Baseball’s Ever-Expiring Secrets

Last week, imprisoned former St. Louis scouting director Chris Correa responded to MLB’s ruling in the Cardinals hacking case with a tweeted statement whose 14 sentences amounted to: “The Astros started it.” Within, Correa doubled down on his prior claim that Houston hacked first, accessing Cardinals data that aided the Astros in replicating “key algorithms and decisions tools” related to player evaluation. Correa also asserted that much of the front office was in on the heist, including current Astros GM and longtime Cardinals executive Jeff Luhnow.

Correa, of course, has Ryan Lochte–level credibility. Now permabanned from baseball, he was already serving a 46-month sentence for five criminal counts of unauthorized access, to which he pled guilty last January. Commissioner Rob Manfred immediately refuted some portions of the disgraced Correa’s most recent claims, and no evidence exists to substantiate Correa’s assertion that the Astros acted improperly.

But Correa’s case, and his Twitter protestations, still raise a couple of questions that aren’t easy to answer from outside the front-office fraternity. First: Hacking aside, what would constitute acting improperly, when someone with knowledge of one team’s intellectual property migrates to another team? And second: What can teams do to preserve their IP when they’re liable to lose their most promising personnel to direct competitors? Clearly, Correa transgressed by using Astros employees’ passwords to infiltrate the other team’s reports on players, notes on trade talks, and more. His repeated intrusions broke both the bounds of convention and federal law. But Luhnow and sabermetrician Sig Mejdal, who followed Luhnow from St. Louis to Houston, wouldn’t have had to hack to access at least some St. Louis secrets; they brought a combined 15 years of Cardinals experience with them when they joined the Astros. The secrets were inside their heads.

To find out how teams try to protect their IP, I emailed dozens of baseball insiders, ranging in expertise from scouts to statisticians and in hierarchy from consultants to GMs. Their responses reveal the wavery, unwritten rules of relocating, but they also show something that might seem surprising, given how tight-lipped teams tend to be in public and how strenuously they seek out innovation: No matter how much teams emphasize secrecy, their secrets don’t last long.

Based on the sources I surveyed, Correa has few fans in front offices. Every respondent who offered an opinion on the scandal agreed that the Cardinals got off light with a $2 million fine and two forfeited draft picks after the first round. But as one statistical consultant to multiple teams observes, the penalty, however lenient, formalizes what we’ve long assumed to be true about baseball’s front offices.

“What this does clearly imply, even to those of us who feel that the Cardinals penalties were light, is that teams have individual pieces of IP worth, at a minimum, millions and millions of present-day dollars,” the consultant says. “Teams have [player] assets, like a cost-controlled no. 4 starter, a well-extended ace like Chris Archer, or a mid-tier prospect like Jorge Mateo. They also own other assets, which in the case of some teams have great value; the mechanisms/systems [that] identify this talent before their competitors do. They’re no less material of an asset than the player.”

The people privy to the latter type of asset are mostly not making millions of dollars, and they’re often on the move, thanks to teams’ widely observed policy of letting other clubs interview their employees for positions that would represent promotions. This fall, the Twins raided a division rival to hire Indians assistant GM Derek Falvey, becoming the latest team to install a former Cleveland executive as its baseball operations head honcho. Falvey then poached Rangers AGM Thad Levine — whom the Rangers had once wooed away from the Rockies, after he’d debuted with the Dodgers. Meanwhile, the Diamondbacks hired Boston’s GM, Mike Hazen (another Cleveland alum), and VP of international and amateur scouting, Amiel Sawdaye, to lead their revamped front office; Hazen soon hired respected Pirates analyst Mike Fitzgerald to oversee Arizona’s statistical efforts. Elsewhere, the Tigers hired a senior Yankees analyst to manage their own quantitative department.

The high-profile vulnerability of Ground Control, the Astros’ internal (and, on the 48 occasions Correa accessed it, external) information system, inspired some teams to beef up their security. “We are developing a new app, and at the start of it, I stressed to the development team that we cannot have an Astros incident,” says the director of one AL team’s quantitative crew. But while many teams have taken steps to reduce the threat of outside infiltration, the real risk of a leak is coming from inside the office.

