How Carlos Santana Has Turned Back the Clock for the Twins

Great article John. Very timely. Thank you. 

Nice job on the article! Honestly, what I've taken away from this is there's no good reason for Santana to be hitting so well, but there may not have been a good reason for Santana to have looked so poorly in 2020-2022.
 wOBA vs. xwOBA
2020 255 PA, .316 vs .372
2021 659 PA, .294 vs .335
2022 506 PA, .316 vs .352
3yrs 1420 PA .303 vs .348
--------
2023 .323 vs .308 <--- Looked cooked. 
--------
2024 .328 vs .323

There's every reason to believe Santana's solid production days were behind him. While 2020-2022 paints harshly unlucky results, when you see a pattern that long over that many plate appearances, it's so rarely a coincidence you just have to reset that to the new norm for a player.

xwOBA of .348 (2020-2022) would be good for about a wRC+ 120ish, instead he managed an actual wRC+ 91 over that span. Then, last year, his expected numbers tanked to an xwOBA of just .308 at age 37 while his actual numbers reversed some of what appeared to be "bad luck" in the previous years.

If you blend everything together, you get a player who is definitely in the twilight of their career, but Santana may be able to sustain close to this level of overall production through 2024. I think he'll regress a bit, yet, but he may be able to put up a slightly better than league average season at the plate.

Using OAA on it's own isn't a good practice, IMHO. It's the worst possible defensive metric to use as a standalone and Fangraphs should switch back away from it ASAP. It's wildly inconsistent, and it's more like a video game "what if" simulation rather than crediting the player with what they actually produced. By most measures, Santana is a good first baseman, but the defensive metrics are actually split on him. Combining them all together gives the view of Santana being a solid first baseman.
DRS = Excellent
UZR = Below Average
OAA = Excellent
RF/9 = Poor

On 7/22/2024 at 8:10 AM, Brock Beauchamp said:

I had no idea Santana was destroying lefties THAT MUCH.

Conversely he's part of the problem against right-handers.  This goes beyond him and applies to most of the guys facing righties from the left side of the plate.  Only because our RHB have fared actually pretty well against the right-handers has it not become a full-blown disaster.  Kepler, Julien and Santana are tasked with pulverizing RHP and they just aren't doing it.

20 minutes ago, Twins_Fan_in_NJ said:

Front office takes a lot of heat - some of it, rightfully so - but Santana at $5M is a complete steal and the Twins' season would look a lot different (in a bad way) without him.

Jose Miranda is better than Carlos Santana. Miranda would have directly replaced Santana's plate appearances. The implication Santana has provided some sort of major lift to the team is looking at his play in a vacuum.

1 hour ago, Schmoeman5 said:

In the past couple weeks you've used several players whose projected numbers don't match up to actual results. Are these abnormalities? Or do they happen more often? Just the last paragraph alone with 4 different ways of measuring how good he is defensively tells me that it's faulty at best. What were Miranda's projections for 2024? I doubt they were as sterling as his results have been. And based off that wouldn't he be expected to have a huge regression? 

It's sample size, luck, and defensive shift changes. xwOBA is like a more wide reaching stat than something like BABIP. Consider it ERA vs FIP, but a bit more advanced. Maybe like ERA vs SIERA.
2024 wOBA vs xwOBA 
Correa = .383 vs .362
Castro = .343 vs .331
Buxton = .354 vs .330
Miranda = .379 vs .345
Santana = .332 vs .326
Jeffers = .339 vs .322
Lewis = .430 vs .400
Kepler = .311 vs .308
Julien = .305 vs .298
Larnach = .330 vs .353
Lee = .297 vs .304
Wallner = .351 vs .313
Vazquez = .225 vs .249
Farmer = .256 vs .277

While you can see a single season variance, especially with this just being a little more than half season so far, seeing a trend where xwOBA diverges substantially from actual wOBA isn't common.

When it comes to defensive metrics, they're measuring totally different things.
OAA replaces the actual player with a fictional, hypothetically average fielder with average reflexes, average, acceleration, average speed and an average arm. Then it simulates what would have happened if the hypothetical fielder was asked to make the play vs. what actually happened, but positioning and instincts can make enormous differences in the outcome of a play.

DRS and UZR are pretty similar. They use IDENTICAL datasets. They section the field into zones, then assign positions to their expected zones. The metrics give bonus modifiers for making a difficult play or botching a routine play, but UZR limits the modifier while DRS does not. UZR also uses scorers to report when shifts are employed and it throws that data out. DRS separates batted ball data more aggressively, but it separates the data sets so much you wind up with super SSSS which might sit at the edges of statistically relevant. Basically, UZR is steadier, but it might blend a bit too much while DRS is less steady, but subject to additional error.

RF/9 says fielders of position x are involved in the making of outs y times. If a fielder is involved in more or less outs than league average, they're better or worse than league average. It's vulnerable to shifts like the Twins employed so frequently because it increases the rate at which some fielders are involved in plays.

3 hours ago, John Foley said:

Defensive measures are noisy and take a long time to become reliable. One detail point of clarification is that Statcast measures from where the player is positioned. If the pre-pitch positioning makes the play easier for the player, they get less credit for successfully making it. If that positioning makes it harder, they get more credit if they convert it.

The other defensive measurement systems don't handle pre-pitch positioning as precisely as Statcast. 

Sorry I didn't make that clear enough.

The fact a fielder who is very good at positioning gets dinged for it vs. a player who positions terribly, but makes up for some of it with athleticism gets rewarded is the greatest flaw in OAA. It's why I feel OAA is the worst metric to use on its own; it doesn't tell you what actually happened, and whether or not a fielder's play results in more real outs than average. It's probably also a big reason why OAA is the least reliable metric with the greatest swings from year to year.

46 minutes ago, jimbo92107 said:

Miranda may be a better hitter, but on balance Santana is a better 1st baseman. Better glove for sure, knows exactly what to do on every play, so he's half a step ahead of Miranda defensively. At the plate, Santana may not bring quite the power Miranda does, but Santana is a switch hitter, and he has one ability Miranda can't match: Availability. Having a steady, capable 1st baseman helps the whole infield. 

Not. Even. Close. The small advantage (if any advantage at all should Miranda see regular playing time at 1B) in WAR for Santana's defense over Miranda's is not worth the better bat Miranda brings. Jose Miranda has all of 50 innings played at 1B the last 2 years.

Santana is a switch hitter, but the only side he can actually hit well is from the right side against lefties. He's been below league average against righties in recent years. Miranda hits both righties and lefties.