LAS VEGAS -- When the Redskins play the Cowboys Sunday, most of America will be aware that Washington is a 14-point favorite. That number has been drummed into the public consciousness all week. Local sportscasters routinely cite it in their reports on the game. The number has appeared daily in the listing of point spreads run in this newspaper. Anyone who bets with a bookmaker, plays a parlay card or takes part in an office pool knows he or she must decide whether the home team can win by more than two touchdowns.

But why is the crucial number 14? And where does it come from?

The creation of the point spread on the Redskins game -- as on all other football games -- is the product of a process that is partly science, partly art, partly capitalism at work. It involves many people, both in this gambling mecca and around the country, but the process started last Sunday in an office just off the Las Vegas strip.

On that afternoon, Michael J. "Roxy" Roxborough was, in fact, preoccupied by the Redskins. Roxborough's company, Las Vegas Sports Consultants, provides point spreads and odds to most of the important bookmaking establishments in this state. The object of a point spread is to attract roughly equal amounts of money on both teams.

But for the Redskins-49ers game last Sunday, Roxborough conceded: "We blew it." The Super Bowl champions had opened as a 6 1/2-point favorite, and money had poured in on the Redskins, forcing bookmakers to lower the spread to 6, to 5 1/2 and finally to 5 in order to attract any significant 49ers money. If Washington covered the spread, Roxborough's clients would take a beating on the game and so the oddsmaker was unabashedly rooting for San Francisco.

He got his wish, of course. San Francisco not only won, but covered the spread, making it the bookies' most profitable game of the week. Even so, neither oddsmakers nor bookmakers want games like that. They want the action on the two teams neatly split. Since bettors must lay 11-to-10 odds, a game that attracts evenly balanced action means a guaranteed profit.

When Roxborough went into business in the early 1980s, this city's preeminent linemaker, Bob Martin, had retired, and various sports book managers were making point spreads as much on the basis of instinct as on solid information. A famous team of gamblers known as the Computer Group was winning millions by employing computer analysis to find weaknesses in the point spreads, mostly on college games. Roxborough brought computers into the fray on the bookmakers' side.

He and a statistics professor, Mike Orkin, devised a simple but effective computerized system that created so-called power ratings for all teams, one that gave a numerical assessment of each team's ability as well as the strength of its home field advantage. It was recalculated after each week's games.

The Redskins, for example, have a current rating of 103.8, Dallas 94.2 -- a difference of 9.6 points. The Redskins' home-field advantage is worth 2.5. The computer says Washington should be a 12.1-point favorite. That is Roxborough's starting point, but only a starting point. "Roxy's strength is that he not only has a good feel for the strengths and weaknesses of teams, but for how the public bets," Orkin said.

When Sunday's games were finished, Roxborough had about an hour to finalize his point spreads on the upcoming week's games. He had made his tentative lines, and now he placed phone calls to two other top linemakers to compare notes. One was to Chris Andrews, director of the Club Cal-Neva in Reno, and they quickly went through the schedule, game by game. They came to the Redskins-Cowboys game and Roxborough said: "I thought 12 1/2 or 13."

Andrews said: "I made it 14. The Cowboys are pretty bad. They should have been 9 points today." (The Cowboys had only been a 7-point underdog to the Giants and had been swamped; the betting line somewhat overrated them.)

"I couldn't make up my mind," Roxborough said, "but the Redskins do have a good chance of annihilating them. It'll be 14."

The number was set, and within an hour it was going to be tested. At 6 p.m. local time, the Stardust Sports Book would post the first point spreads in Las Vegas. The manager of a rival establishment said: "The Stardust does everybody a favor -- they take the initial hits and they iron out the line."

Because of this vulnerable position, the Stardust's manager, Scott Schettler, also tries to create a point spread by consensus. "One guy can't make a line," he said. He makes his own line. He has two employees on his payroll who make lines. He buys Roxborough's line.

A crowd had gathered by the betting windows to take part in the Stardust's "lottery," the establishment's means of imposing some order on the initial wagering. Would-be participants wrote down their names, which were drawn to determine the betting order.

If a customer is in the lottery, he has to bet a minimum of $300 on at least one game. The maximum is $10,000 on a pro game. The bettors were directed to their positions: "Jack F., Window 1, No. 1. Eddie Z., Window 1, No. 2." . . . The betting started, and changes in the lines were announced in rapid-fire succession: "Fresno State 32. . . . Texas Tech 9. . . . Bears 2. . . . "

When it ended 15 minutes later, not one bet was written on either the Cowboys or the Redskins. Nobody saw an edge to one side or the other. It was a solid number.

Of course, this was only the beginning of a weeklong process. As Sunday's games drew nearer, there were moves in the lines, rising and ebbing tides of sentiment for various teams, but the Redskins-Cowboys encounter was not attracting a whole lot of attention.

A couple of bookmakers, with a small surplus of money on the Cowboys, has lowered the number to 13 1/2 to generate action on the Redskins, but for the most part the game inspired indifference. On Wednesday, the manager of one sizable sports book checked his computer and found a measly $1,000 had been bet on what was once one of pro football's hottest rivalries.

The point spread is almost too good, for bettors do not see a real edge on either side of it and they are concentrating their action on other games. But that 14-point spread -- which is neither as subjective nor as mysterious as many people might think -- is the right number, the true number, at least in Las Vegas's view of reality. And even if the Cowboys win in a romp, it was the right number because it split America's wagering opinion on the game.