When Machine Learning Produces the Same Forecast for Everyone, the Perfect Session Becomes a Crowded Session
Bohiney Magazine | The London Prat
The Perfect Forecast Day: When Everyone Gets the Same Information at the Same Time
SOUTHERN CALIFORNIA — The forecast said “excellent” at 7am Saturday. The surf forecast services — Surfline, Spitcast, Magic Seaweed, the various independent forecasters — all agreed: overhead-plus, offshore winds until noon, long period southwest swell, the kind of conditions that justify waking up early and driving wherever the conditions are. Approximately thirty thousand people in Southern California received this forecast on their phones. A significant proportion of them drove to the same breaks. The breaks, which are excellent for surfing in small groups and technically challenging at crowd densities above thirty people, were surfed by significantly more than thirty people.
This is the information asymmetry problem solved in reverse: previously, local knowledge determined who knew about good conditions and who arrived first. The surf forecasting services eliminated the information asymmetry by providing the same information to everyone simultaneously, which is a genuine service and which has also eliminated the friction that local knowledge provided as a natural crowd-management mechanism. When everyone knows the same thing at the same time, everyone arrives at the same time, and the “excellent” conditions are surfed in the way that excellent conditions crowded with frustrated surfers are surfed, which is not the way the forecast description implied.
The Information Commons Problem
This is Garrett Hardin’s commons problem applied to surf forecasting: the resource (the good surf break) is finite, the information that directs users to it is now widely and instantly available, and the result is congestion that degrades the experience for all users. The paradox is that better information produces worse outcomes for the individual user while being genuinely better information. Better market information is supposed to improve outcomes; in a finite natural resource context it concentrates demand in ways that exceed the resource’s capacity. Managing the commons in the information age requires solutions that the information age has not yet produced. The forecast was excellent. The session was crowded. The forecast remains accurate. The paradox remains unresolved.
SOURCE: https://bohiney.com/capitalists-solve-grocery-crisis/