Mark Twain famously described Bermuda as paradise you have to go through hell to reach.

 A chronic seasickness sufferer, he served his time in hell on the dependably bumpy rides to his island getaway.

Twain was no seaman, but even experienced ocean sailors are wary of the 600-plus-nautical-mile passage from the East Coast to Bermuda that features sea conditions energized by a hot Gulf Stream flowing fast through a North Atlantic that is often a speedway for weather systems.

Wary indeed were the skippers of the 63 boats that withdrew from this June’s Newport to Bermuda Race before the race started. They had reason to be. The weather forecasts were scary: a gale blowing contrary to the Stream with resulting 20- to 30-foot seas.

It turned out that sailors on the 122 boats that started the race in spite of the daunting forecasts encountered relatively mild Bermuda race weather with Gulf Stream conditions that were typically uncomfortable but far from dangerous.

This provided fodder for a fair amount of post-race discourse, enlivened by the facts that the pre-race dropouts included most of the largest boats in the fleet (a number of 70-plus-footers) and that those that hung in included the smallest in the race (35-footers) and almost all of the doublehanded entries.

Hindsight delivered its usual clarity along with some uncharitable observations about the withdrawals, but it is probably safe to say that if the forecasts had proved accurate and the race was marred by dismastings, rudder failures and crew injuries, those who opted out would likely have been praised for their good judgment.

Skippers suffering from dropout remorse could always blame the weatherman, and some did. Ken Campbell, president of Commanders’ Weather, reported being asked by a number of them, “What happened to the storm?”

Commanders’ is the go-to weather forecasting and routing service for yacht racers, with a big following among the Bermuda race fleet. Racers who bought Commanders’ Newport-Bermuda package of four days of pre-race weather analysis and briefings ($495) got an 11-page “weather model recap” after the race.

The document offers a detailed explanation of the blown forecast. But the bottom line is that the reason for the mistaken predictions, though it involves weather models based on mathematical equations so complex they can only be solved by the world’s most powerful computers, was very simple: The models that everyone from the meteorologists working for services like Commanders’ to your local TV weatherperson depend on were wrong.

According to Campbell, the European model, produced by the system supported by 31 European countries that is often said to be more accurate than the U.S. equivalent, the Global Forecasting System (GFS), “was completely wrong.” GFS didn’t stumble as badly, but its predictions for the track of a low-pressure system were off enough to change everything.

Commanders’ forecasters put enough faith in the models to be alarmed by what they were seeing before the race. The words “ominous” and “storm” appear several times in the recap to describe models calling for east wind increasing from 25 knots, backing toward the north and bucking the Gulf Stream, resulting in a sea state that would be “very rough and getting worse.” 

Campbell concludes: “Given the same set of weather model output, I would forecast the same stormy weather leading into the start of the Bermuda race. It was the safest and smartest forecast given the intensity and track of the low in the model forecasts.”

Fair enough, but I wish he had foregone the last two sentences of his recap: “But we had the best result possible—nobody was hurt during the race and physical damage to the fleet was minimal! Given the forecast leading into the race, we could not ask for a better result!”

Doesn’t that sort of remind you of the beaming TV weather personality who takes credit for a sunny Fourth of July when he had predicted a rain-out?

In the final analysis, the Bermuda race weather surprise is a cautionary tale with a lesson most of us have learned but tend to forget—weather forecasting is still an educated guess. It’s mostly science, but there is art subject to human error in the interpretation of the science. What’s more, the science isn’t perfect either. Even the mighty computers of the European weather consortium and GFS can’t perfectly solve the physics equations fundamental to their models.

As someone who has spent many hours of his sailing life digesting Commanders’weather analysis and routing advice for numerous races, I try to keep that caveat in mind, but I’m still usually persuaded by the case these marine weather specialists make for various meteorological scenarios. Even when they’re not spot-on, their analysis helps in understanding the enormous trove of weather information available to sailors today. 

On the other hand, sailors who wing it with nothing more than, say, a NOAA forecast and some hunches often seem to get along quite well. 

If a crew starting the 2016 Bermuda race had ignored all weather advice and decided to simply sail the shortest distance between the start and finish lines, they would have sailed the perfect route. The good old rhumb line had the most favorable wind velocity and angle.

You can find evidence of that in the most appealing story of the 2016 Bermuda race. The second boat in the entire fleet to cross the finish line and winner of its class was a comparatively old boat less than half as long as the first boat to finish, the 100-foot Comanche (which broke the elapsed time record by more than five hours). It was a borrowed Tripp 41 entered by the Young America Junior Big Boat Sailing Team of American Yacht Club of Rye, New York, sailed by a crew that included seven teenagers 15 to 18 years old. I don’t know what kind of weather advice they had, but the boat’s track pretty much stuck to the rhumb line.

I love all the weather data sailors can immerse themselves in these days, but sometimes you have to label it TMI—too much information.