Stephen Take The Wheel

Brodie Van Wagenen has shown he will not hesitate to shake things up.

Is that a good or a bad thing? Maybe it’s both?

In an effort to make some sense out of our new general manager’s tactics I’ve enlisted some help.

Enter Stephen Guilbert. You may know him as Stephen Josiah on Twitter. If you don’t already do so, follow him here. He’s an incredible baseball mind so this was a lot of fun, and very enlightening as well. He has agreed to answer a couple of my questions, so without further ado…

Shortly after Brodie Van Wagenen was named the Mets GM, he hired Adam Guttridge as the assistant GM for systematic development and Allard Baird as their Vice President of player development and scouting. Do you worry that the Mets have made these hires simply as a way to cut costs? Is this Moneyball in a big market or have they just fully embraced analytics?

I do not see any way Brodie made these hires simply to cut costs. I think he hired these two men because he believes in their track record and what they bring to the table. I would like to discuss whether or not that is warranted—at least when it comes to Adam Guttridge. But let me first get to your second question about “Moneyball in a big market”. Every team employs what Moneyball revolutionized— using math to find value where others do not.

Every major league team has an analytics department. However, the size, investment, and brainpower in those departments varies wildly from team to team. On the high end, the Yankees have the most people while the Dodgers and Astros invest the most money. The Mets are near the very bottom in both size and investment.

Is the relative apathy towards analytics a way for the Wilpons to save money? While hiring fewer people does mean fewer dollars in salary and benefits, it makes up (or would make up) such a small percentage of the total operating expenses of a professional sports team that I do not think budget restraints hold the Mets back. Rather, I believe it is overall ideology that does. It’s no mystery that the Wilpons have their hand in just about every facet of the team. It’s also no mystery that the Mets’ owners do not understand, nor have the stomach for, advanced analytics. I see that as the primary problem and until Jeff Wilpon learns a bit more math or buys into what successful teams like the Dodgers, Astros, Yankees and Rays have been doing for years, Mets fans are stuck with a big market team with a mid-market payroll and an analytics team that is small, underfunded and punching way outside its weight class.

But let’s get back to Adam Guttridge because even a small and underfunded analytics team can be successful with the right people, the right mindset, and the right implementation. Guttridge has something he touts as the “Automated Prospect Model” (APM) which he describes, in essence, as a projection system designed to predict future performance by combining projection-analytics (performance, aging curves, age-relative-to-league and other inputs he has not publicly disclosed) with scouting. Speaking at MIT’S Sloan Sports Conference, Guttridge talked about his model and its effort to remove emotion from decision-making and standardize players across leagues:

“We have to put a quantitative value on what we care about, how much we care about it and why we care about it. So what we’ve done here…what we’ve achieved here…is to distill the scouting process to the parts we feel are wise. The parts we feel we can defend empirically as well as theoretically. We’ve modeled that wisdom, and, since we’re a robot (and we’re not a human) we can apply that wisdom without any emotional bias, with a perfect memory, and with an omnipotent viewpoint— seeing every pitch, every game, (and) every inning, and apply this judgment equitably with the same brush to all players in the purview of (professional) affiliated baseball”.

http://www.sloansportsconference.com/content/the-automated-prospect-model/

Start at the 10-minute mark for this quotation and context. I want to break this down a bit. First of all, I think the idea is great. As a Michael Lewis fan (the guy who wrote Moneyball. Yes, it was a book before it was a great baseball film), this idea of questioning our decision making and removing emotion and bias from data collection and decision was explored in his book The Undoing Project (I highly recommend it. Chapter 1 covers this topic in the context of basketball and, effectively, basketball’s “Moneyball moment”). Second, putting value on what we care about based on what the data tells us we should care about also seems like a good idea. Lastly, this sounds more like a holistic approach and I largely like holistic approaches in front offices—that is, combining what we can gain from scouting along with what we can see in numbers. However, I have concerns over this model’s effectiveness both from what Guttridge cannot answer as well as systematically flawed problems from the get-go.

First, defense seems to be left out of the equation. We struggle to properly and accurately value defense as it is and it’s even harder to evaluate that with young players and project it years from now. However, we still have to include it in the equation. How does the Automated Prospect Model account for it? Unless Guttridge is being coy or to protect his intellectual property (he fails to answer a question about this very point at the above video’s end)…it doesn’t. That’s problem #1. Problem #2 is a concern I was hoping would be squashed after seeing this regime’s f irst few moves. But it hasn’t. In fact, it’s been exacerbated. Since this model requires data in order to work, what do you with players without any data or very little data? Players in the DSL or recently-signed out of Latin America, recent domestic draftees, young players who were injured early in their career don’t have much “affiliated baseball” data, what do you do with them? How do you project those players? How do you value them? If you don’t have the data integral to the model, the algorithms either take what it can and has a massively high delta, or you put a big ‘ol question mark around it and use something else. I knew this about APM going into this off-season once Guttridge’s hiring was announced. What I didn’t know was whether or not he would be conservative with that unknown (that high delta) or not.

