Behavioral ecology has a fifty-year-old model, the ideal free distribution, that predicts something elegant: given a patchy environment of unequal resources, foragers will spread themselves across it in direct proportion to what’s available, arriving at a stable equilibrium with no central coordinator required. Gig-economy platforms look, on the surface, like a perfect real-world test case — drivers are foragers, city neighborhoods are patches, ride requests are the resource, and surge pricing is the signal telling everyone where the food is. It’s tempting to treat surge pricing as confirmation that human drivers behave the way the ecological model predicts. The more interesting and more honest finding, once you look at what platform researchers have actually documented, is closer to the opposite: drivers left to their own devices tend not to reach anything like the ideal free distribution’s stable equilibrium, and surge pricing exists specifically because that natural self-organization fails.
Scientific Foundation
The ideal free distribution, formalized by Fretwell and Lucas in 1970, predicts how foraging animals should spread themselves across habitat patches that differ in resource quality. The model’s name describes its two core assumptions: foragers are “ideal,” meaning they have complete, accurate knowledge of resource distribution and where competitors currently are, and “free,” meaning they can move between patches without cost or restriction. Under those conditions, the model predicts foragers will distribute themselves in direct proportion to each patch’s resource input, an outcome called input matching, reaching what game theorists would recognize as a Nash equilibrium: nobody can improve their own intake by switching to a different patch. The mechanism driving that equilibrium is crowding itself — as more foragers pile into a resource-rich patch, competition drives down each individual’s share, until the effective payoff there drops to match whatever’s available in the less crowded alternatives. It’s a genuinely elegant self-correcting system, and it’s been extended to account for unequal competitors too: stronger foragers systematically claim the richer patches, pushing weaker competitors toward lower-quality but less contested ground.
Cross-Domain Connection
Ride-hailing platforms operate on a structurally similar premise, at least in their stated design goals. Uber and Lyft divide their service areas into fine-grained geographic zones, track supply and demand within each in something close to real time, and use surge pricing — a multiplier applied to fares when demand outstrips available drivers in a given area — specifically to steer driver behavior toward under-served zones, alongside predictive heat maps that show drivers where demand is expected to spike before it fully materializes. On paper, this looks like an engineered analog of input matching: drivers, like foragers, are supposed to redistribute themselves toward wherever the “resource” (ride requests) is richest, with the price signal doing the work that crowding-based payoff decline does in the ecological model.
What Remains Undemonstrated
Here’s where the analogy needs real correction rather than confirmation. Research into ride-hailing markets has documented a specific, named failure mode that the ideal free distribution’s assumptions are built to rule out entirely: the “Wild Goose Chase” equilibrium, identified in work by Castillo and colleagues and discussed in subsequent driver-pricing research. When a demand spike appears in one area, drivers converging on that signal don’t reliably experience the smooth, self-correcting crowding effect the ecological model predicts — because unlike foragers with something close to complete, real-time knowledge of resource and competitor distribution, drivers are chasing a heat map built on prediction and historical pattern, not perfect present-tense information. Multiple drivers can rush toward the same apparent opportunity, arrive to find the ride already claimed or the demand spike already passed, and end up spending long stretches driving unpaid between pickups — a genuinely worse outcome than staying put, and a clear departure from the stable, mutually-improving equilibrium input matching describes. That’s precisely the scenario surge pricing is explicitly designed to prevent, according to the platform economics literature: not by passively reflecting existing driver behavior, but by actively correcting for the fact that driver behavior, left alone, doesn’t naturally converge on anything resembling the ideal free distribution’s equilibrium.
The deeper reason for the gap is exactly the assumption the ecological model’s name flags as essential and drivers don’t actually have: the “ideal” condition, complete and current knowledge of where the resource and the competition really are. Uber’s predictive heat-map systems are, in effect, an attempt to manufacture that missing precondition through information engineering — trying to give drivers something closer to the complete, real-time awareness animals are simply assumed to possess in the ecological model, precisely because human drivers, working from lagged, probabilistic demand forecasts rather than perfect present knowledge, don’t have it by default. It’s worth noting that even in the original ecological literature, real animals don’t perfectly satisfy the pure ideal free distribution either — experiments across species, including humans, regularly find a well-documented pattern called undermatching, a systematic, milder deviation from perfect proportional distribution. But the Wild Goose Chase problem in ride-hailing is a sharper, more acute failure than ordinary undermatching — a case where the corrective feedback loop can actively break down under a coordinated rush toward the same signal, not just settle slightly off the theoretical optimum.
Why It Matters
That reframing changes what the comparison is actually good for. Surge pricing isn’t validated by the ideal free distribution — it’s better understood as a real-time patch against exactly the market failure that model predicts shouldn’t happen if its assumptions held. Ride-hailing platforms are, functionally, running a continuous, high-stakes experiment in trying to engineer the informational preconditions a fifty-year-old ecological equilibrium model simply assumes for granted, in a domain — human commuters responding to an imperfect, predictive price signal — where those preconditions don’t arise naturally and have to be built, continuously, through algorithms.
Human Dimension
There’s a useful humility in this correction. It would be a satisfying, tidy story if gig-economy drivers turned out to be self-organizing exactly the way foraging birds and beetles do, proof that markets and ecosystems run on the same deep logic. The messier, more honest version is arguably more interesting: humans chasing an app’s demand signal don’t have the one thing that makes the ecological model work cleanly — genuinely complete, current knowledge of where everyone else is headed — and the platforms know it well enough to have built an entire pricing and prediction infrastructure specifically to compensate for that gap. The ideal free distribution isn’t a description of what rideshare drivers already do. It’s closer to a target the algorithms are still trying to build toward.
Sources:
1. Wikipedia — “Ideal free distribution” — https://en.wikipedia.org/wiki/Ideal_free_distribution
2. ScienceDirect Topics — “Ideal Free Distribution — an overview” — https://www.sciencedirect.com/topics/earth-and-planetary-sciences/ideal-free-distribution
3. ScienceDirect — “Putting competition strategies into ideal free distribution models: Habitat selection as a tug of war” — https://www.sciencedirect.com/science/article/abs/pii/S002251930600292X
4. ScienceDirect — “Ideal free distribution of unequal competitors: spatial assortment and evolutionary diversification of competitive ability” — https://www.sciencedirect.com/science/article/pii/S000334722300088X
5. ResearchGate — “Can Ecological Theory Predict the Distribution of Foraging Animals? A Critical Analysis of Experiments on the Ideal Free Distribution” — https://www.researchgate.net/publication/236833936_Can_Ecological_Theory_Predict_the_Distribution_of_Foraging_Animals_A_Critical_Analysis_of_Experiments_on_the_Ideal_Free_Distribution
6. Nikhil Garg (Stanford University) — “Driver Surge Pricing” — https://gargnikhil.com/files/papers/garg_driversurge.pdf
7. Brainforge.ai — “How Uber Uses Predictive Analytics for Ride Matching” — https://www.brainforge.ai/blog/how-uber-uses-predictive-analytics-for-ride-matching
8. Columbia Business School — “Ride-Hailing Networks with Strategic Drivers: The Impact of Platform Control” — https://business.columbia.edu/sites/default/files-efs/citation_file_upload/msom.2023.1221.pdf
9. Duke Fuqua School of Business — “The Algorithms Behind Pricing Your Ride” — https://www.fuqua.duke.edu/duke-fuqua-insights/algorithms-behind-pricing-your-ride
Idea originated at artificialideas.org. Article researched and written by Claude Sonnet 5. Published at artificialideas.org.
Artificial Ideas 