Tracking real-life trades, or: BTV vs. A.J.
In the interest of transparency, it’s time for an update on how the BTV model is doing this offseason relative to real-life trades — with one caveat: A.J. Preller.
Our model assumes most GMs will act rationally, and seek to get the best overall value possible in a deal, just as in most other walks of life, which means most deals are pretty close to fair. But we also know that’s not always the case. GMs have different motivations depending on where they are in the win cycle, and the most notable example this offseason is A.J. Preller of the Padres.
The San Diego Union-Tribune reported in late September that Padres Executive Chairman Ron Fowler has mandated that the time has come for the Padres to have a winning season in 2020 — or else “heads will roll.” The implication there is that Preller will lose his job if he doesn’t produce. He’s clearly gotten the message.
So far, every trade Preller has made this offseason has been imbalanced in favor of the other team, as you may have noticed. This is not a coincidence. He’s in win-now mode, and every other GM knows it. He cares far more about upgrading his MLB roster than winning a trade.
And because he has so much prospect capital to use, he’s spending it like a drunken sailor. Trade a top infield prospect for a 4th outfielder? Here you go. How about two prospects (one a legit hitter) for a guy who was about to be non-tendered? No problem. Trade a solid outfielder with four years of control for a pricier one with only two years of control? Sure — AND have one of my top prospects to boot.
He’s a little nuts right now.
And according to the MLB Executive Burner Twitter account (which appears to be an anonymous front office source), “Teams w/ trade assets hoping #Padres GM AJ Preller calls or will place outbound call. Everyone around game knows Preller’s seat is hot & willing mortgage future assets of leagues top 3 farm system to win now. Told “He’s not consolidating prospect base, he’s giving away players.”
So we think it’s fair to show our numbers two ways: with A.J. and without AJ.
With A.J. (all trades from late October through Dec. 6, 2019):
Percentage of all trades accepted by our model: 76.4%
Average variance: +/- 5.4
Percentage of all trades except Padres accepted by our model: 86.7%
Average variance: +/- 1.3
So we’re doing okay with A.J. We’re doing much better without him. We know he’s going to do a few more spendy trades, so we’re just gonna roll with it.
A few other trade observations:
Narvaez: We consider this a strong win, as his BTV value ($5.1M) closely matched his return (estimated at $4.1M, but variable due to the draft pick). Lots of baseball writers were surprised by the low return, but we weren’t.
Villar: Similar story. We had his value at -0.1, and he returned a non-prospect (estimated at 0.1). There was never any surplus there, and the preponderance of cheaper 2Bs on the market suggested no one was going to pay much for him.
Marisnick: Again, no surplus. We had him at 0, and he returned two non-prospects (we added both to the database at 0.1 each, but that may be generous).
In addition, the model feels like it’s working well in other types of transactions: non-tenders and free agent deals:
Our non-tender predictions did well, as you can see: we predicted quite a few big names would be non-tendered, and we were right.
And so far, most of the free agent deals are tracking well relative to the model’s fair value lines (we publish each player’s trade value after each deal is announced, and it’s usually close to zero, which means the salary was fair; however, we can’t take credit for these since we don’t publish the predictions in advance, as we’re not in the free agent salary prediction business). Examples include Zack Wheeler, Travis d’Arnaud, and Jake Odorizzi.
All told, so far so good. Except for A.J.