We do now because we’ve had four trades already so let’s take a look at those and see if we can figure this question out:- Traded Jared Kelenic in a package for Canó and Díaz. This trade has been debated and argued six ways til Tuesday and I won’t make a case for the trade being “good” or “bad” but I will point out that Van Wagenen was quick to trade arguably the Mets best prospect—and certainly the player with the highest raw upside in the system— very early in the winter for an aging infielder coming off of a suspension and a relief pitcher. While the value might turn out to be fine in the long run (I’m doubtful but we’ll see) the fact that Brodie and Guttridge didn’t take some time to see what Kelenic&Co. could have gotten from other teams (Realmuto? Kluber?), leads me to believe he’s not being conservative with those young players.

But that’s one trade. What happened next? Traded Adam Hill, Bobby Wahl and Felix Valerio for Keon Broxton—a human highlight reel CF and a fine major league bench piece but certainly not a “must have” cog and someone who is slightly redundant with a similar…ish player in Juan Lagares already on the roster. But look at the return for the Brewers here: Hill, a 2018 draftee and Valerio, an intriguing OBP-promising infielder who spent all of 2018 as a 17-year-old in the Dominican Summer Leagues. Both players with practically no data that the APM could use to project.

Then he traded Kevin Plawecki for Walker Lockett and Travis Haggerty. Two players with a lot of data in the minor leagues but highly limited upside. I can defend this because of their MLB-readiness but I also do not see this as a great return for a steady MLB catcher (those are just so rare in baseball.)

And now for the worst offense: J.D. Davis and Cody Bohanek for Luis Santana, Ross Adolph and Scott Manea. Ross Adolph, yet another 2018 draftee, had a wonderful debut for Brooklyn but, as you might be able to see from the theme here, also has very little data because he was just drafted. Luis Santana, same story but from international origins. Losing Santana hurts. This is a confident, gamer 2B who put up .348/.446/.471 in the Appalachian Leagues as he was turning just 19 years old. But again, only 204 at bats in stateside baseball. Despite the success, the positive reviews from rivals and experts, he was still cannon fodder in a battle to accrue bench pieces and quad-A players. Kiley McDaniel in a recent chat stated that the Mets were the only team to not scout rookie leagues in 2018. So, in recap, the Mets do not scout the lower levels of the minor leagues, they haphazardly trade away rookie league players for pennies on the dollar, and they trade major league assets for low-ceiling, low-reward players.

Fantastic. Needless to say, my fears about his system have been justified and the question over whether or not Guttridge&Co. would be conservative with the data-light young players in the system was a resounding “no”. The counterargument, if there is one, is that the Mets are “going for it” (as much as a Wilpon-owned team can) and you need MLB-caliber players for when things go wrong like injuries and underperformance. You can argue that J.D. Davis and Keon Broxton mean more for the Mets window now than even an elite talent like Jared Kelenic in three or four years when the window might be closed. I understand that point. However, let me share something with you from the very man who is parting with talented youngsters like they’re quarters at an arcade:

“General managers, unlike their counterparts in industry, never really used the balance sheet, primarily because we’ve never viewed major league players, minor league players, draft picks, and playoff value within a single plane of tangible value (both present and future). Those who learn to do so effectively will gain a massive leg up in the information wars.”

This was Guttridge from back in 2009 (https://tht.fangraphs.com/the-guttridgewang-trade-model/ ) in a piece in which he argues for value-based models across the board in front office decision making. (Yes, I did dive this deep into the rabbit hole in case you were wondering).

So, I’m confused. Does this “omnipotent” prospect model (I really hope he meant “omniscient”, by the way, otherwise my goodness he’s arrogant) contradict valuebased decision processing? Or does that model somehow tell you that trading players like Santana, Adolph and Manea for fringe-MLBers in a competition window will “pay off”? And, if so, how is it telling you that? Because it sure as heck ain’t based off of the Automated Prospect Model.

What offseason deal could you see paying the most dividends?

