As the Seattle Seahawks and the New England Patriots meet again on football’s biggest stage, the rematch inevitably pulls everyone back to one moment. Different rosters. Different seasons. Same unresolved question. With a championship on the line, memories of the one-yard line come rushing back, not because the teams are the same, but because the decision still divides fans, analysts, and armchair quarterbacks.

As a brief recap (apologies to Seattle fans for reopening old wounds), late in the fourth quarter of Super Bowl XLIX, Seattle trailed by four points with roughly twenty seconds remaining and only one timeout. On the one-yard line, the decision was widely framed as obvious: hand the ball to Marshawn Lynch (aka Beast Mode), arguably the best power running back in the league at the time. But the game situation significantly narrowed the Seahawks’ options. With just one timeout left, a run on second down that ended in bounds would have forced Seattle either to burn its final timeout or rush into a highly predictable third-down play under extreme time pressure. Bill Belichick understood this constraint. From a clock-management and optionality perspective, a pass on second down was not reckless, it was rational. An incompletion would stop the clock and preserve both third and fourth down. The Patriots were not guessing. They were reading the situation.

When Russell Wilson released the ball, the defense was ready. Malcolm Butler jumped the route, intercepted the pass, and ended the game. In hindsight, the interception hardened the narrative. But hindsight obscures the real lesson. The play call did not fail because it was irrational. It failed because high-reward decisions carry inherent risk, especially when the other side understands the constraints driving the call.

That same dynamic plays out when companies decide whether to patent an invention or keep it as a trade secret. On the surface, patents often look like the riskier choice. A patent application will be published, usually eighteen months after filing, meaning competitors get a clear window into how the invention works. A patent also has a finite life, generally twenty years from the earliest filing date, after which the invention falls into the public domain. Unlike a trade secret, which can theoretically last forever, a patent comes with a built-in expiration clock that starts ticking the moment you file.

More importantly, patents are not guaranteed. Filing an application does not ensure that claims will be allowed, that the scope will be meaningful, or that the rights will survive challenge. During prosecution, claims may be rejected, narrowed, or abandoned altogether. That uncertainty is especially acute in software and AI, where eligibility concerns, obviousness rejections, and functional claiming issues routinely derail applications. In that sense, pursuing patent protection is very much like throwing a pass at the goal line. The play might succeed and win the game, but there is always a real possibility that it gets intercepted by an examiner unconvinced that the claimed invention passes the patentability goal line.

Trade secrets, by contrast, feel like the power run. No publication. No term limit. No examiner second-guessing claim scope. If secrecy is maintained, protection can last indefinitely. For certain categories of innovation (e.g., manufacturing processes, server-side algorithms, internal workflows, or secret recipes and formulations that are difficult to reverse engineer), this can be exactly the right call. But just as a run up the middle only works when the blocking holds, trade secrets only survive when a company puts real protection in place. Reasonable secrecy measures are the IP equivalent of an offensive line and protection package: access controls, compartmentalization, NDAs, monitoring, and training are what keep the defense from blowing up the play. Once that protection breaks down, through employee turnover, a leak, reverse engineering, or independent development, the ballcarrier is exposed, and the protection is gone instantly, with no replay and no appeal.

Belichick’s insight at the goal line is the same insight sophisticated companies bring to IP strategy. He understood not only what Seattle wanted to do, but what Seattle could do given the constraints. In IP, those constraints include how visible the invention will be once commercialized, how likely competitors are to independently arrive at the same solution, how fast the technology will evolve, and whether enforceable rights against third parties are more valuable than indefinite secrecy. On the trade‑secret side, those constraints also include how much “surface area” the technology exposes: the number of employees who must know it, the degree to which it shows up in customer‑facing outputs, and how much operational friction the company can tolerate in order to keep it locked down. In software and AI, where patents face higher prosecution risk but trade secrets can evaporate the moment code behavior becomes observable, the choice is rarely obvious.

The real lesson of that Super Bowl play is not that passing was foolish. It is that good decisions carry risk when the other side understands your limitations. Patents disclose and expire. Trade secrets endure, but only as long as the protections around them do, and those protections can collapse overnight. Filing a patent may result in powerful rights or nothing at all. Keeping something secret may work for decades, or fail the moment the information leaks, the protection measures break down, or a competitor figures it out independently. IP strategy, like football, is not about choosing the move that looks safest in hindsight. It is about making a defensible decision in real time, with imperfect information, knowing full well that even the smartest call can still end with the ball going the other way.