High tech hits the high seas in a windblown battle between Craig McCaw and Larry Ellison. Carl Hoffman sets sail with Team OneWorld in the race to take back the America's Cup.
There are few pieces of technology simpler than a sailboat. It is, after all, among the slowest vehicles on earth, and its basic design has changed little in thousands of years. Hand Christopher Columbus the tiller of today's hottest racer and he'd sail off to America, no problem.
Sharon Green. Two Team OneWorld racers, dubbed USA-65 and USA-67 for the America's Cup, sail in the Hauraki Gulf off the coast of Auckland, New Zealand.
It took Columbus 10 weeks to reach the New World. By 1905, thanks to better maps and streamlined hulls, the record for crossing the Atlantic stood at 12 days and 4 hours, for an average speed of 10 mph. Over the next 75 years, skippers carved just 2 days off that record. Then something remarkable happened. As the 20th century ended, a surge in data-processing muscle and the advent of carbon fiber gave naval architects the power to craft dramatically lighter hulls with less drag and more efficient sails. At the same time, more accurate weather models told sailors how to exploit optimum wind currents.
In October 2001, millionaire adventurer Steve Fossett's PlayStation put it all together, racing from New York to England in 4 days and 17 hours, shattering the decade-old record by 44 hours. Fossett sailed at an average speed of nearly 26 mph, a pace even most blue-water powerboats can't match. Fossett's secret: He waited for weeks on his 125-foot carbon-fiber catamaran until his weather guru said go, then surfed across the pond on the curl of three powerful low-pressure systems. The forecast was so accurate that, in 2,900 miles, he never tacked once.
Sailing, so long an intuitive art, has become a contest of technology. It's a sport where a slight tech edge can truly make a difference — sailboat races are routinely won by seconds. Yet the transition from art to science is proving almost as hard as predicting the stock market. A yacht remains one of the most difficult crafts to model. Its only source of power is the fickle wind. Its sails flex and flap. The boat heels and bounces. The water itself is always moving. It's hard even to take an accurate measurement of the wind or currents.
Nowhere is the science of sailing more extreme than in the quest to win the oldest trophy in international sports, the America's Cup. Beginning October 1, nine teams — from Britain, France, Italy, Sweden, Switzerland, and the US — will race in the qualifying Louis Vuitton Cup in Auckland for the chance to beat defender Team New Zealand in the main event, which starts in February. Held roughly every three years, the Cup is, as eight-time competitor Dennis Conner once said, "a war game, a space race," where history's richest men and biggest egos deploy armies of engineers, mathematicians, physicists, programmers, and weather experts to create a sailboat a hair faster than the rest. This year, the defender and nine challengers will drop a total of about $700 million, with each team spending almost seven times what Conner's spent 15 years ago. For some teams, R&D alone will cost $30 million to $40 million, ten times what it takes to build the boats themselves, and will call on the biggest names in technology, from Compaq to Ford, BMW to Xerox.
The red-hot center of the tech battle is the fight between two American teams, OneWorld Challenge and Oracle BMW Racing. It's Craig McCaw versus Larry Ellison; billionaire versus billionaire; the Seattle Yacht Club versus the Golden Gate Yacht Club. The two CEOs are reaching deep into their technological holds. Oracle touts its "sophisticated design tracking system" that uses database software to retrieve detailed reports about the performance benefits of subtle design changes, which is then shared with engineers and designers all over the world: in California, New Zealand, and Annapolis, Maryland. For his part, McCaw, the pioneer in wireless communications, outmaneuvered the competition early on by buying up the high-capacity wireless LAN spectrum that covers Auckland and its waters, leaving Ellison and the other teams with the indignity of communicating by much slower radio and cell phone. McCaw then raided Team New Zealand, hiring six of its best sailors, along with designer Laurie Davidson, a national hero to fellow Kiwis for creating boats so fast they humiliated the
US in 1995 and Italy in 2000. Says McCaw: "We did things so stealthily that Larry thought the Seattle challenge had to be coming from Bill Gates and the Microsoft guys."
The Cup is a war game, a space race — and a tiny tech edge decides the outcome.
Sharon Green. OneWorld Challenge under full sail.
