“When it came to getting dressed to play roulette,” said Ingrid, “there was a real ritual, like being a secret agent. To get everything on and hooked up took an hour. Then you conducted tests, and something invariably wasn’t working. So you had to strip down all the way to your antennas. You tried to keep everything in discrete units, so there would be parts of the system and little piles of clothing scattered all over the room. People would be completely serious and upset about something not working, but they’d be standing there in their underwear clicking the computer with their toes. It was ridiculous.
“I had trouble finding a costume because the batteries were so bulky. I scrounged a pair of burgundy pants out of the basement and borrowed a top from Lorna—a floral print on rayon that crossed over in front with a big sash, like a Japanese kimono. I got dressed in the following order. First I put on a Playtex midriff bra that hooked in front. It had twenty hooks, and I invariably missed one in the middle; so I’d have to start hooking them all over again. The sacroiliac belts would have flattened me out completely, so for holding the computer and batteries, I stitched pockets onto the bra. Then I stuffed them with washcloths to fill them out. I figured if I were stacked I would look better, but mainly I needed to hide the corners. I also tried to keep the equipment away from my skin, because if I sweated I had a real problem with getting shocked, which was no fun at all. To prevent this, we later wrapped the computers in plastic bags. Then I put on the antenna T-shirt, which had connectors running down the front to the computer. Another set of connectors ran from there down to the solenoid plate, which I wore on my stomach under a red and white polka dot girdle.
“I looked like a battle-ax on top. What made it even funnier was that I just didn’t move in any natural way. The wraparound blouse was supposed to be flamboyant and sexy and loose, and I was wearing it loose, but I didn’t want to reach over the table too far because someone might see my corners. I even took to ratting my hair, because I thought it would make my head look bigger and my body smaller. Thinking I’d try to fit in, I wore mascara and blusher and carried a handbag. When I was all dressed up, I was in an altered state of consciousness, and I was always cold, because when I’m nervous, even if I’m sweating, I get the chills.”
“As you can imagine,” said Doyne, “with so many roulette freaks at large in the house, the pandemonium was incredible. Norman was staying up all night working frantically on the receivers. At one point he said they were finished and took off for Portland to see Lorna, but I suspected they weren’t really working. Jim Crutchfield was trying to debug Patrick and coming up with glitches all over the screen, until he later discovered one bad bit in the PROM. People were finishing end-of-the-quarter exams. It was crazy. Very crazy.
“Like always, we were way behind schedule. John Loomis had to pull out, and if we’d been smart, we would have scrapped the spring trip. But I felt as if we had built up so much expectation that we had to get out of town and do something different from living on top of all this stuff.”
With Patrick still on the fritz from the New Year’s trip to Las Vegas, and only Harry available for practice sessions, the Projectors decided to pack up the wheel and finish training themselves in Nevada. Riding in the Blue Bus—converted as it was into a mobile electronics workshop and casino—were Dan Browne, Charlene, Ingrid, and Doyne. Alan Lewis and his girlfriend, Molly, drove up in another car, while Marianne had gone ahead to meet Ralph Abraham at what he called his “cabin”—a three-bedroom condominium with Jacuzzi that he rented for the winter on the north shore of Lake Tahoe.
At the wheel of the Bus, speed-rapping with Dan Browne about cybernetics, Charlene circled twice around Hayward and got hopelessly lost outside Davis. Four hours later than expected, the Projectors reached Ralph’s “cabin” in the middle of the night. “They rolled up and took over the place,” said Marianne. “In no time they had the wheel set up and the house overrun with paraphernalia.” They spent two days training themselves, and then late on the evening of the second day Doyne, Ingrid, and Marianne drove across the state line to Reno. With Patrick not yet debugged, and Norman’s radio transmitters on the blink—which left the computers Renata and Cynthia unusable—Doyne would have to play solo with Harry.
