being well executed pitches overall.
The way in which we summarize distributions can hide valuable information. This is often cited in relation to average launch angle, but Michael A. Taylor is a great example of how this concept also applies to exit velocity. A 🧵:
A comparison of my Stuff+ grades vs.
@enosarris
Stuff+ grades is interesting. His models tend to prefer pitchers with higher xCSW%, while my models slightly prefer those with higher xGB%.
Here's the same viz, but with location quality. The
#Dodgers
and
#Rays
have consistently been among the league leaders the past three years, while the
#BlueJays
improved 19 spots from 2020 and the
#Rangers
dropped 26 spots.
A look into how pitch quality progresses throughout the season. Stuff+ is at it's highest at the beginning of the season, Location+ is pretty steady throughout, and Pitching+ peaks at the end of the season.
Here's a look at how team stuff quality progressed over the 2023 season, teams like the
#Cardinals
and
#Twins
took a step back at the end of the year while teams like the
#WhiteSox
and
#RedSox
actually saw their stuff tick up around September/October.
Justin's comment on Kimbrel's volatility made me curious as to which players have the most consistent pitch quality outing to outing. First, the leaders and laggards in Stuff+ consistency, measured by coefficient of variation:
One of the biggest problems with Avg. EV is that it is non-linear with outcomes. As
@_Ben_Clemens
explained it, "production on contact doesn’t vary linearly with exit velocity, so you end up conflating unlike things". To adjust for this, I created a new stat, EV+. 1/x
Although there isn't a clear pattern at the league level, the residual between inferred axis (from pitch movement) and arm angle (for sinkers) is more meaningful at the player level. Players with an absolute residual >= 28.5 (those highlighted) had an average Stuff+ of 167!!
@Drew_Haugen
@Mind_OverBatter
@EliBenPorat
would be interested to see arm angle plotted against “observed movement” spin direction (i now call it expected movement direction)
i think that’s where you’ll find the interesting stuff. what movement direction you expect based on arm angle versus how the ball actually moves
Interesting point here by
@enosarris
. A prime example of this is Yordan Alvarez, who saw the second-largest jump of p(XBH) based on pitch characteristics from '21 to '22 in the league, one of the reasons behind his ISO jump of 68 points. Pitch selection📈 = quality of contact📈
tldr: Using pitch quality data, home run probability can be estimated, which lends a better representation of pitcher talent and home run control.
Google Sheet with data for all pitchers who threw at least 300 pitches in 2022:
h/t
@Mind_OverBatter
@EliBenPorat
for their work on arm angles, which are super fun. I wanted to look into the importance of deviating your movement from your slot (arm angles as a proxy for slot) as Robert said. It's pretty important! For FF, vMov residual vs Stuff+ R = 0.5!
@enosarris
@jokeylocomotive
@fangraphs
Yea a former organization I played with stressed the importance of having your movement deviate from your arm slot. SSW makes that easy to do.
I put together a swing decision metric based on the expected run value of a swing vs. a take, and the Mariners had a big improvement this year. Cal Raleigh has the biggest improvement of any hitter in baseball with at least 500 pitches seen in both 2021 and 2022.
Part of what makes splitters fun, outside of their rarity in MLB and insane effectiveness, is, as Nick said, their volatility. In 2022, MLB splitters had a Stuff+ standard deviation of 52.2, while no other pitch type was over 50.
#BlueJays
3rd baseman Matt Chapman has been tearing it up in the final year of his contract. He has an insane 67.4 Hard-Hit% this year as well as a launch angle distribution 7% better than league average, which are both big factors in his 100th percentile .473 xwOBA.
I wanted to compare how hitter's "preferred" attack angle lines up with their performance on good stuff, elevated four-seam fastballs. For the sample of hitters who have seen at least 65 good elevated FF, there is a decent correlation between attack angle & Whiff% (R = 0.44).
I built an xgboost model to predict launch angle to see how much swing decisions are influencing hitters' launch angles. Walker actually has a an above average expected launch angle based on the pitches he's hit (86th pctl.) but continues to struggle to lift the ball.
re: jordan walker's demotion, his low launch angle, and "working on his swing":
(image 1) he likes low pitches
(image 2) on average, pitches at the bottom of the zone yield LAs 20°+ flatter than pitches at the top
so, did he work on his swing, or did he work on swing decisions?
Within the first week of the season, you have a really good idea of how good a pitcher's fastball is. First week Stuff+ has an R^2 of 0.70 with rest of season Stuff+.
Stuff+ is great because it stabilizes super fast, especially for fastballs. Already, among fastballs that were thrown at least 250 times last year and 25 times this year, YOY R^2 = 0.65. This means big changes are meaningful, like Kris Bubic's, who added 75 points of Stuff+!!!