It's a rainy 50 degrees, the beginning of the austral winter on the blustery waters of New Zealand's Hauraki Gulf. The qualifying rounds won't start for another five months, but OneWorld is testing two of its yachts, designated USA-67 and USA-65 by the America's Cup technical director. While they look roughly like any other sailboat, everything about them is calculated to precisely capture the wind's capricious power. Eighty-five feet long, their one-piece carbon-fiber hulls weigh less than a Volkswagen Beetle yet support a 110-foot mast and a narrow, 14-foot-deep keel dangling a 20-ton lead bulb. Sleekly gliding over the water, they accelerate rapidly when the slightest breeze hits their 6,500 square feet of sail, and they turn as fast as bobsleds. Sensors and digital cameras feed real-time performance data and GPS coordinates across the wireless network to the crew and to OneWorld's land-bound designers and engineers. A second computer network sucks in meteorological data from five weather boats, dozens of weather stations, and OneWorld's proprietary neural net. The goal: to predict the tiniest wind shifts. Custom software aboard the yacht calculates everything from the fastest sailing course to the average speed and angle of the wind up and down the mast.
The 65, which flies an oddly shaped new mainsail so secret that I'm forbidden to describe it, is going head-to-head with 67 to determine which boat's mainsail is faster. I'm following close behind in a 60-foot chase boat with OneWorld designer Phil Kaiko and engineers Masanobu Katori and Richard Whitaker.
The mainsail's shape is so secret that I'm forbidden to describe it.
Suddenly Kaiko, 42, grabs binoculars and scans the horizon. Wearing blue shorts and a gray sweatshirt, with a boyish face and shaggy hair, he could pass for a college student.
"Anybody know who that is?" he asks, as an inflatable orange Zodiac speeds within range. When the Kiwis beat the US to win the Cup in 1995, they vowed to clean up what had become a game of relentless espionage among teams. Now "persistent observation" of another team's boats is explicitly forbidden until two months before the first races. The team bases in Auckland are patrolled by security guards and closed-circuit cameras; photographs anywhere within a team's compound are verboten; and when 65 and 67 are withdrawn from the water each night, their hulls and keels are shrouded by gray skirts. "Every team is curious about what the other is doing," Kaiko says. "You try to psych them out, to make them feel like they can never do anything without being watched."
Sharon Green. The brains of the OneWorld operation (from left: Kaiko, Whitaker, and Katori) crunch performance numbers aboard a chase boat.
Eventually, attention on the boat returns to 65 and 67, whose sensors and strain gauges are tracking 200 different parameters every second and sending the information across OneWorld's LAN to its chase boats and offices. Then the info gets dumped into a Microsoft SQL database, where it's sifted to pinpoint the effects of sail and hardware experiments. Unraveling all the input is, in the words of OneWorld engineer Richard Karn, "nearly impossible." Take the keel's trim tab, an adjustable surface that's analogous to the flap on an airplane wing. Simply finding the optimum trim settings for each wind speed requires comparing each flap setting against all others, using one boat as a control. The combination of settings and wind speeds is mind-boggling.
Making a sailboat fast is unlike making any other kind of vehicle fast. Cars and airplanes, for instance, operate within a single medium, but a boat plows through air and water, creating what's known as "free surface" between the two that dramatically raises drag. Every parameter of a Formula One car can be targeted and modeled, including the track on which it's racing. Airplanes can be so well modeled that some engineers even argue against building prototypes. And cars and planes are all powered by engines with precisely known and controllable horsepower. "Our power comes from the wind," says Kaiko, settling down once again at his laptop, as we roll gently in the swell. "It's variable in height, speed, direction, and density, and it's totally erratic and random." Ditto the track on which they race — waves and currents. And the yachts require not one pilot but 16 crewmen, each of whom is an independent variable.
