Continuing its string of super-accurate NHL postseason predictions, UNU correctly picked 3 of 4 major NHL awards given out last night in Las Vegas to the top individual performers of the 2015-16 season, including League MVP Patrick Kane of the Chicago Blackhawks. UNU of course is the Swarm Intelligence that outdid the experts picking the Oscars, and more recently, predicted the Superfecta at the Kentucky Derby (1st through 4th finishers).
As documented in this blog, a CNET reporter challenged UNU to predict the winner of the Stanley Cup Playoffs. At the time, there were 4 teams remaining–Blues vs. Sharks in the West & Penguins vs. Lightning in the East. Researchers at Unanimous A.I. assembled a swarm of 29 hockey fans. That swarm predicted the Penguins would win, which of course, they did.
Predicting the winner of a 4-team tournament would be a decent result, considering there was little consensus from experts. But that alone would hardly be remarkable. But here’s where it got interesting. The swarm went for extra credit. They predicted not only the winners of each series, but also the number of games.
UNU nailed the Semifinals perfectly, predicting the Sharks would beat the Blues in 6 games, and that the Penguins would beat the Lightning in 7. They came within 1 game of perfection in the Finals, predicting the Penguins would take 7 games to beat the Sharks, when in fact Pittsburgh needed just 6 games to dispatch San Jose.
Even so, UNU predicted 20 of 21 games accurately — 95% accuracy. But the Unanimous A.I. researchers weren’t done. They challenged UNU to go beyond the Stanley Cup, and forecast the winners of the major NHL Awards.
This UNU Swarm of 29 hockey fans correctly pegged NHL scoring leader Kane as MVP. UNU also nailed the Calder Trophy winner, given to the Rookie of the Year, correctly picking Kane’s Blackhawk linemate Artemi Panarin.
For good measure, UNU also predicted the Vezina Trophy winner, given to the league’s best goalie. Yep, UNU got that one right too. The swarm accurately predicted that Washington’s Braden Holtby would beat out fellow keepers Ben Bishop of Tampa Bay, and Jonathan Quick of the L.A. Kings. UNU wasn’t quite perfect, missing on its Norris Trophy pick for best defenseman.
All in all, UNU’s NHL predictions were stunningly successful. This swarm, made up of 29 average hockey fans, crushed the experts. What’s more, as a group, the swarm was far more successful than any of its individual participants. Here is the breakdown:
- UNU Picked 20 of 21 games correctly (95%) & 3 of 4 NHL Awards (75%) — TOTAL: (23 of 25 = 92%)
- The average individual participant scored 15 of 21 (72%) & 2 of 4 on Awards (50%) — TOTAL: (17 of 25 = 68%)
- Out of the 29 participants, no individual achieved better than 84% accuracy. This means the Swarm Intelligence performed better than every single participant who comprised it. In other words, every individual would have been better off betting with the swarm…
We’d love to have you join our next swarm. Come check it out here.
