Anatomy of a Point Spread

in the trenches with the master bookies of las vegas

The week before the Super Bowl, the National Football League will cite some staggering statistics to demonstrate the sport's popularity: The game will be seen on television by three quarters of a billion people, including almost half of the population of the United States. Not everyone will be watching for the sheer enjoyment, of course: The day's wagering, legal and illegal, should reach about three billion dollars. And at the heart of it all will lie the numerical Holy Grail of sports: the point spread between the Super Bowl combatants.

Conceived by a Midwestern bookmaker in the Forties, the point spread transformed even the worst athletic mismatches into intriguing, bettable contests. If the 49ers, say, were playing the Broncos, few might bet Denver to win. But a point spread favoring the 49ers by eight and a half would stimulate action on both sides. People who bet on the 'Niners would win only if the favorite won by nine points or more; those who bet on the underdog would collect only if Denver lost by eight points or fewer or won the game outright. Bettors lay odds of 11 to 10, the difference being called the juice, or vigorish, or "vig"--the margin of profit that enables bookmakers to drive Cadillacs.

As the money flows in the week before a game, Las Vegas becomes the nerve center for the entire nation. Although odds in Las Vegas influence the nation, ultimately, it's a case of the tail wagging the dog. The magazine Gaming & Wagering Business reported that in 1989, legal wagering in Nevada on sports totaled 1.4 billion dollars. Americans' illegal betting on the same events was estimated at 29.5 billion dollars.

The surging popularity of sports betting isn't hard to understand. It's a fair gamble: Those 11-to-10 odds (bet $110 to win $100) calculate to a 4.5-percent edge for the bookie, a reasonable figure compared with horse races, in which the tracks take 17 percent or more, or lotteries, in which the state confiscates 50 percent. Betting on sports can reward skill--as opposed to, say, craps or roulette. And it makes a perfect marriage with television. Bet on a game that might otherwise put you to sleep and you have three hours of riveting excitement.

•

A good example of how the system works--and its crucial place in the sports culture--is the September 23, 1990, game between Tampa Bay and Detroit. Even in the third week of the N.F.L. season, this is considered a low-intensity conflict between two chronic losers with little charisma. Yet on this evening, 3,200,000 households and virtually all of the barrooms in America will tune in to see the Buccaneers face the Lions. Even in cities with no allegiance to either team, those barrooms will erupt in cheers when a score is threatened.

This is no surprise. In Nevada, more than $3,000,000 will be legally wagered on the Bucs and the Lions--the largest amount on any game this day. Nationally, illegal wagers on the game will probably exceed $100,000,000.

Although this is the season when the N.F.L. forbade its television announcers from mentioning the point spread during pregame and game broadcasts, most of America is aware that Tampa Bay is favored by three points. The number has appeared in hundreds of newspapers and is routinely cited by local sportscasters. Anyone who has taken part in an office pool, made a friendly bet or telephoned his local bookie has had to decide whether or not the home team can win by more than a field goal. However, not many of them have an answer to the pivotal question: Why three points?

The process to determine that figure--indeed, all of that week's point spreads--begins the previous Sunday in a cluttered office just off the Las Vegas Strip. There, Michael "Roxy" Roxborough and his staff spend the afternoon watching all the games on television while referring to teletype and computers for reports on the rest of the day's action. Almost as soon as the afternoon games are finished, Roxborough's company, Las Vegas Sports Consultants, will have to release opening point spreads for the next week's schedule to clients that include most of the major bookmaking establishments in Nevada.

Roxborough went to Las Vegas like so many young new arrivals: as a scuffler who dreamed of making his living as a gambler. In 1975, he was able to parlay an only-in-Vegas skill--success betting on the total number of runs that would be scored in a baseball game--into a career. A bookmaker in Reno took note of Roxborough's astuteness and asked him to help set betting lines on the games, then gave him other sports. In 1982, with fortuitous timing, Roxborough formed his own company. Betting on sports was booming, and the industry had no central source of point spreads.

At the time, the gathering of data and the creation of point spreads were crude processes. Information--especially on college teams--was so hard to come by that odds makers would pay the clean-up crews at the Las Vegas airport to bring them newspapers left on incoming planes.

