Before the official June 1 start of the Atlantic hurricane season, we invited you to enter our client hurricane contest—purely for yucks, with the coveted prize of bragging rights. And while the season officially ended on November 30, tropical cyclones obey laws of physics rather than arbitrary calendar dates and sometimes develop after that date, so it is our custom to stop the counting and close our contest at the stroke of midnight following the ball drop on New Year’s Eve. Now that we have passed that milestone, we can note the official totals for 2021 and announce the results.
Although not quite as active as 2020, Atlantic hurricane activity was quite lively in 2021 and set a few records too. The most memorable feature of this year’s storms will likely prove to be the record rainfall that accompanied hurricanes Henri and Ida, most of which impacted the U.S. Northeast more severely than the Gulf Coast. In total, 21 named storms formed, the third highest total on record after 2005 (28) and 2020 (30). Of these storms, seven developed into hurricanes, four of which became major hurricanes. Despite the high number of named storms there were just two U.S. landfalling hurricanes, only one of which—Ida—was a major hurricane. Unsurprisingly, Ida proved to be the costliest hurricane of the season, thanks in part to the flooding it caused in the Northeast metro corridor.
There we have it. The challenge set in our contest was to predict how many named storms, hurricanes, major hurricanes, and U.S. hurricane landfalls the North Atlantic basin would see in 2021, and the numbers realized were 21, 7, 4, and 2 respectively. Once again, we had a great response to the contest. 2021 was also the first year we jettisoned our rule about choosing a winner from the set of folks with tied lowest error points (defined as the sum of absolute count errors across the four predictions) using a purely random tie breaker.
The result? A seven-way tie for the title between our skilled and experienced participating client base! We extend our hearty kudos and congratulations to Marco Rohrer and Laura Sime of Axis Capital, Eva Horackova of Guy Carpenter, Yurie Budhu of Travelers, Colin Wildey of Fidelis, Matthew Boyce of Hannover Re, and former AIR employee Anna Neely, now of Tiger Risk. All these champion 2021 hurricane count forecasters submitted forecasts with just four error points.
So, how did AIR staff do in the parallel contest we run internally? Well, some of us ran complex computer simulations, others consulted their psychics or simply picked random numbers and we had the usual wide range of responses. The ensemble of AIR forecasts had a mean of 10.4 error points—beating the ensemble client forecast mean of 11.2. Our internal champion for the 2021 season contest was Kulbhushan Joshi, whose forecast bottomed out the AIR error distribution with just a single error point. John Weis and Byron Seese each garnered only two error points, but Byron managed to squeak ahead into second place thanks to the official contest tie-breaking rules.
If we think of this as a contest within a contest, AIR staff outperformed client entrants because they scored fewer error points. You will have your chance to fight back when we announce our hurricane contest for 2022 in May!
This contest focuses on the numbers of storms, and in 2021 the total of named storms was significantly above average but the figures for hurricanes, major hurricanes, and U.S. landfalls were broadly in line with the long-term average. It is important to remember that an active season doesn’t necessarily mean high losses, and that an inactive season with just one major storm—even a bypassing one—can cause catastrophic losses.