The U.S. Department of Defense regularly runs war simulations on all sorts of interesting scenarios, which are unfortunately (for us) classified. The Brookings Institution, however, which is chock-full of former U.S. officials, did run a recent simulation. Check out the indispensable David Sanger's report here. Below is a summary of his observations:

1. By attacking without Washington's advance knowledge, Israel had the benefits of surprise and momentum - not only over the Iranians, but over its American allies - and for the first day or two, ran circles around White House crisis managers.

2. The battle quickly sucked in the whole region - and Washington. Arab leaders who might have quietly applauded an attack against Iran had to worry about the reaction in their streets. The war shifted to defending Saudi oil facilities, and Iran's use of proxies meant that other regional players quickly became involved.’

3. You can bomb facilities, but you can't bomb knowledge. Iran had not only scattered its facilities, but had also scattered its scientific and engineering leadership, in hopes of rebuilding after an attack.

4. No one won, and the United States and Israel measured success differently. In Washington, officials believed setting the Iranian program back only a few years was not worth the huge cost. In Israel, even a few years delay seemed worth the cost, and the Israelis argued that it could further undercut a fragile regime and perhaps speed its demise. Most of the Americans thought that was a pipe dream.