The betting industry then was propelled by rumor as much as by fact. On one Sunday morning in the Seventies, a gambler named Jolly Joe Sarno received a phone call from a well-informed source who told him that there was a blinding snowstorm in Denver. Sarno immediately began placing bets that the total points scored in the Broncos game would be fewer than 37--the prevailing line. As the money poured in, bookmakers lowered the over-and-under line to 36, to 34, and then to 32 before they finally stopped taking bets altogether. When the Las Vegas wise guys turned on their television sets that afternoon, they saw a crowd of Denver fans in shirt sleeves enjoying a balmy fall afternoon. The score was 21--17 before half time.

It was a time of opportunity for sophisticated, informed gamblers to beat the point spreads, especially in college games. And in the early Eighties, a team of bettors--led by a moonlighting orthopedic surgeon--did just that. Employing computer analysis while odds makers were still using antiquated techniques, this so-called Computer Group bet and made millions.

Roxborough, for his part, brought computers onto the side of the odds makers, enlisting Mike Orkin, a professor at California State University at Hayward, who still teaches Statistics 2088, Games of Chance. Orkin helped Roxborough devise a method of calculating numerical measurements of teams' strength, known in the gambling world as power ratings.

Each team goes into a game with an assigned rating; the difference between those ratings, adjusted by a home-field advantage calculated by the computer, is a raw version of the point spread. When the game is played, one team will cover the spread (and the other team will fall short of it) by a certain a number of points. That's the "error." The error is multiplied by .093 and the resultant figure is added to or subtracted from each team's power rating for the next week.

Suppose the Rams and the Giants start the season with power ratings of 100 each, and they are playing in Los Angeles, where the home-field advantage makes the Rams a three-point favorite. The Giants win, 21--0. That's an error of 24, which, multiplied by .093, equals 2.23 points. This number is added to the Giants' power rating, which now becomes 102.23, and subtracted from the Rams', which drops to 97.77. "It may be simple," Orkin said, "but it's not naïve. That constant of .093 is based on simulations of how the error should be factored in. Other people who make power ratings use passing yardage, rushing yardage and a variety of variables, but from the perspective of the odds maker, they're just muddying the waters."

In the system that Roxborough devised, the maximum error recognized by the computer on any game is 24 points, and the maximum adjustment that can be made to any team's rating in a single week is 2.23. There is a reason that number seems so small: It recognizes the way that professional gamblers--as opposed to the general betting public--look at the results of N.F.L. games. The average fan watching a 40--0 blowout on ABC's Monday Night Football will conclude that the winner is the team of the decade and that the loser will never win another game. And he will bet accordingly the next week. But neither the computer nor the professionals go overboard on the basis of one game. So even after Tampa Bay dominated Detroit 38--21 in their first meeting, the power ratings still had the Lions a better team. Give the Bucs points for the home-field advantage and they figured to be a slim favorite.

What complicates the line maker's task--especially in pro football--is that he must do more than decide who is better than whom. He must gauge the inclinations and biases of bettors. In the September 23rd Chicago--Minnesota game, Roxborough believes that the Vikings are superior. But he also knows that there is huge popular sentiment on behalf of the Bears, and his point spread will have to favor the Bears more heavily to entice people to bet against their hearts. The Tampa Bay--Detroit game is even trickier. Making the home team a three-point favorite would be safe and predictable. But, Roxborough says, "Tampa Bay went to Detroit and clobbered the Lions; now Detroit is going to Tampa. The public says, 'They'll kill them again,' but the wise guys love these reverses. If I were a bookmaker out of town, I'd probably open this game at three. But I'm going to try to rob the wise guys and make it two and a half. That's a big half point, two and a half versus three."

Roxborough has calculated his tentative point spreads for the upcoming games. Now he consults two other experts before deciding on which numbers to release. He telephones Eugene Buonantony, another Las Vegas line maker, and admits that he is still uncertain about the Bears--Vikings game. He is inclined to make the game even; because Chicago has the home-field advantage, that line would reflect Minnesota's superior manpower.(continued on page 140)Anatomy of a Point Spread(continued from page 116)Buonantony dissents vigorously. "I have the Bears one and a half," he says. "Last year, you had the same match-up and the Bears crushed them, 38--7. And the bottom line is that Minnesota doesn't win on the road. They've won two out of their last ten." The men compare notes on each game. On Tampa Bay vs. Detroit, there isn't much discussion. Buonantony is neutral. "Two and a half or three. Either way."

