Just reading some of the comments from the Constellation used-car sale pitch going on in Huntsville today. One of the topics discussed was how Ares-V enables manned missions to Mars. The Marshall guys put up a chart showing that depending on whether we go with NTRs or chemical propulsion, Ares-V could place the needed mass into orbit in only 7-12 launches.
To quote the former senior NASA official who was poo-pooing depots in the article I linked to earlier
Rocket malfunctions are not uncommon, and the more launches are needed for each moon mission, the more likely it is that something will go wrong, a former senior NASA official told New Scientist.
ESAS had a whole section slamming lunar architectures that used more than two launches as having too high of a probability of losing the mission. The ESAS study pointed out that the odds of not losing any of the launches is much lower than the odds of not losing any one given launch. The probability of successfully doing N launches without any failures is the probability of success for a single launch raised to the Nth power (ie 6 launches with a 98% reliable vehicle only has 88.5% chance of not losing one vehicle). They also went into rendezvous reliability, launcher availability, boiloff losses etc…
…but for some reason that doesn’t matter for Mars missions. Why are large numbers of launches, rendezvous, and long duration cryo storage considered perfectly acceptable for Mars, but completely unnacceptable for the Moon?