Braess’s Paradox
Adding a new path to a shared system can make everyone worse off when people independently reroute around it and collectively settle into a worse equilibrium than before.
What Is It?
Mathematician Dietrich Braess demonstrated in 1968 that adding a new road to a congested traffic network can, under the right conditions, increase everyone’s travel time rather than reduce it. The mechanism isn’t a fluke of one particular map. It’s structural: when a new route is added, individual drivers each choose whatever path looks fastest for them at that moment, with no way to coordinate their choices with everyone else making the same calculation. If enough drivers are drawn onto the new route because it looks like a shortcut, the route that seemed fastest in isolation becomes congested by the very people who chose it for that reason, and the network as a whole ends up slower than it was before the option existed. Every individual driver is optimizing correctly given what everyone else is doing. That’s exactly what produces the collectively worse outcome. Removing certain roads from an already-congested network can, by the same logic, speed everyone up.
Why Does It Matter?
Organizations add options constantly, believing more choice can only help: a new approval fast-track for urgent requests, an additional communication channel, a new escalation path, an alternate way to get a decision made. Each addition looks purely beneficial from where it’s added. But the paradox isn’t simply that the new option gets busy while the old path sits idle, it’s that the new option changes everyone’s routing incentives at once, and the whole system can resettle into a genuinely worse equilibrium than before the option existed, where every individual is still making the locally sensible choice given what everyone else is doing, and nobody acting alone has any incentive to switch back. This is a different problem than simply having too many options. The paradox specifically requires that people are choosing individually rational routes without anyone coordinating the system-wide effect of everyone choosing at once, which exposes a gap between what’s individually stable and what’s collectively efficient: nobody has a reason to change their own route unilaterally, even though everyone would be better off under a different overall pattern.
What Changes Once You See It?
You stop assuming that adding a new path, process, or option is automatically an improvement, and start asking how people will actually redistribute themselves across the whole system once the new option exists, not just whether the option itself looks faster in isolation. You start watching for a new “fast track” or “shortcut” quietly becoming the new bottleneck, precisely because everyone rushed toward it for the same reason it looked fast in the first place. You also become willing to test whether removing a seemingly useful path could improve an already-congested system, since the paradox runs in both directions: a badly-timed addition can produce a worse equilibrium, and in some congested networks, fewer available routes produce a better one.
Common Misunderstandings
- It isn’t a claim that adding capacity or options is always counterproductive. The paradox only appears under particular network structures and patterns of decentralized choice, adding capacity doesn’t automatically make performance worse.
- It isn’t the same as simple overload from too many choices. The mechanism is about how a new option redistributes existing traffic across the whole system, not about decision fatigue from having more choices to consider.
- It doesn’t mean centralized control is always the fix. Coordinating everyone’s choices centrally can solve the paradox, but so can more targeted interventions, closing a specific bottleneck route, that don’t require controlling every individual decision.
- It isn’t limited to physical infrastructure. Any system where people choose their own path through shared, congestible resources, approval queues, communication channels, support tiers, is a candidate for the same dynamic.
Diagnostic Question
If everyone responds to this new path the way that makes sense for them individually, what does the system look like once everyone has adjusted, not just where the new path itself sends traffic?
Explore Further
Field Notes
None yet.
Related Field Guide
Origin
Dietrich Braess, “Über ein Paradoxon aus der Verkehrsplanung” (“On a Paradox of Traffic Planning”) (1968); the underlying mathematics draws on earlier work by John Glen Wardrop on how individually optimizing traffic reaches equilibrium.