Should it come to war, U.S. and allied forces will almost certainly win on the battlefield. Iraq stands alone; we do not -- allied forces come from 27 countries. While Iraq suffers from economic sanctions (embargo, boycott, freeze), we can draw on all the world's resources. Iraqi troops suffer low morale, have recently ended eight traumatic years of war with Iran; allied forces are ready to go. Iraqi arms are good, but our military technology is better in every category. Rarely have the military odds been so lopsided in advance.

Just as important and much more problematic, however, is what happens when the shooting stops. Winning the peace requires a set of clear political goals. With dispassion, we must decide before war begins the ideal outcome for the United States. This exercise can lead to surprising, even unpalatable conclusions.

Because of the allies' enormous advantages in power over Iraq, the U.S. government has a wide range of options before it. Moving from the least to the most ambitious, it can aim to get Iraqi troops out of Kuwait, to destroy Iraqi nonconventional weapons, eliminate the Iraqi military machine, bring down Saddam Hussein, establish democracy in Iraq or divide the country among its neighbors.

Which is optimal? Assuming that the U.S. goal is to foster a stable, defensible and nonbelligerent Iraq, two imperatives immediately rule out most of these options. First, Saddam must not keep his chemical, biological and nuclear weapons programs. The allies must do more than eject Iraqi troops from Kuwait and return to the status quo ante; they must apply force in Iraq too.

Second, U.S. forces must not occupy Iraq. Tempting as it is to blot out Saddam and his foul regime, doing so would create chaos. Because the Baathists have liquidated alternative leaderships (with the single exception of the Kurds), ousting Saddam probably implies an occupation of Iraq by American and allied forces -- and this spells near-certain disaster. Much as they goaded the Lebanese against U.S. Marines in Lebanon in 1983-84, the Syrian and Iranian governments would urge Iraqis to acts of terror. American troops would again find themselves victimized by suicide attackers, car bombers, snipers and a range of other unpleasantries. Our Arab allies would desert us, too.

The proper course of action, then, is to do more than take back Kuwait but less than occupy Iraq. Within those parameters, here are some specific steps for American planners to consider:

Resolve to use ground troops only in Kuwait and southern Iraq; use only air power in the rest of Iraq.

Affirm the territorial integrity of Iraq within its present borders and impress this commitment on would-be aggressors (Syria, Turkey, Iran).

Announce that the U.S. government does not intend to bring down Saddam Hussein. However distasteful this decision, it is necessary because it gives Saddam more incentive to end the fighting, and it gives Washington the flexibility to deal with Iraqi regime.

Destroy Iraq's nonconventional facilities from the air. As there are only some two dozen sites, this should not take long nor be costly in American lives.

Leave as much of the country's basic infrastructure standing as possible. Excessive destruction would create a power vacuum that either we or a neighbor would be obliged to fill.

Insist that the Iraqi government reduce its armed forces to about 200,000 soldiers, enough to defend the country from its neighbors but not more.

To ensure this diminution of Iraqi power, reach an agreement in advance with allies about maintaining a military embargo after the war. Because the allies might default on their promises, the U.S. government should keep the option open of further military action.

Some of these steps are counter-intuitive, others are painful or difficult to achieve. But all are necessary if Americans are to emerge from this crisis with gains from the sacrifices they are likely to make. The writer is director of the Foreign Policy Research Institute in Philadelphia.