The 2021 trade deadline was bonkers. It was also a heck of a lot of fun. 

But more importantly, how did our model do this year? Starting with the Joc Pederson trade, and ending at the deadline, we counted 55 deals in total. Of those, 4 were miscoded due to human error or late information, so we can’t count them (they would have been accepted by the model if not for those issues). So with those excepted, there were 51 measurable deals, of which 46 were accepted by our model, for a success rate of 90.2%. 

Overall, since August 2019, we’ve tracked 250 trades, of which 238 were accepted by our model, for a success rate of 95.2% and an average margin of error of 1.7.

Biggest wins for the model this deadline?

• We were very close or dead-on on these fair deals: Scherzer/Turner, Bryant, Baez, Gibson/Kennedy, Rizzo, Pederson, Schwarber, Escobar, Tyler Anderson (twice!), Gomes/Harrison, Hill, Hernandez, Rosario, Happ, Marisnick, Galvis, and most of the relievers and minor swaps.
• We were in the ballpark on these reasonable overpays: Gallo, Graveman, Duvall, Norris, Soler, Heaney, Lester, Hand
• We missed on these: Kimbrel, Castillo, Frazier, Marte, Rodriguez

 

*The four trades we are not counting because they were miscoded in the model due to human error or late information were: Berrios, Luplow, Straw, Cruz

Observation #1: It was a seller’s market for good starting pitching

This was no surprise, as we had previously observed that the starting pitching market was in shambles. Almost every contender had holes to fill in its rotation, and there just weren’t a lot of good available options on the trade market. The supply of quality starters was low, and demand was high. So it was no surprise that Jose Berrios went for an overpay.

But note that Kyle Gibson did not. Tyler Anderson did not. The non-elite starters went for fair prices, as they should. And curiously, Max Scherzer did not, although it’s hard to decouple him from Trea Turner -- but together, they yielded what we thought was a fair return. In his case, though, we had already priced in a market overpay, which is probably why. In Berrios’ case, since he’s a cut below Scherzer in reputation, we did not -- but the market did it for us.

Observation #2: The market hates 2Bs

This is a trend we’ve been seeing for a while now, and with each data point, it becomes clearer and clearer that 2Bs are practically being given away. When we first noticed this trend, we conservatively applied a position adjustment of 10% below market value. Then it kept happening, and we kept having to increase that haircut. We missed on the Adam Frazier deal, we think, because we had only applied a 20% positional discount to his value. The Pirates traded him for 50% below field value. 

This was also true of Cesar Hernandez (although we had already docked him at 50% below, based on his offseason contract, which is why we hit on that one). Kolten Wong’s offseason discount was also about 50% below market. 

This is probably why the Royals could not get an offer they liked for Whit Merrifield -- and GM Dayton Moore said as much when they pulled him off the table. It would also explain the inclusion of Nick Madrigal in the Kimbrel trade, at least to a degree. 

That is, we noticed that veteran 2Bs on higher salaries were getting undercut, but we haven’t noticed it enough with pre-arb players making league-minimum salaries, because we figured that teams were thinking they could replace a higher-priced 2B with a lower-priced one easily. Therefore, the lower-priced ones should not be discounted, right? Or should they, because eventually they’ll get expensive, too? We’ll keep investigating this question.

Observation #3: Our baseline assumption about QO draft picks was correct

We’ve found that teams with rental players who might be offered a QO at the end of the year are valuing the draft pick associated with that QO at around $9M. That means, in theory, that if the player can return more than that in a deadline trade, they should trade him. If not, they should hold onto him and take the draft pick.

You’ll notice that Kris Bryant was traded for roughly $11M in value. Trevor Story was not traded, we think, because the offers did not exceed $9M. There are likely reasons for this, as Story was not having a great year, and there was virtually no SS market, which means that teams would have to play him out of position (such as 2B, which is lesser-valued), which in turn lowers what they’d want to give up for him.

Observation #4: The position-player market was lukewarm

Big names like Gallo, Bryant, and Baez went for packages that were just… okay. Anthony Rizzo’s return was a touch high, but still fair. The inclusion of Trea Turner in the Scherzer deal may have actually watered it down a little, because it came in a touch light according to our model (the two minor prospects after Ruiz and Gray weren’t worth much). The only player who went for an overpay was Starling Marte, but even that’s an arguable case, as Jesus Luzardo has been struggling badly of late, and his stock may have dropped even further than where we had it.

Observation #5: The market loves relievers with the hot hand

Valuing relievers is the trickiest problem we have with our model. How much do you weight recent performance vs. track record? Since we launched in 2019, we’ve been tweaking the model so that it weights recency more and more heavily, as that’s what has correlated with the market. We track this not only with trades, but with DFAs as well, as we notice time and time again that teams have short leashes with underperforming bullpen arms, even if they have a strong track record. It’s volatile, folks. But you knew that.

So what to make of the extreme overpay for Craig Kimbrel? It means the market did not care about his struggles in 2019 and 2020. It only cares about the now. Most contenders were also facing budget constraints, either imposed by their owners (A’s, Rays) or by luxury-tax implications (Yankees, Red Sox, Astros). So we figured that Kimbrel’s enormous salary would also be a hindrance, which is why his surplus number seemed low. But clearly the White Sox are not worried about him turning back into a pumpkin, or about the money. They just want to win a World Series, so they’re overpaying for him like the Cubs did with Aroldis Chapman in 2016.

On the flip side, we saw a big underpay for one of the most successful relievers in baseball, Diego Castillo, whose value was high because he’s been elite and because he has three more years of control after this one. If Kimbrel yielded a huge return, why didn’t Castillo? Did the Rays even shop him?

We have noticed that Castillo’s velocity has been dropping over the past three years, so perhaps that was the red flag that he was starting a decline phase, and the Rays wanted to sell high. Perhaps they’re higher on Austin Shenton than the public evaluators we follow are. This one will likely take a while to measure. The Rays, of course, excel at playing three-dimensional chess in the trading game, so don’t write this one off yet.

Most of the other (non-elite) relievers were traded for fair returns, though some were on the higher end of fair value, and most of the ones who moved were the hot hands, who have been performing well of late. We had recently downgraded Richard Rodriguez a touch due to a performance decline post-sticky stuff, but apparently that scared the market even more, as he went for less than expected.

It’s possible that the Kimbrel and Castillo cases were just outliers on each end of a volatile market. It’s generally a bell curve, so in this case, you can picture Castillo at the left, Kimbrel at the right, and most everyone else in the middle, albeit with a skew to the higher-end right side. That would suggest the model is fine in this regard. We’ll take a closer look at the way we value relievers, though, and see if we find any consistent patterns that would warrant a tweak.

Summary

So:

• *The model did well (apparently, at least we think so) -- it got most of the big ones right, and almost all of the little ones
• *It was a seller’s market for good starting pitching, as expected
• *2Bs are clearly out of favor, and we’ll be adjusting down their values even more
• *Story wasn’t moved because the offers didn’t beat the QO draft-pick minimum
• *Hot-hand relievers are always in demand at the deadline (no surprise); special ones generate overpays; but watch for red flags that might scare off others.

We’re exhausted, and we’ll be laying low for a bit in August. Enjoy the rest of your summer, folks.