SABR Day 3: Evolution versus revolution

PHOENIX -- As the third and last day of SABR’s third Analytics Conference wound down, I wanted to focus on the question that I came into it with, asking whether or not sabermetrics today is more revolutionary or evolutionary. Moneyball’s pop culture status has been safely achieved with Brad Pitt mugging for an analysis revolution already won. Have all the great discoveries already been made with the proliferation of sites like Baseball Prospectus and FanGraphs, and with every organization employing their own teams of analysts?

This seemed like an especially relevant question to ask after talking with Voros McCracken after the first night’s proceedings. McCracken is the inventor of DIPS theory, the most important discovery in sabermetrics in the last 25 years -- and perhaps also the last great discovery. Making an observation of that magnitude, something that changes the way all of us see the game, including those of us who have been in the analysis community for decades, is extraordinarily rare. If you’re familiar with Thomas Kuhn’s "Structure of Scientific Revolutions", it seems like we’ve been more engaged in doing the basic science and communicating its lessons to a growing audience since Voros gave us the game’s last big paradigm shift.

Take valuation models: There's nothing really new about the idea of assigning dollar values and arguing at length about how to price a player's performance and his expected performance; it’s cool, but it all basically spins from the work the late Doug Pappas was doing in the 1990s. Value over replacement and WAR? I thank Pete Palmer for getting us started on that track more than 30 years ago.

What about this new research on catcher framing? The work delivered by Ben Lindbergh, Baseball Prospectus or new Indians hire Max Marchi is radically changing our understanding of the value of catchers as receivers. It’s new and compelling, but you could argue that the original observation goes at least as far back as Earl Weaver choosing Rick Dempsey for his skill as a receiver over a better-hitting backstop brick like Earl Williams almost 40 years ago, and then subsequently writing about it in "Weaver on Strategy."

So I turned to several sabermetric thought leaders in attendance at the conference to ask them, are the new analytics just documenting previously observed phenomena? Has sabermetrics become more evolutionary than revolutionary?

Rob Neyer, Fox: "Wow, it’s a fantastic question. I just recently started re-reading Kuhn. I think Kuhn would say we’re still in the evolutionary stage, that we are waiting for something else to be revolutionary. I think that we’re just sort of going step by step. Even the new data that we’ve got coming is just more steps in that process. We can’t know what the revolution will be until someone does it. I think there will be something, but no, we are not revolutionary. As exciting as the new data’s going to be, there’s going to be something even more wildly exciting after that, but something which a lot of us will reject. When I think back on Voros McCracken and DIPS, I was impressed when Bill James not only didn’t reject it but embraced Voros’ findings, and said, ‘I should have seen this a long time ago.’ Bill hasn’t always been immediately willing to embrace new information, but he was all over that one, and I admire him for that. But you’re right, it’s been a while. In fact we’re still grappling with Voros, but you can’t talk about pitchers without talking about what Voros discovered. Maybe the new data will create an environment or inspire someone, but right now I think we’re just talking about using the new data to further the things we already understand."

Keith Woolner, Director of Baseball Analytics, Cleveland Indians: "I think there are elements of both. I think that there’s a constant refinement of something simpler, like offensive models -- those have gotten better as our data has gotten better. Every once in a while, you’ll get something that really creates a new area -- especially at this conference, that’s been catcher pitch framing. That was something that we had the data for a while, but it was not publicly released, discovered or discussed. That is, I think, the revolutionary step, just because it was not an area where there had been a lot of focus before. I think that, as we get new data sets, the catcher framing and all of the PitchF/X analysis at the event came about because there is a new source of data, that told us things about what was going on in the game that we didn’t have before. That tends to be the impetus for creation of the more evolutionary things. And because we’ll have this new tracking system from MLBAM, that could be the next source of some evolutionary advances in sabermetrics. It’s less about the technology, and more about the information content.”

Cory Schwartz, VP of Stats, MLB.com: "I would still qualify it as evolutionary, but the pace of the evolution has increased considerably over the last 5-7 years, obviously with the influx of PitchF/X data, but PitchF/X data has enabled research into other areas using existing datasets. I think once we are able to roll out the complete field-tracking system and start to introduce some of that data into public space to whatever extent it might be, I think that will further increase the pace of evolution and perhaps bring about what we would consider revolutionary turning points.”

