(The first part discussed, in terms of a dice game, the initiation of expense budgets of department heads.)

A sales manager gets to play the budget game with a sales projection, too, and a second roll of the dice. This phase is more exciting for the sales manager, because it relates to his or her primary responsibility, and may even be tied directly to compensation.

At first glance there's an excruciating dilemma. An ambitious projection would seem to help create an aggressive, company-oriented image. But the final sales budget will become the principal yardstick for measuring the manager's performance, and it's easier to stand tall next to a short yardstick.

The key to playing is for sales managers to know how they stand with their bosses.

Situation One: The High-Rated Sales Manager.

The best bet is to estimate just a modest increase in sales, or even a slight decrease if economic conditions are dire. The sales manager will only lose points with a dice roll of two or three, because his or her immediate boss usually understands this part of the game.

Of course, the sales goal will be increased if the dice add up to anything from two to 10, but starting with a low projection does help hold down the final budget.

Situation Two: The Low-Rated Sales Manager.

Different rules apply. This player must try to offset the points that he or she lost before the game ever started, so a more ambitious sales projection is mandatory. But it won't be revised upward much if the dice roll is higher than seven. A boss with little confidence in the sales manager doesn't want the target to be too difficult; they'll both be blamed if it isn't met.

So much for the easy part of the game. The stakes get higher as the various budgets wend their way upward. And the rules get too complex to describe here--partly because individual personalities become even more of a factor. But the essence of the game remains the same: a mix of budget needs, wants and guesses . . . modified by role playing . . . and subject to luck.

I have barely touched on the amount of estimating/guesstimating/ forecasting/fortune-telling/fiction (pick a word or supply your own) that is inherent in the budgeting process. But every company's composite budget is unavoidably riddled with assumptions--on what will happen to the cost of materials, services, money and more, and on its sales climate as well. Items are calculated to the dollar that can change by thousands if one of the major assumptions is off by a percentage point or less.

Even more ironic is the fact that companies that try to do the best possible financial planning generate some of the flakiest figures, because they have a penchant for long-range budgets.

Business forecasting, alas, is like weather forecasting: reliability goes down geometrically as the time span increases.

All in all, the amazing part of the budget game is not that revisions are inevitable. It's that good budgeting frequently happens. (It occurs whenever the chief executive officer rolls a nine, 10, 11 or 12.)