OneWorld's designers began hammering out hulls almost immediately after the 2000 Cup. Unlike yachts in races between identical sailboats, the International America's Cup Class yachts are created around a complex formula that lets designers trade off sail area for length and weight. The formula has been the same since 1992, so making a boat faster than your competitors takes ever more powerful technology for even the smallest improvement. "Yet, if you can make your boat even a fraction of a second faster," says OneWorld designer Bruce Nelson, "that'll give your team the edge it needs. So the question is, How do you find that?" He answers himself: "If you really want to explore a boat, you need to run a 200-point matrix, and each point has to go through up to 100 iterations, so running one hull might take a day or two." OneWorld worked through hundreds of possible hulls to find this year's shape, a design as closely guarded as the curves on the Pentagon's stealth aircraft. "In '92, we looked at dozens of hulls, but none had this one's accuracy and sophistication," Nelson says. "Much more physics has been added: Now you can predict the performance down to a hundredth of a knot." Oracle, for its part, riffled through thousands of hull configurations on a Compaq AlphaServer SC supercomputer.
Sharon Green. A OneWorld crewman checks weather data via a wireless touchpad.
But tactical decisions intrude on what otherwise might seem a straightforward analysis: Do you want a hull that functions better upwind or downwind? Primed for 15 knots or 8? Less drag or more stability? "There are so many trade-offs that you have to be really smart," Kaiko says. The penalties for erring are high, only the least of which is ending up with a slow boat. "The wind loads on the rigging and hull are huge," says Kaiko, "and they're predicted, not known, so equipment breaks and people get hurt." In the 2000 race, the Australian yacht (designed by Kaiko) broke in half midrace. So far this year, two of Oracle's training boats have lost keels, and OneWorld lost a mast. Dennis Conner's boat, Stars & Stripes, sank.
After 10 minutes of straight-line sailing, 65 is two lengths ahead of 67. A decisive victory for the radically new rag, I suggest. "Maybe," Kaiko says, checking the numbers flowing in. "The winds are really too light and shifty, the data too inconsistent."
Suddenly there's the Zodiac again. The radio crackles, and in seconds the sail comes down and a more conventional one goes up; then another test begins. So it goes for the next eight hours, as the 65 and 67 race each other using various combinations of sails, and the data pours in. Every few minutes a crew member snaps photos of the sail from digital cameras mounted on the deck, to be scrutinized later. Periodically, sailors swap boats, to better distinguish between a yacht's performance and that of its crew. Ideally, over the course of the day, speed advantages can be traced to one particular sail, which can be tweaked yet again. Over months, the fastest permutations of sails and trim will be studied. Conclusive data is hard to come by, however. During any given moment of any particular test, the vessels weren't even sailing in the same wind despite being a mere hundred yards apart.
"It's a statistical process," says Katori, the team's lead programmer, as we take the boats in tow and head back to shore at the end of the day. "You have to build a lot of very subjective data before it begins to mean anything, and that's especially true in light wind. But over time you do build real numbers."
On race day, those numbers become a tactical weapon. In 2000, the yachts were so closely matched that 80 percent of the time the first boat to cross the starting line sailed to victory. Devising a plan for getting across the line first falls to navigators like Kevin Hall and Mark Chisnell, for whom turning the vagaries of wind and sail into hard data is a relentless puzzle. That's because the start of a sailing match is unlike the start of any other kind of race. You can't, after all, just poise on the edge of the line and press the accelerator when the gun fires. Instead, in what's known as the prestart, boats duel during a wild five-minute process of jockeying for the best wind relative to the starting line, while denying that same advantage to opponents — who are trying to do the same — all while circling and tacking and jibing in constant motion amid variable winds in a finite space aboard an engineless vehicle. And all that technology? Forget it. America's Cup racing is ruled by the Corinthian notion that a yacht at sea should be on its own, so at prestart every craft must sever its connection to the outside world. Cell phones, two-way radios, and the wide-area network hardware are sealed in a special bag and tossed overboard. Transmission of those all-important weather and wind forecasts ceases. What's left is the onboard network that links two Panasonic Toughbooks tucked under the deck; one holds predicted weather data, another keeps performance parameters for various wind speeds and directions. The computers feed select pieces of data to small monitors at the base of the mast and at every sailor's station on the boat.
The neural net sucks in data right up to the prestart. Then the umbilical is cut.