Ingrid had never been inside a casino before, although she remembered when she was young that her mother on trips into the mountains used to stop in Tahoe to feed the slots. Ingrid would wait in the parking lot, and once her mother had returned to the car with a purse full of coins. Even though she had never seen the real thing, Ingrid knew the layout of a roulette table cold. Working from a cloth model painted by Alix for her own maiden trip to Reno, Ingrid and Marianne had developed lightning reflexes for covering the layout with chips. “To pick up the solenoid buzzes and get chips on four numbers at once, you had to be really fast,” said Ingrid. “Even for three numbers you had to be quick. But we got good enough to do four numbers without any problem.”
The Blue Bus pulled into Reno on a cool night in April. “We split up to check out casinos,” said Ingrid. “We would meet later to compare notes on wheel tilt and other conditions for playing. Walking into my first casino—I don’t remember which one—I was trying to act real casual, but I was so nervous I couldn’t even get up the nerve to go over to the roulette tables. I was hanging around the slot machines, trying to watch from there, and the house cops must have thought I was underage and loitering. Two of them came over to ask for my ID, and I wasn’t carrying one. So they threw me out for being underage. I felt terrible. If the first time I go into a casino I get kicked out, I figured I just wasn’t making it in the Project.”
When they regrouped later, Doyne reported finding a good wheel at Harold’s, the biggest and flashiest of the downtown clubs. The three of them walked up to the roulette tables and Doyne bought into a game. He would play alone as data taker and bettor, while Marianne and Ingrid compiled statistics. “We did our best to look glamorous,” said Ingrid. “Marianne was wearing makeup and trying to use her shoulders. I was dressed in a rabbit fur coat, and we were hanging on Doyne’s arms like a pair of groupies. Doyne was nervous and staring intently at the wheel without being able to say a word. It wasn’t obvious why anyone would hang on to that sort of guy. We stood around trying to look casual, and then Marianne and I would giggle and run off to the bathroom to pull out our notebooks and write down data.” At four in the morning, down a few dollars and getting shocked by a flaky computer, Doyne called off the session. They drove back to Tahoe and reached the north shore of the Lake at dawn.
Stripping off antenna T-shirts, sacroiliac belts, solenoid plates, pimp shoes, midriff bras, wires, and switches, they jumped into the hot tub outside Ralph’s condominium. They watched the sun come up and turned pink in the water and pinker still when they got out to roll around in the snow. Later in the morning, after loading the wheel and computers into the Blue Bus, they made their way over Donner Pass and down the Sierra Nevada to Davis, where they stopped for a pizza. Inside the restaurant, after staring into his water glass for a long time, Doyne reached into his pockets and pulled out all his money and keys. Handing them to Ingrid, he said, “You take care of these. From now on, you make all the decisions. I’m not taking any more responsibility for the rest of the trip.”
“Up to that point,” he said, “I had been trip leader and camp counselor. I had been coping with everything from sewing holes in my socks to rewiring computers. I was totally fried.”
8
Exploring the Envelope
Lest men suspect your tale untrue
Keep probability in view.
John Gay
Alan Lewis was teaching a graduate course in electricity and magnetism when he and Norman, chatting after class one day, first talked about roulette. Of medium height, with brown hair and eyes, Lewis is given to understatement. Comprehending his jokes demands a minimalist aesthetic. He speaks in a monotone so precise that one sometimes feels the urge to record his voice and play it back at faster speed
. But Lewis knows how to relax in a world in which there are more days you can drive your car to the beach with the top down than not. Born in Tucson and brought up in various parts of California, he is a westerner of the laid-back and unassuming sort.
He is also an expert in statistical mechanics. This is the branch of physics that attempts through mathematical models to analyze motion and the forces that cause it. Developed in the late nineteenth century by James Clerk Maxwell, Ludwig Boltzmann, and Josiah Gibbs, this statistical approach to physics—which originally described the atomic behavior of gases—has lately found new applications in gambling theory. As far as Lewis is concerned, the most interesting area of research for statistical mechanics is the stock market.
“In playing the stock market,” he explains, “the trick is to take the gambling idea seriously. One is then led to a wealth of ideas from the well-developed mathematical disciplines of probability theory and statistics.” With his theories on how to model the stock market like any other physical system, Lewis eventually left the university and moved to Newport Beach, California, as a stock market analyst. It is not your average stock broker who is interested in employing a physicist, and Edward Thorp once again reappears as a figure in this story.