Now Roxborough calls Chris Andrews, director of the Club Cal-Neva in Reno, and Andrews is in a funk. "The squares have been beating me all day," he laments. The teams that the public likes to bet (such as the Bears and the Giants) have been covering the spreads.

Roxborough commiserates, then quickly gets down to business: "Eugene had the Bears one and a half and I had it even," he says.

"I had it one," Andrews says.

That figure seems to be the consensus. "That's what I'll use," Roxborough says. He tells Andrews he and Buonantony are leaning toward a two-and-a-half-point spread on the Bucs--Lions game.

"I knew you guys were going to make it less than me," Andrews says. "I had it three. Tampa Bay's not all that bad."

"All the guys in Vegas are trying to outthink all the other sharpies," Roxborough replies. "We're going to use this at two and a half."

When all of the numbers are finalized, they are sent via teletype to Roxborough's clients throughout the state. In little more than an hour, they will be subjected to their first test. At precisely six p.m., Las Vegas time, betting on the next week's games begins at the Stardust Hotel, the place where America's betting lines originate. Its competitors along the Strip are happy to concede this distinction, because if the line is off the mark, the Stardust will take the initial hits and allow everybody else to make the necessary adjustments.

In front of the board where the point spreads will be posted, rows of men sit in desk chairs, like schoolboys waiting for the bell to ring. Would-be bettors have submitted their names, which will be drawn by lot, assigning each to a position at one of four wagering windows. A participant in this lottery has to bet at least $300; the maximum is $10,000 on a pro game. Bettors can bet as many games as they please, and the Stardust is free to change the line in response.

"Allen J., window one, number one! Rich S., window one, number two! Dave Z., window two, number one!" The gamblers take their positions and start betting, and changes in the point spreads are announced in rapid-fire succession. "Fresno State thirty-two! Bears two!" Someone has immediately concluded that the one-point spread is too low and has wagered the maximum $10,000 on Chicago. "Texas Tech nine! Akron six! Rams five and a half!"

At the top of the town's sports-betting hierarchy, sports-book managers say, are five or six men whose opinions and money command a degree of respect that borders on reverence. "You look at the print-outs of their records," says one manager, "and it's unbelievable, astronomical." When members of this small elite make a serious bet on a game, bookmakers may re-evaluate the point spread.

It is the flow of money that ultimately determines the point spread, causing frequent ripples and occasional tidal waves. The aim of sports-book managers in Las Vegas is to make sure they have roughly the same amount of money on each side of a game. They want to avoid taking risky positions on any game. So, as the bets come in, they adjust their point spread, trying to balance the action.

On Monday, some of the first significant movement of money comes at the Mirage Hotel. The Raiders open as a four-point favorite over the Steelers, and when sports-book director Jimmy Vaccaro taps his computer, he sees these totals flash on the screen: Raiders $33,495, Steelers 0. He will raise the line to five and a half points and knows he may have to go as high as six to attract Pittsburgh money. The one fear that may constrain bookmakers from moving a line too much is the fear of being "middled"--losing both to the bettors who take the Raiders at minus four and to those who may jump in with the Steelers at plus six--which is what will happen to Vaccaro if the Raiders win by five.

The most memorable middle in history occurred in the 1979 Super Bowl, which opened with Pittsburgh a three-and-a-half-point favorite over Dallas and closed at four and a half. When the Steelers won, 35--31, virtually all of the bettors won and all of the bookies lost. "That was like the day Kennedy got shot. Everybody remembers where he was," said Vic Salerno, owner of Leroy's Horse and Sports Place. "I was getting ready to get on a plane when that s.o.b. caught the pass and made it four."