Hall, an intense red-haired 33-year-old from Ventura, California, graduated with degrees in math and French literature from Brown University, and twice failed to make the US Olympic sailing team by a whisker. Chisnell, 40, was a member of Britain's 1987 America's Cup team. Known as Chizzy, he has degrees in physics and philosophy, and, between yacht races around the world, has written seven books on sailing and a novel for Random House. The morning after the long day of sail tests, I find Chizzy and Hall at OneWorld headquarters, in an office littered with two-way radios and foul-weather gear, brainstorming with Katori over how to write a piece of software that can more accurately advise the skipper and tactician as they maneuver during the prestart.
In a perfect prestart, you fight for position, sail away from the starting line, turn, and then sail back under full power, crossing the line just as the gun fires. "It all comes down to time and distance," says Hall. "The boat's computer can tell us where we are in relation to the line, and it can guess how fast we'll sail to it based on boat speed, but there are a lot of assumptions that you need to make that may or may not be correct." Assumptions, in fact, about nearly every aspect of the boat's performance. During the wild prestart gyrations, "you may never go in a straight line," says Chisnell. "Boat speed is measured by a little paddle wheel in the water, but you might be going sideways and so it won't work at all. Same with the compass heading. You might be pointing 180 degrees but actually going backward. You're tacking and jibing constantly, swinging through all the points of the compass. And all these instruments deliver data at different rates and cope with the violent motions of the yacht differently, so the inputs get out of whack and you get spikes in your data."
Sharon Green. OneWorld navigator Chisnell aboveboard with a laser range finder.
Hall leaps up, excited. "And when you're heading away from the starting line," he says, "every second the question is, Should we turn around?"
"But how long does it take to turn around?" interjects Chisnell, using his hands like two boats fighting for position. "Is it better to make, say, a tight 10-meter-diameter turn that takes less time and burns off boat speed, or a wide 50-meter turn that takes longer and keeps your speed up?"
"Ideally," says Hall, "at any point in the prestart you'd be able to tell the tactician exactly how long it'll take to turn around."
They lead me over to the shed for a tour of the yachts. As we walk past the gray-skirted boats, Chisnell warns against peering under the occasional opening and talks about the vexing dilemma of simply measuring true wind speed and direction, the two most important pieces of information in a race. "The navigator is trying to give the most accurate wind information he can to the tactician, who's using shifts in the wind to get the boat to the next mark. We've got this huge mainsail, which distorts the wind as it flows over the sail. The actual angle of the breeze might be 15 degrees different from what it registers at the top of the mast. We've got this 110-foot-high column in which wind direction and wind speed may vary, so the power the wind gives to the rig varies enormously." We climb a set of stairs and step onto a carbon-fiber deck, a small pit crowded with lines. "Should you accept the number at the top of the mast?" he says, squatting by a pedestal that serves as his navigator's station. "The smallest bit of wind makes a huge difference. It's an anal problem, but it's central to winning, and it's been occupying me for two years."
To solve it, Chisnell has been experimenting with mechanical and sonic anemometers placed up and down the mast, and even on an adjustable 25-foot pole at the stern. Katori is writing a program that crunches the measurements and creates a "wind profile number — an implied wind," in the words of Chisnell, that the navigator can base his calculations on. "You've got so much information, and so little of it is precise, and you've got 16 guys trying to sail a powerful boat that you can barely control, all in, say, 20 knots of wind that is trying to tear the rigging out of the boat. If you can nail down some of those numbers, you can gain a huge advantage."
If there's any area where the sailor's art has long prevailed, it's anticipating the weather. An experienced eye can spot gusts blowing across the water, and even predict how wind will rise, drop, or change direction when a cloud passes over the sun. That's extra valuable in the Hauraki Gulf, where the winds are especially strong and shifty.
Sharon Green. A shroud conceals the hull design from spies when out of the water.
While intuition remains critical, there's been big progress in the area of weather analysis over the past 10 years. "The percentage of the budget being spent on weather has increased more than any other item," says weather padrone Hamish Willcox, surrounded by computer monitors and maps of the Gulf covered with little red dots. Willcox, 40, tall and tan with sleepy green eyes, might be OneWorld's Willard Scott, but like each member of the team, he's an elite sailor who started young. At 17, he won the first of three world championships in Olympic 470-class small boats. He raced in the 1996 Summer Olympics, and then coached the New Zealand national team for five years. Just as important, he's an Auckland native, bottle-fed on the Gulf's erratic winds, and a veteran of Team Prada's 2000 Louis Vuitton Cup victory.
