We found John Cassidy’s essay in the Oct. 5 New Yorker, “The Real Reason that Capitalism is so Crash-Prone,” to be illuminating about the challenges of managing in an irrational context, like the recent credit craze or the more distant dot-com and telecom bubbles. Cassidy argues that, even if managers know that they are in the middle of a bubble, they have little choice but to go along. Boards and investors tell them: “Do it, or move aside so that someone else can.” Few can resist such pressure.

Cassidy rightly casts the pressure as a form of the prisoners’ dilemma, where two suspects are separately given the choice to confess to a crime in return for a lighter sentence. Each suspects knows that the other is getting the same offer, and that if neither confesses both will go free. The problem is that if only one confesses and implicates the other, the one who didn’t confess will get a much harsher sentence. The best option would be for both to remain quiet. However, the almost inevitable outcome–and the most rational from a purely risk-reward standpoint–is that both will confess to avoid having to trust the action of the other to avoid the harsher sentence. This is true even if they are innocent.

So, imagine that you’re a manager in the days before the Great Recession and worry that the frenzied markets are a credit bubble that will soon pop. But your rivals, showing no such concerns, are recording record growth and profits. Your board, investors and analysts are complaining that your company performance is lackluster. What would you do? Like almost everyone else, you’d probably jump in, with devastating results.

What should you do instead? Based on our analysis of recent failures, here are some thoughts:

–First, make sure the pressure is real, and not just a rationalization for your aspirations. We’re thinking of Motorola, Inmarsat, Loral, TRW and others who raced to build multibillion-dollar satellite systems for voice and data communications in the 1990s. The companies argued that they had to get in the game, and quickly, or lose out. In reality, they were looking for excuses to try to commercialize their military technology at a time when defense businesses were under pressure. As we say at length in “Billion-Dollar Lessons,” each justified its moves by arguing that its approach was better than the others’. Yet none were viable–and just a bit more care in evaluating the technologies involved would have spared the companies billions of dollars of fruitless investments.

If you conclude you have no choice but to participate, do it with your eyes wide open and at least follow these caveats:

–Don’t be the last one in. If, like Merrill and Citigroup, you buy in late into a bubble, that can be very costly. Obviously, it’s hard to know when a bubble will burst. But you’ll be better off if you make assumptions about how long a cycle will last, then monitor how those predictions play out. You’ll at least be more aware that an end is coming, possibly soon. There’s a saying in poker that if you look around the table and don’t see a sucker, you’re the sucker. Don’t be the sucker.

–Always be ready to get out. As Lehman and Merrill learned, and as Greentree Financial learned in an earlier bubble, liquidity can disappear in an instant. All the financial models were based on the idea that it’d be possible to sell investments. The only question was at what price. In fact, as we’ve learned the hard way, some markets can have no buyers, at least for a time. So, it’s smart to play a treacherous market like Goldman did, making sure you understand your entire exposure to a market and hedging your risk.

–Whatever you do, don’t double down when things start looking bad. Take your losses and get out, rather than betting on a quick recovery. Leave the doubling down to the folks that don’t seem to understand they’re in a bubble, like Bank of America when it bought first Countrywide and then Merrill Lynch. Make sure you live to fight another day.