The Tampa Bay--Detroit spread lies dormant until early Tuesday, when wire services report that three of the Lions' defensive starters are injured and will miss the game. None is a star, and the news will not necessarily prompt an automatic change in the point spread. But on Tuesday morning, two sports books, at the Riviera Hotel and at Little Caesar's, take substantial wagers on Tampa Bay. Both establishments promptly raise the line on the game to three, seeking to attract counterbalancing action on Detroit. Having monitored this action, Roxborough's company advises his other Nevada clients, who still have the game at two or two and a half, that Tampa Bay should now be favored by three.

As Vegas goes about making its small adjustments to the line, the nation follows suit--though most of the action is outside the gambling capital. In fact, with a huge illegal market absorbing 25 times as much money as the legal market, gamblers betting large sums may prefer to do so anywhere but Las Vegas. If, for example, a syndicate wants to bet $200,000 on Arkansas minus 12 versus Texas Tech, it can't do so in Las Vegas without causing the point spread to skyrocket. And the change in the Las Vegas number will trigger changes from coast to coast. But the syndicate can lay its money down illegally before the illegal-bookmaking network has had a chance to react and, say, raise the line to 14. When that happens, gamblers will see potential for a middle by betting Arkansas minus 12 in Las Vegas and Texas Tech plus 14 elsewhere. Such arbitrage can account for millions of dollars in wagers.

While the flow of money and information may be orderly before game day, it will be frenetic in the final hours before kickoff. At every sports book in Las Vegas, there will be crowds and long lines at the betting windows. There will even be a line at the $5000-Minimum-Bet window at Caesars Palace. During the last-minute frenzy, bookmakers may be adjusting their lines rapidly in an effort to balance their books.

On September 23, the Chicago--Minnesota game attracts the most interest--and a surge of Bears money. The Mirage's Vaccaro has seen it coming. "The Bears are back in the driver's seat as America's team," he said earlier in the week. "You won't stop the average guy from betting on the Bears. But the wise guys will never let the line go to three points, because they know the game should be even." He's right. Public money drives the line up, but when Vaccaro adjusts it to two and a half, the wise guys jump in on the Vikings. Vaccaro winds up with a perfectly balanced book--about $60,000 on each team. Some of his counterparts aren't nearly so lucky. The Stardust is overloaded with Bears money, and loud cheers erupt whenever the Bears score; Vikings successes bring boos or silence. When the Vikings' punter fumbles a snap in the last 90 seconds and the Bears' Kevin Butler kicks a field goal that enables his team to win 19--16, covering the spread by half a point, the bettors erupt in cheers. Scott Schettler, manager of the Stardust's sports book, is shaking his head, and he is still shaking it at the end of the afternoon. "This was a players' day," he says. "I've got to tell my boss that we lost a hundred twenty-eight thousand dollars for the afternoon."

But the day, of course, is not over. For the same reason that the Monday-night game is the biggest betting event of the week in pro football, the Sunday-night game is becoming the second biggest. It's the chance for winners to press their winnings, for losers to bail out. Unlike the Chicago--Minnesota game, however, the point spread for Tampa Bay and Detroit remains as solid as a rock. Not a single establishment has moved the line from three points. The action is evenly split, with the public favoring the Bucs and the wise guys the Lions.

The two teams slog through three dull periods, managing to score only one touchdown between them, with Detroit leading 13--9. But the final quarter demonstrates why betting on football can be so riveting--and so excruciating. Tampa takes advantage of a fumble recovery to score and go in front, 16--13. Detroit counters by driving for a touchdown to regain the lead, 20--16. With less than five minutes to go, quarterback Vinny Testaverde engineers the drive that gives the Bucs the winning touchdown. The extra-point attempt is irrelevant to the outcome of the game, but it will have a $100,000,000 impact on the nation's economy. When kicker Steve Christie converts, the final score is 23--20, and betting on the game is a push, a chop, no decision. The only people who will cash a bet are the few sharpies who took Tampa Bay minus two or two and a half earlier in the week. Bookmakers will refund everybody else's wagers. But if this is an anticlimactic conclusion to a day of football, nobody will have to wait long for more action. The point spreads for the next weekend's games have already been posted on the big board at the Stardust, Roxy Roxborough and his cronies are glued to their phones and computers and the money is starting to flow.

"Where the point spreads will be posted, rows of men sit like schoolboys waiting for the bell to ring."