Some clubs create cultures that employees are less likely to leave, but the incestuous shuffle of scouting and front-office talent that we’ve seen since the fall takes place every offseason, driven by the layoffs, promotions, and regime changes endemic to a public-facing, results-driven industry. And every time these employees take their skills elsewhere, lured by a raise or a more prestigious title, they pollinate a new franchise with their old franchise’s seeds. One idealistic R&D director portrays those promotions in a positive light: “To some extent we want them to be able to use what they’ve learned as that is a part of the process of growing and moving into positions of higher responsibility,” the director says. Another analytics director probably speaks for more of their peers when talking about a brain wipe: “You would love to have the Men in Black thing, where you could just flash it and they wouldn’t remember anything from the last X number of years.”

Sometimes, teams hire employees who have preexisting IP, such as a proprietary projection system. Those employees have historically retained the rights to their creations, although now that every team has embraced analytics, fewer would risk losing a vital tool by agreeing to such an arrangement instead of buying out the creator or developing their own solution. Even in cases where an employee retains the rights to a specific product, anything the person produces while employed by the team is legally that team’s permanent property. Contracts for full-time front-office employees are standard, and nondisclosure agreements are common for full-timers and interns alike, although noncompete clauses that extend beyond the term of the contract are rare, if not unheard of.

One strategy that some teams have tried is walling off access to sensitive information on an employee-by-employee basis, which doesn’t help when a GM or AGM leaves but can offer some protection in the event of an intern or more junior staffer’s departure.

“The best thing a team can do is simply allow the minimum possible access to its employees,” one scout says. “If you’re an amateur scout, there is absolutely no need for you to read anybody else’s reports, or [player development] reports, or medical reports, or pro scouting reports or international scouting reports.” A scout’s personal reports occupy an undefined area: Technically, the reports a scout produces belong to his team. But based on some responses, it’s not uncommon for scouts to keep copies of reports they submit, maybe because they know teams probably wouldn’t go through the trouble of prosecuting someone for hoarding his own subjective opinions about players.

“I think that every organization (the smart ones at least) assumes that anyone leaving will bring their knowledge of the organization with them and accordingly restricts access to … secrets by creating ‘information silos,’” says a former scouting executive. When I interned for the Yankees from 2009–10, I couldn’t view scouting reports and some other sections of the team’s information system, BASE, even though I’d signed a (since-expired) NDA. But building walls comes with costs.

“I think the silo-ing thing is better in practice than in reality,” says one AL director of baseball operations. “It creates real morale issues in the staff if they are walled off from things, particularly once you get into director and higher levels. Everyone doesn’t need to know every piece of information, but if you start excluding department heads from certain things in the fear that they might leave, you are sort of inviting them to leave for somewhere else where they will be more involved and more trusted.” A department head with a different AL team says they haven’t restricted subordinates, for similar reasons. “I know people who have worked in that environment and it really impacts job satisfaction negatively,” the department head says. “I want my team to feel like they’re a part of something bigger…they work too hard to not get a feel for how their work is impacting the ultimate mission.”

Limitations on access make more sense for some teams than others. A few early adopters have been developing their databases for more than a decade, and they cycle through interns so fast that they can’t give them all unfettered access without jeopardizing whatever advantage they’ve gained through their years of innovation. (The Yankees belong to that category; the building of BASE began shortly after the 2005 hiring of current AGM Michael Fishman, and, when I worked there, the height of the summer sometimes meant more interns than office space.) Those teams tend to target new employees with specialized skill sets, who work on specific tasks within their areas of expertise.

Not only does that increased security run the risk of impairing the front-office version of clubhouse chemistry, but it can lead to less production. “Some teams such as the Yankees do a better job with consultants by walling them off from [the] bulk of the internal IP, but that comes at the very steep cost of lowering [their] value and impact,” says Christopher Long, a former senior quantitative analyst for the Padres who’s since consulted for multiple teams. A walled-off employee can’t make as many direct contributions, and the smaller the pool of potential peer reviewers, the more likely it is that mistakes will survive. “Sometimes a great idea starts to fall apart when opened to wide-scale scrutiny, and when that idea is private, very few people have the ability to give that idea the necessary scrutiny,” says Voros McCracken, who pioneered defense-independent pitching stats and went on to consult for teams.

