The Law of Escalation: How Rational People Create Irrational Outcomes

The Problem With Looking at Decisions in Isolation

Something that genuinely shifted how I see things is how often we judge events like they exist in a vacuum. When something goes wrong, we immediately want to find the cause. The decision that started it. The person who messed up. The exact moment things went sideways. It makes sense now that we’re wired to think in stories, and stories need a beginning and a villain.

But that’s rarely how things actually play out in real world. Most situations aren’t the result of one bad call. They’re the result of a whole chain of decisions, where each one shapes the next. By the time you can actually see the outcome, it’s already the product of dozens of reactions stacked on top of each other. Hunting for the "original cause" can actually get in the way of understanding what really happened.

Game theory kind of broke my brain in a good way here. Instead of just asking "why did this happen," it pushes you to ask "how did people respond when it happened." That small shift in framing changes a lot.

We do this with big societal changes too that point to a new technology, a law, some major event. And sure, those things matter. But what matters just as much, maybe more, is everything that happens after. People adapt. Organisations shift. Competitors react. Incentives quietly rearrange themselves. And slowly, the whole environment looks nothing like it did before. Simply putting it - Actions matter, but the reactions? They often matter more.

Why Escalation Feels Rational

Here’s the thing about escalation that I find genuinely surprising that it almost never feels irrational while it’s actually happening.

In hindsight, a lot of escalation cycles look absurd. You look back and wonder why nobody just... stopped. Why didn’t someone see where this was going? But the uncomfortable truth is that most of the people involved were probably acting completely rationally. That’s what got me. I always assumed irrational outcomes had to come from irrational decisions. Game theory keeps suggesting otherwise. Sometimes reasonable people making reasonable decisions collectively produce something nobody actually wanted. Picture two competitors. One pulls ahead, the other responds because not responding means falling behind. The first one reacts to that, because now their position is threatened. Neither side is being reckless. Both are just following the logic in front of them. But somewhere along the way, they’ve both ended up worse off than when they started. The irrationality isn’t in any single decision. It lives in the system. That’s also why it’s so hard to catch in real time. Every step feels justified. Every reaction feels necessary. Everyone has a perfectly reasonable explanation for what they did. The problem only becomes obvious once you zoom way out and look at the whole pattern.

We expect irrational outcomes to come with obvious warning signs. Most of the time, they come from people behaving exactly as you’d expect.

The Internet Didn’t Become Extreme Overnight

The internet is probably the most visible example of escalation in everyday life.

People talk about online culture like someone designed it to be this way. I don’t think that’s true at all. It didn’t become attention-driven and emotionally charged because of one decision. It got there through millions of small decisions all bumping into each other over years.

At some point creators figured out that certain content got more traction. Platforms noticed engagement kept people around. Audiences naturally drifted toward stuff that hit harder emotionally. None of that was wrong, exactly. It was all accurate. The problem is that everyone started adapting to the incentives everyone else was creating.

If you’re making content today, you’re not competing against the internet from a decade ago. You’re competing against an environment that’s been shaped by years of this cycle. The baseline keeps shifting. What used to feel bold becomes normal, so the next wave has to push further just to get noticed. And here’s what gets me is that most people weren’t trying to make things more extreme. They were just responding to what was already there. But by responding to it, they helped push it further. Nobody’s fully in control, but everyone’s contributing to the direction it moves.

The Hidden Feedback Loop

The more I think about escalation, the more I see feedback loops underneath almost everything. A feedback loop kicks in when an outcome starts influencing the conditions that created it. The result feeds back into the process. Once that happens, things stop being static and starts evolving on their own.

What makes them tricky is that they’re almost invisible up close. You only see your small piece of the system. You react to what’s right in front of you. Very rarely does anyone have a view of how millions of similar reactions are all interacting at once.

That’s where escalation gets really interesting. Every response changes the incentives for the next one. Every adaptation creates new conditions someone else has to deal with. Over time, the game itself changes, even if nobody officially changed the rules. You think you’re reacting to the environment. But you’re also building it.

I think this is one of the most useful things I’ve taken from game theory. We’re obsessed with decisions ,who made the right call, who was responsible, who should have known better. But a lot of the most important outcomes are shaped less by the decisions themselves, and more by the chain of reactions they set off. That’s why escalation is so hard to stop. By the time you notice it, it’s not a single event anymore. It’s a process. A self-reinforcing cycle where everyone’s responding rationally to a world that everyone else helped create. And that’s the strange paradox at the centre of all this. Irrational outcomes don’t always need irrational people. Sometimes they just need rational people reacting to each other, each making a call that makes total sense on its own, while collectively producing something none of them ever intended.

The ideas are all preserved, but the phrasing is looser, more direct, and reads like someone actually thinking through it rather than presenting it.