In the fall of 1978 Lewis went to work for a small company called Analytic Investments, which manages $500 million for customers such as New England Telephone and Yale University. Located in one of the suburban haciendas strung along the freeways of Orange County, AI occupies a modest seven-room suite decorated with seascapes and wandering Jews maintained by a plant rental service. The company is run by Sheen Kassouf, a courtly New Yorker of Lebanese descent who specializes in playing the options market. With a degree in economics, Kassouf had hung out in New York for many years as a self-professed “boardroom bum” who spent his days camped in front of ticker tapes while looking for the American El Dorado: a stock market system that would allow him to win whether the market went up or down, on bear and bull days alike. He finally found what looked like a winning strategy through the use of bets known in stock market parlance as hedges.
Thinking along similar lines as Kassouf, and arriving at the same conclusion, was Edward Thorp. Then a professor of mathematics at UC Irvine, Thorp had moved on from blackjack and roulette to look at schemes for playing the stock market, which he viewed as another favorable game whose odds, under certain conditions, run strongly in favor of the informed bettor. Thorp and Kassouf met in 1965 when the latter was being interviewed for a job at Irvine. Sharing their ideas, they worked up a winning stock market system based on warrants and other kinds of hedges.
“I got interested in taking my gambling winnings and going into the stock market,” said Thorp, “because it seemed like a larger-scale gambling game with a lot of extra interest to it and a lot more potential. I wouldn’t run into the problems I had in Nevada with blackjack cheating and that sort of thing. If there was cheating, it wouldn’t be directed at me personally, but at everybody who happened to be playing. So I did a lot of reading, and in the summer of 1965 I came across the idea of warrant hedging. I met Sheen as a prospective faculty member and found out almost immediately that he had the same idea, and had actually been doing some of this for a few years in a fairly crude but effective way. So we put our heads together and tried to improve the product, which developed into Beat the Market.“
Written as a sequel to Thorp’s first book, the work he co-authored with Kassouf is titled, in full, Beat the Market: A Scientific Stock Market System. As Thorp put it, “Our general styles of doing things are somewhat different,” and one senses this disparity even in the title of their book. The soft-spoken Kassouf contributed the “science” and “system” to the subtitle. But when it comes to “beating”—whether it be dealers, markets, or other running dogs of the capitalist system—Thorp’s your man. He walks on his toes with the nervous lope of a street fighter looking for action. Others have seen in him a resemblance to Clark Kent. Remove the glasses and briefcase, and out of his office could emerge a muscleman from the beaches of Santa Monica, or a high roller just in from Vegas. Tanned and nervous, clothed in short-sleeved shirts and a full suit of what Wilhelm Reich called character armor, this eminent academician and winning gambler knows as much about tenure committees as he does about casino mobsters, and he probably thinks of them in similar terms. Thorp has never adopted the intellectual’s disdain for worldliness. He wants to be rich and famous, and he makes no bones about it.
After their book appeared, Thorp and Kassouf split up to incorporate themselves into companies. Kassouf went public, while Thorp the high roller looked to private individuals for capital. He started a hedge fund called Princeton Newport Partners, named after the two cities in which its principal offices are located. To avoid governmental regulations, Princeton Newport works with unlisted phones and keeps its membership below a hundred investors. “Few and big is what we have to have,” said Thorp. “People come banging on our door, and if they have large enough bankrolls, we take them in.” He is vague in public about its assets and actual gambling strategy, saying only that Princeton Newport has accounts in the “many tens of millions” and a return on investment of 20 percent a year.