Willcox points at a map on the wall, then hits a few keys. Up on the monitor pops a chart of the Gulf and the surrounding land, covered with small markers in a rainbow of colors. All told, 57 weather stations and buoys surround the Gulf (most are exclusive to OneWorld, but some are shared with other teams), continuously milking wind direction and velocity, temperature, barometric pressure, and humidity, and other variables Willcox refuses to mention — a dizzying wealth of information pouring into OneWorld's weather database. That's not all: Every day for the past two years, five OneWorld weather boats have headed out into the Gulf to harvest data. On race days, they are stationed around the course — one at the upwind mark, one at the start, one on each side, and another deployed as needed.
The key to the forecast is a proprietary neural network developed by the Ford Research Laboratory in Dearborn, Michigan. "The basic idea," says Lee Feldkamp, a Ford scientist, "is to take the information measured upwind and predict what will propagate down the course." Every day, the mountain of weather data collected is fed into the program. "If you have, say, 12 knots of wind moving in a certain direction and you go downwind from that point 300 feet, what happens?" asks Willcox, peering over his office dividers at the weather outside. "Does the air rise? Fall? Dissipate to the left or right? That's the mystery we sailors are forever trying to solve." And that's where the neural net can help, says Feldkamp: "We're trying to find patterns, to see that one set of conditions tends to result in something else. We don't know why, and we don't need to, because the answer is in the data." The program runs continuously, sucking in information from all the stations and buoys and weather boats, interrogating it right up to the prestart; and then, the moment the umbilical is cut, the computer spits out how the wind is expected to shift over the next seven minutes.
The wind, which today has been blowing too hard, improves enough so we can set sail. Peter Gilmour, OneWorld's skipper, herds us out to the boats. As we're towed out to the Gulf behind the chase boat, Gilmour, a 42-year-old native of Perth, Australia, and a five-time America's Cup veteran, perches on the gunwales, the occasional cold drop of rain smacking his ruddy face. "Sailing is a strange sport," he says, bracing himself against a wave. "At one end you need physical stamina, and at the other mathematical ability. Technology is extremely important. It helps you find that sweet spot and builds confidence. But on the racecourse, it all comes down to a feel thing, from the experiences you've been able to build up."
Which is evident the moment we drop the towline and start sailing in gusty 15-knot winds. There are two boats to test and two teams of 16 sailors to hone into machinelike crews. Every moment on the water counts; even if conditions are marginal, the yachts go out. Cup boats are so high-strung, however, that even in relatively light winds they are on the edge of chaos. Gilmour calls out the gusts he spots coming our way, shouting over the eerie wail of carbon fiber under tension. Each tack is a sudden and violent change of direction that leaves the squad of finely tuned athletes frantically hauling lines, grinding winches, and clinging for their lives. Over on the comfortable chase boat, Kaiko and the engineers are gently rocking on the waves, watching raw data pour in. Aboard 67 it's all unbound energy; art and nature challenging science, a foot of water surging over the transom.
Then a carbon-fiber batten — a slat used to stiffen the mainsail — breaks. The winds rise and great rivers of cold sweep across jade-colored waves. The skies open up, a frigid winter rain that even custom Gore-Tex can't keep out. There's nothing to do except give up for the day and head home.
As we're towed back to the dock against the current, waves of ice-cold spray pour into the boat. One by one, the crew drops through hatches fore and aft. Finally, I join them belowdecks, leaving only Gilmour and James Spithill, the two helmsmen, at the wheel. An America's Cup yacht is no pleasure cruiser. Below, it's pitch dark, with no place to sit — nothing but raw carbon fiber, ropes, wires, a slanting tub — barely an inch and a quarter thick — banging on the waves. Fourteen sailors, the lowest-paid earning close to $80,000 a year, huddle together with me like wet rats in a barrel as seawater runs down our backs. The rain and wind have won. This is sailing, after all, and all the technology in the world will never change that.