Analytics laggards and teams with smaller budgets don’t have the luxury of parceling out access selectively. “The smaller the department, the more you kind of have to trust people and give them access to everything, because there’s just so much work to be done that it’s very difficult to just silo someone off and give them one project,” says one R&D director. “So it’s like, ‘OK, here’s our database, and go wild, and sign the NDA, and I hope you don’t screw me over.’”

Teams do enjoy some degree of built-in protection, in that some projects are often so complicated that even if everyone in the front office is aware they exist, no single contributor can master all of their intricacies. “If you have multiple people work on the infrastructure (taking care of different parts) then no single person would be able to recreate the whole system,” one scout says. Inevitably, though, some employees do discover enough to be dangerous. Unlike computers, people don’t always have to be hacked to surrender their secrets. Some of them like to talk.

“Our data is very secure on the technology side, but the biggest challenge is always the social aspect,” says one R&D director. “People freely share information that we’d prefer they not share. Advance reports are occasionally traded with other clubs — sometimes with both clubs’ (or at least advance scouts’) consent, sometimes because a player or coach carries it from one club to another.” In other words, an advance scout for an eliminated team might share information on a playoff contender’s upcoming opponent, either out of friendship or in exchange for future favors.

One longtime front-office type tells me that in theory, a team could try to patent its algorithms and models. But to do that, it would probably have to disclose a full description of the idea in the public patent application, which wouldn’t be worth the dubious protection it might derive if the patent were approved. “The reality to every situation is there isn’t any protection for the development of intellectual property rights,” says former Rockies GM Dan O’Dowd. “Most of the time in offices that are more inclusive by nature you will be exposed to the development and actual usage of the systems you develop and as such when you leave you take that with you. In fact, in most cases that is part of the reason you are being hired to begin with.” A team whose farm system is struggling might recruit a new director from a team that’s excelled in scouting and player development, and a team that’s struggled to assemble a quantitative department might pluck an employee from a team that’s known for its analytical acumen. Implicit in moves of that sort is the expectation that the new recruit will implement a process that resembles the successful one.

When deciding where the line is for acceptable recycling, most respondents draw a distinction between ideas and products — the eureka moment and the countless lines of code you write to make that moment matter. “I believe standard business practice would be if they can remember it you cannot stop them from taking it with them,” says Twins director of baseball research Jack Goin. One of his counterparts with another team expresses it this way: “My understanding is that concepts will travel with people.” And as the AL baseball ops director says, “My feeling on this is that forms or format in which information is presented is roughly fair game. If there is a particular style of depth chart, graphic display or way in which a platform looks and acts it’s nearly impossible to prevent people from bringing that from one place to another. Whether they take the actual form or just remember it and describe it for re-creation it’s the same result. On top of that … if there’s an organizational focus on bio-mechanics or a philosophy on shifting or types of amateur players that they believe in, again it’s nearly impossible to monitor.”

The nightmare scenario is a complete data dump. “One end of that scale is going into the source code for whatever system you have and hitting copy and saving it to your Dropbox and then taking that to some other team,” says an NL R&D department head. “That’s in my mind, like, the worst thing you could do, or doing a ‘Select *’ from the database and dumping that into a bunch of files and just taking all the data and all the models and everything. That’s a thing where it’s like, we would prosecute you for that.”

Some sources say that the Correa precedent will make them even more scrupulous about being aboveboard if and when they change teams. But there’s still a strong incentive to engage in some Jonah Lehrer–style self-plagiarism, despite the specter of Lehreresque career consequences. “In the middle I think there’s a situation where it’s like, you do the project, you write the code, you do everything, and then you might still have access to that code that you wrote just because you have it in your email or whatever,” says the same NL R&D exec. “Do you use that for your new team or not? And that’s kind of where the gray area is. And I think teams would strongly prefer that you not reuse things that you’ve done, but hey, you know, my intuition is that many people will do that.” Ironically, each team’s impulse to keep every other team in the dark about what it’s doing makes it easier for staffers to get away with this sort of light lifting. “It would be hard to enforce anything short of a data breach or stolen code since there is so little transparency between clubs,” says one of the quantitative department heads. “I don’t see a practical way to stop it.”