It was during his final year of teaching at UC Santa Cruz, before he headed south to Orange County, that Alan and Norman discovered that Lewis himself, back in his undergraduate days at Cal Tech, had tried to predict roulette by computer. In 1972 he had loaded a camera case with silicon chips mounted on a printed circuit board. Built before microprocessors were generally available, Lewis’s computer required numerous chips to perform its various functions, and it operated at a lower level of sophistication than the machines later built by Eudaemonic Enterprises, but it probably ranks as the first digital computer played in a casino against roulette. Powered by lead acid batteries and operated by two buttons sticking out from the bottom of the camera case, the device gave its predictions by means of a diode display that flashed through a window cut into the top of the case. Like the Eudaemonic system, Lewis’s employed a radio link between data taker and bettor. One player manipulated the buttons and then whispered the computer’s prediction into an FM transmitter built into his tie clasp. The second player picked up the prediction through an earphone connected to a cigarette pack that was actually a radio receiver, and placed the bets.
Lewis found the best place in Las Vegas to play his computer was Circus Circus. From a second-floor balcony under the high-wire act, he could look down onto the roulette wheels, enter data, and broadcast signals to a partner standing below him at the tables. The program for their computer came from trial-and-error approximations fit on a case-by-case basis. From his balcony overlooking the gambling floor, Lewis in advance of a playing session would clock a roulette wheel with his computer, accumulate data on a miniature paper-tape print-out, and then return to his hotel room to fit a curve to the wheel. After programming the curve into the computer’s memory, he and his partner would head back to the casino to play roulette.
“We did some betting,” he said. “We didn’t make any money, but everything functioned.” Lewis’s problem, he realized later, lay at the level of theory. At the time he had no idea about the variability of roulette wheels, particularly the differences in their degree of tilt. “Our system wasn’t really tested on anything sensible. The existence of tilted roulette wheels is crucial to having a predictive system that works, and we had no idea of it at all. We made other errors and suffered generally from lack of time.”
On hearing Alan’s story, Norman told him he was working on a more sophisticated version of the same idea, and he invited him over to take a look. “I was surprised to find a miniature commune of physicists,” said Lewis of his first visit to the Riverside house. “They were a really interesting and impressive group of people. I liked them all. Just to see a bunch of crazies working on something like this was pretty amazing. I had absolutely no reason to suspect that anyone in Santa Cruz was doing this sort of thing.
&nb
sp; “Even the fact that they had such a good roulette wheel was impressive. I knew how expensive they were, and I could tell from that how seriously they were taking the Project. The wheel was a work of art, although by then it had a story behind every burn and scratch mark on it. Given all the time we spent staring at it, the wheel would become for us a great mandala inviting everyone to bow down and worship it.”
On joining the Project, Lewis took on the assignment of figuring out a Eudaemonic betting strategy. How big a bank did they require? What percentage should they bet on each play of the game? Should they place their bets on one or more than one number at a time? And what conditions—in terms of rotor speed and tilt—were tolerable for playing a winning roulette computer? To answer the last of these questions, Alan and Ingrid spent several hours a day hunched over the “great mandala” set up on the picnic table in Doyne’s bedroom.
“We ran the computer through thousands of trials and compiled success histograms every afternoon,” said Ingrid, “before Alan went to the beach. Of all the people I worked with on the Project, I liked being around Alan the most. He had a sense of humor about it and was more relaxed than anyone else. He never viewed the Project as a place to do his stuff and be a star, which is how a lot of us saw it, including me.”
On finishing their research, the Lewis-Hoermann team presented their findings at one of the ad hoc Project meetings called whenever anyone had anything interesting to say. “Thorp calculated an advantage for a computer playing roulette of over forty percent,” said Lewis. “But the odds are always changing. Sometimes they’re higher than that. Sometimes you’re actually losing money. And a lot of it’s out of your control. But there is in fact a set of ideal conditions for playing a roulette computer. You want a croupier who spins the inside wheel slowly and the ball very quickly, so you have a lot of time to make measurements. You want the table relatively quiet, not hectic, so you have room to place your bets. And you want these conditions to persist for a period of time. This is not an unusual set of requirements, but you’re dealing with an uncertain environment. When the conditions are ideal, the odds are quite high in your favor. Who knows what they are? Twenty percent or forty percent or ten percent? You may not know the exact figure, but what you do know is that any advantage of this order, if it persists, will make you a fortune.”
The Eudaemonic Pie Page 16