-35% swing prob
- 58% whiff prob (if swung at)
- 52% CSW prob
- 4% prob of being put in play
- -2.91 xRV/100
- 0.16% (that's 0.00157) prob of being hit for a homer
That's the 2nd lowest p(HR) of any pitch actually hit for a homer this year.
Awesome observations here from Lance, here are the most notable of these changes to me:
-Ryan SL to ST, Stuff xRV/100 2.2 runs better than SL
- Keller SL to CU/ST, takes advantage of supination bias with two distinct breakers, ST Stuff xRV/100 1.4 runs better than SL
-Springs📈
April 2 pitcher shape metrics of note.
#MLB
My favorite edition so far.
Quick thoughts on Jeffrey Springs, Joe Ryan, Noah Syndergaard, Graham Ashcraft, Tyler Anderson, and Brad Keller.
The lowest p(HR) pitch hit for a homer this year? Who else but Shohei Ohtani with a bomb off this Ken Waldichuk sweeper that had just a 0.039% probability of being hit for a homer (10% swing prob, 2% prob of being put in play).
-35% swing prob
- 58% whiff prob (if swung at)
- 52% CSW prob
- 4% prob of being put in play
- -2.91 xRV/100
- 0.16% (that's 0.00157) prob of being hit for a homer
That's the 2nd lowest p(HR) of any pitch actually hit for a homer this year.
An update of Max's viz, for the 2022 season. The
#Dodgers
and
#Rays
led baseball with five different pitch types with better than average Stuff+ and Location+, or the top right quadrant of the graph. The
#Marlins
and
#Orioles
were just behind at four each.
#Brewers
pitcher Bryse WIlson has improved his command significantly this season. His Location+ per FG/
@enosarris
has jumped from exactly average in 2023 to 108 in 2024. He is elevating his 4-Seam very well, burying his curveball down, and getting his sinker down and in to RHB.
Brock Stewart's fastball is more impressive the more you look into it. It gets whiffs at the highest rate of any pitch (relative to its pitch type), and gets called strikes at an above average rate. Pitch types with >= 175 SwStr+ and 115 Called Strike+: Stewart FF and Snell CH
I've been taking a look at python/pandas recently and have finally started to get a bit more comfortable with data manipulation and aggregation, here's a leaderboard of pitcher pitch type CSW% this season, created in pandas:
I've been playing around with the mlbplotR package lately, which is a super cool tool from
@k_camden
. Here is Stuff+ above average by team and pitch movement for one of my favorite pitchers, Matt Brash, with his headshot. Can't wait to use this package more!
#rstats
I took a look into some hitting metrics for the great American states 🇺🇸. New Jersey (Mike Trout's home state) leads all states in xwOBACON, Delaware (Paul Goldschmidt's home state) leads in avg. exit velo, and Minnesota leads in SwStr%.
Pitch modeling is huge for talent evaluation, get a look into some prospect pitch quality (Stuff+/Location+/Pitching+) with some graphs, tables, and a full leaderboard in the article below!
Kyle does some of the best work on this app and is super great about offering advice/sharing code, notebooks like these are a fantastic resource for anyone interested in data manipulation + visualization in python.
Splitters also have the most volatile location quality of any pitch type. The unique grip and release of a splitter is what makes the pitch so effective, but that grip/release also makes it difficult to create a consistent movement profile and have consistent command.
Part of what makes splitters fun, outside of their rarity in MLB and insane effectiveness, is, as Nick said, their volatility. In 2022, MLB splitters had a Stuff+ standard deviation of 52.2, while no other pitch type was over 50.
All of the worst xRV pitches are high HBP prob pitches, so these are the 5 worst xRV pitches with a swing prob > 50%. These pitches had an avg. EV of 102 and .615 xwOBA, but just a .250 wOBA. Even on the worst pitches, pitchers often still come out on top.
Pitching+ is relative to each pitch type, so teams that throw more low xRV pitch types have better xRVs than their Pitching+ would suggest. Teams like the
#Yankees
and
#Astros
beat their relative pitch quality by throwing fewer fastballs.
Very American.
@mason_mcrae
Interesting note here from
@drivelinekyle
, because it's something that might not seem intuitive. Here's what the data shows, with Fujinami highlighted:
Notice that Fujinami threw more strikes as his velocity sat 100-101.
Taking something off your fastball changes your mechanics to an unfamiliar pattern. Why would that lead to better control?
Operate at the same effort you train at to get the best results. Change focal points.
Hunter Gaddis has some good breaking stuff and a great changeup, as well as really solid command, but his fastball stuff quality is pretty poor. However, if he can maintain the command of the fastball he's had so far it can still be a good pitch.