Here’s the humbling truth behind the clandestine behavior: Baseball secrets don’t matter as much anymore. It was one thing when only a few teams were devoting time and money to new-age analysis; a single big insight might mean a years-long head start. In an era when the number of in-house sabermetricians (and scouts) has exploded and the public sphere is more sophisticated than ever, the available edges are smaller and erode much more quickly. In spring 2009, the Yankees measured the impact of catcher framing, possibly earlier and more accurately than anyone else. Within two years, the same insight was all over the internet, and six years later I was writing about it as if it were ancient history, with the team’s knowledge and blessing.

“It’s not a perfectly efficient market, but the catcher framing stuff comes out and then two years later every team’s doing that,” says the NL R&D head. “The spin rate stuff comes out and then every team’s looking at it. There’s a very short amount of time where there’s a competitive advantage to almost anything, I think, especially when it’s something that can be reproduced with open-source data.”

As more and more Statcast data trickles out to the public, we’re approaching a point when the stat-savvy fan can form a nearly complete picture of a player’s performance on the field. Teams have already reached that point, but they also have a handicap that narrows the gap between public and private knowledge: They lack the public’s ability to crowdsource with a brain trust whose size is limited only by the moderate difficulty of communicating online. “Now more so than ever,” the former scouting exec says, “secrets are harder to keep secret from the public at large. An enterprising independent analyst is likely to come closer to reverse-engineering insight into a team’s roster-building process than many clubs would be comfortable admitting.”

McCracken adds, “I think a lot of folks would be surprised at how little ‘secret’ information there really is at any one team, and even further surprised at how little benefit that ‘secret’ information confers.” Mariners director of analytics Jesse Smith sees it the same way: “Hypothetically, by the time someone has taken a statistical method elsewhere, has been able to implement it and is in a position to use that information to influence the decision-making of other teams, we would probably be onto the next thing.”

As the former scouting exec explains, information that seems as if it should be priceless actually begins to depreciate the second it’s obtained. “Knowing what the Red Sox’s free-agent preference list looks like only does you good for one offseason,” the exec says. “Knowing that the Dodgers loved a kid in A-ball doesn’t mean that they will like the same guy the next year when he is posting a .245/.280/.360 line in Double-A. Knowing that a particular player is an alcoholic or a bad locker room presence is not a secret, just ask anyone he’s ever played with or any clubhouse attendant/boot room manager who has worked with him. The real secret is in the discipline that it takes to implement strategy.” Another NL R&D director stresses, “The execution of the ideas is far more important than the ideas themselves, and so that tends to be the most difficult thing to reproduce in another organization.”

Baseball’s statistical frontiers aren’t entirely settled, but if today’s sabermetric movement were a Western, it would be the latter-day kind, where the trains are running on schedule, civilization is on the ascendant, and the outlaws find themselves facing a Gatling gun. When every front office knows more than the smartest team did a decade ago, spying on a stat page or stealing a scouting report can’t compare with the benefits of buy-in from the top of a team to the bottom. The new gold rush is the GM who works well with the manager, or the hybrid ex-player who’s fluent in both stat- and scout-speak and can translate from front office to field or from field to front office.

“That is the key: how you maintain … consistency in the face of success, how you remain open-minded enough to change despite your success, how you continually have personnel that leak nothing to the media,” says the former scouting exec. “That requires dedication and commitment that can’t just be thrown on a flash drive or backdoored by someone with knowledge of a system; the real secrets are the ones that can’t be stolen.”

Maybe that’s the lesson to take from this whole sordid story. We’ve known for some time that Correa’s crimes were illegal, unethical, and punishable by many months in prison. What we might not have known makes the story sadder still: In baseball’s current climate, it’s not even clear how much hacking helps.