This might be off-topic, but has anyone heard of the Aurora Mars mission from the CDF Study Report Human Missions To Mars?
The full report (400+ pages) can be found here:
http://emits.sso.esa.int/emits-doc/1-5200-RD20-HMM_Technical_Report_Final_Version.pdf
From a quick skim, one thing to note is they assume an Isp of 345s for the storable stages, stating "Note that actual storable technology for such thrust levels provide an Isp of 325 s. Here an optimistic approach has been taken."
They also assume use of Energia for most of the launches (21 out of a total of 25 over a minimum period of 4.6 years). This is a report written in 2004, whilst Energia last flew in 1988 after a grand total of 2 launches, but they make the assumption that "the effort to make it operational is smaller than the one to develop a launcher of such performances or even higher from scratch." If SLS is any guide, it is questionable how much saving re-use of existing designs gives you, and at least with SLS all the main shuttle components were still in production, which hasn't been the case for the Energia core for a couple of decades. The costs to re-start the production plant and recreate the launch infrastructure for Energia would be insanely high, and it's not a promising indication that the Russians can't even get their act together to finish Vostochny for Soyuz.
The study goes on to state that smaller launchers could be used, so to launch the necessary 1355 tonnes of hardware and propellant they would need about 68 Ariane V launches based on the 20t payload assumption in the document. That's assuming every launch is practically at capacity, which would not be the case, so probably 80 launches would be closer to reality, meaning a launch cost in the region of $15B (Wiki puts an Ariane V launch at $165–220M). Since its first launch in 1996, Ariane V has made
94 launches, with 4 failures. So this mission alone would need to almost double the number of launches. The maximum Ariane V launch rate has been 7 per year, so you're looking at over a decade to launch the mission to LEO, even assume Ariane V is used for nothing but Mars mission launches in that period, and that no replacement rocket for commercial spacecraft needs to use the Kourou launch facilities at the same time. Of course, you could just build an entire parallel launch infrastructure dedicated to this mission ("The first rule of government spending: why have one when you could have two at twice the price?"), but that would push an already unaffordable mission into the realms of total financial fantasy.
Of course things have moved on since 2004. Shift that 1355t of mass to LEO onto Falcon Heavy and you could do it in about 25-30 launches, coming in at under $3B launch costs and possibly still fit within the 4.6 year timeframe. Or, if you hate yourself, use 15 SLS Block 1B launches and pay 5x as much over a decade and a half. BFR/BFS of course would do even better (if it flies as promised), but given that's designed to do a full Mars mission anyway, using it to launch components of a less capable mission would be kind of pointless.
Still, I think the number of launches are the biggest problem for this mission, along with the complexity of integrating in LEO. These are problems going back to von Braun's very first Mars mission, and indeed this ESA study seems to take pretty much the same approach: Work out what you want to deliver back at Earth (the crew plus some samples in some kind of return capsule), work out all of the major manoeuvres needed to acomplish that, then add chemical rocket stages for each manoeuvre. The ESA study is even using pretty much the same propellants as von Braun proposed in the 1940s (except for the cryogenic departure stage). It produces a credible design using achievable technology, but it's not an elegant solution, and financially and programmatically has many issues.