Navigation is critical to predictions, specifically in calculating space over time. Errors (or uncertainty) in navigation can have life-and-death consequences — though not in Harvard Yard. Good thing, too: The students ended up all over it, though they were following the same directions. In the follow-up class, Goodman shed more light on the statistics of navigation and let the students know who came closest to the actual locations. The winning pair got a historical map of Harvard.
“The goal of the full exercise is to demonstrate why and how uncertainty around navigational predictions was and still is so important,” Goodman said. “When asked to report an estimated uncertainty in their final position, nearly all students grossly underestimated how far off their final position was likely to be. When they finally see everyone’s position, with estimated uncertainties, on a map and realize how common overconfidence is, I think they get a very good idea of how and why estimates of a prediction’s uncertainty can be just as important as the prediction itself.”
Students described the task as fun and challenging, and a good reason to get outdoors at a time most academic work is happening online.
“This is the only class activity which has gotten me out,” Khurana said. “It’s really cool. It reminds me of pre-COVID times a little bit.”