Worst-Case Scenario First

The bank gets the best case. The buyer keeps the worst.

Every deal gets two financial models, and only one of them is allowed to decide. The best case goes to the bank, because a lender wants optimism and that is the appropriate frame for borrowing. The worst case stays close to the chest, and it alone determines whether the deal proceeds. "If the worst-case scenario doesn't work then you better not do the deal, because 90% of the time it's the worst-case scenario that happens."¹ [18:30] The asymmetry is the whole point: the bank sees the best case, the buyer decides on the worst, and the two never trade places.

The worst case is not the apocalypse

Tilman Fertitta has run this discipline on every deal he has done, and it is the governing rule underneath the Platform Consolidator archetype. The private model stops short of doomsday: it assumes the marginal locations close and only the anchors survive, then asks whether the deal still works. If it does, he proceeds. If it does not, no best-case upside can rescue it.¹

Six months later, $450M cheaper

The clearest case is Rainforest Cafe. Fertitta tried to buy it, was outbid by a teachers' fund in Wisconsin, then bought it six months later for $450M less, with $15 to $20M of cash sitting on the balance sheet at close. His worst-case model asked one brutal question: if only the five highest-revenue anchor locations survive, three at Disney, one at the Mall of America, one in Chicago, does the deal work? The answer was yes, so he proceeded. Twenty years later, 26 locations were still operating.¹ The discipline reflects a real asymmetry in acquisitions: the upside is capped at the cash flows a buyer can collect, while debt structures and operational failures can destroy far more than the purchase price. Modeling the worst case is the minimum discipline that keeps a buyer solvent through the outcomes that actually arrive.

Knight would have failed this gate

The rule fits high-capex, low-reversibility environments, which is exactly where hospitality and gaming sit. In fast-turning domains it can be actively wrong: Knight built a massive company on the opposite philosophy, emptying the bank account every month and doubling every order, because running-shoe inventory is far easier to reverse than a casino floor.¹ The other failure is discipline rot. The model only protects the buyer if the best case never seeps into the decision, and the moment an operator starts believing the bank presentation, the worst-case gate is already open.