Sean Forman, Baseball-Reference.com: "I would agree it’s evolutionary. I think it’s revolutionary in the sense of the amount of data that we’re able to collect now. The thing is, you don’t see revolutions coming, so I’m sure there’s going to be something that comes along and blows us away, probably in the next five years. But right now I think batting is a fairly solved problem, I feel like with pitching, we know that we’re not going to know a lot about pitching and that it’s inconsistent, and obviously with fielding we’re expanding our knowledge there. I guess I would lean to the evolutionary side. Pitch framing could be the next big discovery, but I’m hesitant to put much on something that’s predicated on fooling an umpire, and whether that’s a long-lasting effect that we’ll actually see. It’s true, in terms of paradigm shifts, I feel like Voros’ discovery is one of the last ones we’ve seen."

Vince Gennaro, President of SABR: “I think we’re operating on two planes. Some of the concepts that have been around for a while, we’re continuing to evolve them, which is healthy. At the same time, I think technology is helping us revolutionize some other areas. I think when we see what MLBAM unveiled a couple weeks ago, you’re going to see dramatic changes in the way we measure defense. I think there are revolutionary aspects -- largely technology-driven or data availability-driven -- which are starting to tap into our things we’ve always had on our ‘do’ list, but then the good news is that we’re continuing to evolve the tried and true sabermetric concepts.”

Andy Andres, sabermetrics teacher at Boston University: "A really hard question. Certainly, everything has been evolving throughout time, going back to Harry Chadwick on up. But in the past era, created near Moneyball, you’ve seen real revolutions in technology (PitchF/X), you’ve seen real revolutions in defense (the BIS database), and analytically this idea of DIPS and FIP was a big deal, changing how we thought about pitching. If you define those as revolutions, we’ve had lots of revolutions. Now we’re bumping up to a period where there’s not a whole lot new besides catcher framing -- potentially another revolution -- so if I define revolutions as real, significant changes in analytics and our understanding of the game, we will continue to have these revolutions. Some people might say this is just the evolving nature of science, but however you want to define it, we’re never going to stop learning something about the game.”

Kevin Tennenbaum, junior at Middlebury College (and presenter on a panel about Bayesian forecasting): "I think it’s a little bit of both, but we’ve reached an evolutionary stage where we’re kind of using basic math or economics to find the value of different players and different strategies within the game, and we’ve kind of plateaued on that. We’re getting more and more people that are in the game who can really work with more advanced mathematics. That’s where the revolution will come, where we’re looking at things like Bayesian statistics and more advanced decision sciences, machine learning, computer-science algorithms that allow us to analyze these new large datasets.”

Dan Fox, director of baseball systems development, Pittsburgh Pirates: "Generally speaking, as player tracking from MLBAM comes online, the insights that we gain from that might be revolutionary, because any new dataset is going to bring with it stuff that we haven’t thought of yet. An example like what people have done with pitch framing and PitchF/X years after it was released is probably an indication that there will be things there that we won’t know for several years, but people eventually figure out that are much different than what we think today. So, player tracking? Probably revolutionary. Other stuff, up until then? Probably evolutionary, other than with teams’ willingness to invest more in their analytics departments, the teams are probably going to continue to move much faster, so there may be insights when you combine the medical, scouting and performance data that they have internally, that we just haven’t had previously. Who knows? There’s probably insights that some teams have had that we haven’t had that we’d think were revolutionary."

Which, going forward, is perhaps the fundamental challenge. From its inception, going back to the founding of SABR in the '70s and beyond, sabermetrics has relied on open-source information that has engendered and perpetuated its dynamic conversation about the game. The extent to which a generation of researchers have profited from Major League Baseball’s willingness to share that information through PitchF/X has extended that tradition. If all of us will continue to profit from the insight of sabermetricians pushing the boundaries of their science beyond evolution and into new revolutions, much will depend on that continued free sharing of information. Sabermetrics’ legacy of victory in the marketplace of ideas demands nothing less.

Christina Kahrl writes about MLB for ESPN. You can follow her on Twitter.