Thanks for the shout out
@BenLindbergh
and
@megrowler
! The
#Dodgers
had the best plate discipline in baseball by a good margin, a big reason, as stated in the podcast, they led baseball in offensive production. Check out their great discussion of plate discipline/swinging less⬇️!
On our latest episode: Pitchers are calling their own pitches, hitters are seemingly swinging too much,
@FabianArdaya
previews the Dodgers,
@sahadevsharma
previews the Cubs, and we remember when the White Sox wore shorts.
One of my findings in this article is that although exit velocity is higher to the pull side, launch angle tends to decrease significantly. Take a look at the full article to see more!
Latest Down on the Farm article is out!
@Drew_Haugen
explores batted-ball spray angle and how it can inform a player's approach at the plate. Free and unlocked for everyone to read! 🔥⬇️
@BaseballBros
In the 870 possible games played since Corey Seager's first full season, he has played in 636. How can that be compared to the record holder for most consecutive games played?
Matt Brash's plot is a stark comparison, far more blue than deGrom. SL Stuff is obviously amazing but definitely not optimized for chases due to break differences. Hammers home the point made by
@Tieran711
,
@MaxwellResnick
,
@ydouright
,
@WillSugeStats
,
@eli2722
, and others.
Put together a little viz to show the effect of stuff on chases. It compares swing prob from the SL location only model to swing prob with location + deGrom SL stuff characteristics. Stuff induces ~8% more swings off the outside corner, but loses ~15% under the zone.
The
#Orioles
’ Gunnar Henderson has all the tools, but if he can't fix his GB problems, it's going to be difficult for him to truly be a star. He profiles similarly to Christian Yelich and Tommy Pham, two other hitters also held back by their launch angle distribution.
Here's the same viz, but with location quality. The
#Dodgers
and
#Rays
have consistently been among the league leaders the past three years, while the
#BlueJays
improved 19 spots from 2020 and the
#Rangers
dropped 26 spots.
A look into how pitching staffs have developed, in terms of stuff quality, from 2020 to 2022. The
#Yankees
have led the league in Stuff+ the past two seasons, while the
#RedSox
saw the biggest jump in the league from '20 to '22, and the
#Dbacks
saw the biggest drop-off.
I was curious so I put together a viz of the leaders in sum IVB in 2023. Last year, Spencer Strider's fastballs had a total of 2789 feet of ride, or 7.75 football fields. Tyler Rogers' rising slider stands out and leads all sliders despite being thrown far less than the others.
@CardinalsReek
Nice. I created a stat called LA+ (launch angle+) that was essentially a rescaled xwOBACON with launch angle as the only input to quantify exactly how valuable a hitter’s launch angle distribution was, and unsurprisingly Arraez led the league by far.
I used this same game info, which also contains the game umpires, to analyze different home plate umpire tendencies. Guys like Angel Hernandez and Vic Carapazza are trigger-happy on the edges of the zone, while guys like Dan Bellino and Edwin Moscoso are more conservative.
I figured out how to pull info on weather from the mlbstats api so I used that data to see which ballparks are the best to go take in a game at. Lots of California teams at the top.
*Note: The way I defined good weather doesn't include domes/roofs, which penalizes TB/MIA/HOU/etc.
I wanted to investigate this relationship with my swing decision metric, and although not a super strong relationship, there definitely is a connection between discipline and power production.
@DolphHauldhagen
@NotTheBobbyOrr
yes exactly, the heat map of the strike zone still shows barrels in the middle, so discipline has to be good. confounding to me, though, that I'm looking at 1900 player seasons in Statcast era with a .07 r2 between z-o swing and barrel rate.
Here's the same data, but broken out by pitch group. Secondary pitches from slow working pitcher have suffered the most, especially curveballs and changeups. This could be based on the amount of effort different pitch types require, but that's just a theory.
Good point here from Anthony, in terms of decrease in Stuff+, the bottom quartile of pitchers in tempo from 2022 have been hurt the most, very likely due to the new pitch clock. Interestingly, there is a non-negligible relationship between pitcher tempo and stuff quality.
The best single-game framing performance of 2021, by far, goes to
@HipHipJose5
on April 27th. Added over 13 more strikes per 100 takes than expected and in total was worth 1.29 runs in just 72 takes. 1/2
Recently been looking at strike probability using a GAM for framing, pitcher command, and a couple other things. Below is the changing called strike probability of pitch based on batter handedness. The only variables are horizontal and vertical plate coordinates.
This awesome thread inspired me to create a new stat, Wiggle Room+: the amount of area in inches a certain pitcher pitch type can be thrown to and expect good results (a negative xRV), compared to the average for that pitch type. In other words, the blue area on a heat map.