Right now for me it’s signing Wilson Ramos. We know Ramos doesn’t play the best defense but not many can do what he can at the plate and this is a team that desperately needs offense. As far as trades go, I could only make a case for the Díaz/Canó trade being a good one IF and only if the Mets get a ring out of this the next few years OR Jarred Kelenic is Bubba Starling 2.0. Or if by some twist of wonderfully magical fate, J.D. Davis becomes Justin Turner and Keon Broxton becomes Carlos Gomez. I’m still waiting for the savvy move. The one that has little to no risk but huge upside. Trading for Byron Buxton has long been on my list of a move a smart team should make. Minnesota can still get good value for him and he’s a “needs a change of scenery” sort of player and a stat cast god if there ever were one.

Internally, the Mets have some intriguing infield options. Dilson Herrera, TJ Rivera, Luis Guillorme and Gavin Cecchini. Could we get a bold prediction from you from those group of players?

This is a fun one and you asked for bold so let’s go bold: If Dilson Herrera gets more than 200 at bats, he will out-WAR Jeff McNeil.

I’ll give you another one: Gavin Cecchini does not play 100 total career games for the New York Mets.

Question 4. Kinda related to question 2. There has been a flurry of trades where we sent our lower level prospects to teams in exchange for depth pieces. Is this the new regime evaluating the minor league system they have inherited and not liking what they see?

As I stated above I think this is an inherent flaw in the APM and it’s probably some level of “Well these aren’t our guys so we’re going to undervalue them” as well. I can’t help but think of the APM as a glorified KATOH system. And to be harsh here for a second, I don’t even use KATOH anymore for fantasy baseball. As in, offense only, fake, fun fantasy baseball. If that gives you a sense of how I view the APM in light of all of these trades.

Finally. The Mets seem to be fixated on ISO. Is there a place for Jeff McNeil on this team?

Of all the things you could be focused on, you could do worse than ISO but I’m not sure that’s what the team is doing. If they are, it sure beats batting average. Doesn’t this sort of remind you of when Sandy Alderson brought out that statistic about teams who hit a home run win x% of their games and just tried to load up on sluggers while chronically sacrificing defense on a team built around pitching?

I hope that’s not the case now and I do hope that the internal metrics—data we’re not privy to that comes from the MLB-sourced trackman and Statcast data—is informing them on these decisions and Brodie is acting on them from a position of knowledge and not ignorance.

When it comes to McNeil, I have my reservations. I have not been too public about this because fans absolutely adore him and I acknowledge that I might be simply wrong on him but I am very wary about two things:-Small sample size-The ‘hidden’ metrics within that small sample size.

Allow me to explain. Jeff McNeil had a wonderful debut but it was only 225 at bats worth of data and I just never put too much weight in that small a sample size. But, let’s say we do want to draw a conclusion and the best data to use is simply what we have. And what we have is really good, right? 140 OPS+, great battingaverage, good walk rate, low strikeout rate…definitely promising, yes? Well, partly yes and partly no. On one hand, something like O-swing and Z-swing rates and contact% stabilizes rather quickly and McNeil’s is hella good. That excites me. However, he also doesn’t hit the ball particularly hard. Of 332 players with 150 batted balls in 2018, McNeil ranks 290th by average exit velocity. The only player on the Mets worse last year was Jose Reyes and that range of exit velocity is right around the speed players (i.e. batters who use bunts and slow ground balls to leg out hits. McNeil is not that player).

When you contrast that low exit velocity against his batting average on balls in play, there’s contradiction. McNeil’s .359 BABIP tied for 16th in the MLB last year amongst players with 225 PAs or more, tied with Paul Goldschmidt and around the likes of Nick Castellanos, Freddie Freeman, Mookie Betts, Ronald Acuña, and others who hit the ball much much harder than McNeil did in his debut. In fact, only one player with as many batted balls and PAs as Jeff McNeil had a higher BABIP and a lower average exit velocity and that was now-Mariner Mallex Smith—one of the fastest players in the game. The reason I might be wrong about McNeil is that he does an extraordinary job putting the bat on the ball. He runs well. He plays defense well. He’s versatile. And perhaps there’s more noise than signal in his batted ball profile that’s distracting me from the fact that he had one hell of a year last year. However, I’ll end with this: The statement I most agree with Brodie (that he has at least acted on) is minimizing “what ifs”.

Jeff McNeil was a big question mark for me this year and by addressing second base, Brodie has just about eliminated that “if”, while allowing the framework for McNeil to get at least 100 games next year at multiple positions to prove his offensive debut wasn’t a fluke.

Thanks for having me as a guest today. This was fun.

Thank you sir!!!

 

Leave a Reply

Your email address will not be published. Required fields are marked *