🧵
Ken Waldichuk is legit. He put up a 110 Stuff+ or higher on four pitches and has a legit whiff offering in the sweeping slider, comps include Sale SL, Berríos CU, Manoah SL, Ottavino SL, and E Phillips SL.
#Reds
reliever Alexis Díaz doesn't get talked about enough. He releases his fastball from the same height as his brother, Edwin Díaz, (one of his closest comps) which Stuff+ loves (159.5, 1.4 points higher than Edwin) and pairs it with a great slider (140 Stuff+).
The relationship between arm angle and inferred spin axis is, in general, stronger for four-seamers than sinkers, but cutter-like four-seamers such as Keller's, Steele's, Bradish's, and Suter's are big outliers.
Jump in Stuff+ ranking from 2020 to 2022 and jump in run prevention ranking moderately correlate, with Stuff+ jump vs. FIP jump showing a slightly stronger relationship than Stuff+ jump vs. xFIP jump.
There is a strong relationship between average estimated bat speed and average exit velocity (R =.85, R^2 =.73). However, certain players overperform/underperform expected EV due to EA, aka collision efficiency.
Big things could be on the horizon for the
#BlueJays
Yusei Kikuchi if he can just get his fastball command close to league average. Great 127 Stuff+, not so great 88.4 Location+. Finding a way to elevate and get out of the heart of the zone would be huge for him.
Here's a comparison of where Bundy throws his fastball vs. where he should. He locates up and away to lefties often but the model prefers the FF up and in, and he locates in-zone in 2k counts far more than he should, as the IVB leads to positive expected results above the zone.
Was interested when I looked into Cleavinger and saw that he had a poor SL Stuff+ from my models, but he really turned it on at the end of the season. Took his slider from more of a slurve to a true sweeper and saw all his pitches improve with a lower release point.
Top 5 RP in the new
@PitcherList
PLA QP% (Quality Pitch%, PLV >=5.5), min. 300 pitches:
1. Fairbanks (62.8%)👀
2. Clase (62.1%)
3. Munoz (57.6%)
4. GARRETT CLEAVINGER (56.5%)
5. Diaz (56.3%)
Cleavinger flashed 3 elite pitches in 2022, with a top-10 LH SL by PLV. Don't sleep.
The
#Royals
are going to be a ton of fun in 2023. Many of their rookies showed flashes last season and are primed to put it together next season.
@VPasquantino
has an especially impressive profile, look at all that red!
Excited to showcase my recently finished pitch quality models, especially Stuff+, on
@MSportingStudio
! I focused more on model results and what Stuff+ likes, so feel free to dm me with any questions on my process/methodology!
p(HR) is the probability a pitch is hit for a home run based on its stuff characteristics and location, along with count, batter hand, and pitcher hand. For example, the following pitch had the highest probability of being hit for a HR of any pitch last year at 7.8%:
Check it out and let me know what y’all think, in this article I introduce a few metrics for evaluating launch angle that are more insightful than current options like Avg. LA, Median LA, and even Sweet-Spot%.
What makes a launch angle good? Contextualizing batted-ball profiles by understanding the relationship between batter launch angle and exit velocity 🔥⚾️⬇️
@Drew_Haugen
p(HR) is valuable because, as it only looks at pitch properties, factors like batter skill, weather or wind, and ballpark dimensions are ignored. Because of this, pitcher xHR (expected home run, or the mean of p(HR)) rate was far more stable year over year than observed HR rate.
Corey Seager is going to turn it around for the
#Rangers
in 2023. Although not bad in 2022, he had an xwOBA 41 points lower than his real wOBA. Additionally, his 15 best comps had an average wOBA 20 points higher. Especially with shift restrictions, a 145 wRC+ isn't unreasonable.
On the flip side, this towering shot from Jean Segura had the lowest p(HR) of any pitch hit for a homer last year at 0.00384%, in part due to its swing probability of 28% and just 10% probability of making contact, given a swing:
Here's a comparison of where Bundy throws his fastball vs. where he should. He locates up and away to lefties often but the model prefers the FF up and in, and he locates in-zone in 2k counts far more than he should, as the IVB leads to positive expected results above the zone.
Dylan Bundy had a top 10 IVB on his four-seamer last year yet elevated below a league average amount.
I wonder would happen if he sat at the top of the zone more. If he could do that + find a way to get back to sitting 91, i think he could have a very good FF/SL combo
Shohei Ohtani is absolutely alien. He is one of only two pitchers in MLB with six pitches above average by my Stuff+ models (Darvish is other), with his "worst" pitch, his curveball, still an impressive 120. His new sinker is the 5th best SI by Stuff+ in the league already!