Somehow I missed this comment from Shaun Moss back in July. In this post I will discuss hardware, costs and risk mitigation. My approach is modular and can be scaled up or down. It's not enough to get to mars, we need to give our colonists a chance to survive in style which Jerry Pournelle correctly pointed out in the 70s requires more than survival rates of power and energy. Sending less than a dozen colonists puts them at too much risk (IMHO) of not having essential skills should some die. This also reduces the per colonist cost.
As much as possible we should include existing hardware or modifications of existing hardware. Vaperware on your critical path is not a good thing. Unfortunately, modifications of existing hardware may be more like new hardware rather than just a tweak here and there.
Shaun has a variation of NASA's reference mission. Mine is one way. An ERV is a bonus luxury item that may be included but is not essential. Mars colonization deserves the commitment.
SpaceX has two landers in the works; Dragon 2.x and MCT. MCT could potentially bring costs way down (sending more at a time reduces costs per) but it's too early to plan on using it.
So, working backward, we send enough Dragon v2.x to mars orbit for the crew arriving in orbit to make it to the martian surface. These landers would go on a minimum cost trajectory containing the personal property of two colonists that will ride them to the surface (each gets an allotment of half the lander's 2,500 kg total capacity.) So if a dozen crew go to mars orbit, six landers should be waiting. But we didn't go back far enough, before that happens we send supplies directly to the martian surface by the most economical means. We have existing hardware that could do that but more economical hardware is available once the Dragon v2.x is available (we have to test land them anyway before doing it with humans, this is two birds with one stone.) Each lander costs $150m (to NASA at a bargain basement of a billion each, is that too mean?) to mars (orbit or surface.) Six landers waiting in orbit and another six on the surface totals $1.8 billion. Just one lander on the surface will have enough food for 300 days for the entire crew. One lander will include two tractors (one day assembly for either) for digging trenches to create very large personal mansion and common mall habitats connected by underground shirtsleeve passageways (another reason why less than a dozen crew makes no sense.) The motivation is they own it so why not? Mine is not a marxist world view. The other four supply landers are just because we should. The six landers are positioned to surround the landing ellipses (but if Elon is right about pinpoint helicopter style landing we're just over compensating.)
That leaves getting to mars orbit. We either launch directly or refuel at LEO. Launching directly would be like the MCT option, not here yet. So we launch for mars from LEO but not in battleship galactica.
The ship I propose is launched on the existing Falcon 9, massing less than 13 tons and having a habitat volume of about 180 cubic meters (think Bigelow's Sundancer.) It actually replaces the F9 upper stage. It is a continuously reusable ship that once launched never lands. It does not require more than a little development because it is essential just an F9 upper stage and Sundancer permanently mated (plus fuel bladders for extra delta V and radiation shielding.) No on orbit assembly required (although the bladders may be carried internally to be deployed externally once in orbit.) We configure it with life support for twelve. This provides 15 cubic meters of private (and shared) space for each crew member. It should cost less than $200m dry to orbit. We need one for the first mission and more for subsequent missions (unless MCT does become available in which case we sell them to the used, in space already and ready to be refueled, ship market.) This brings our cost to $2b.
This ship is actually a profit center. Before sending it to mars it will require shakedown cruises around the moon. We rent LEO space to tourists and researchers while making certain it will keep our colonists alive for 500 days. 75% of life support is air and water. 6 kg x 250 days x 12 crew is 18,000 kg. Food (hydrated and freeze dried in palatable proportions) would be 2 kg x 500 days (during free return option) x 12 crew is 12,000 kg. Crew and equipment would be another 3,600 kg. So dry departure mass would be 13k ship + 30k life support + 3.6k crew equal to 46.6k kg. Water for a free return comes from recycling.
To get fuel, crew and supplies to orbit would require about 6 Falcon Heavy launches so total and final cost of our mission is $2.6 billion to get 12 colonists to mars. If our ship is a profit center earning $200m per month (10 or so customers at any given time) that's just over a year to pay for the entire mission. Tell me again... why exactly are we waiting?
Update: $20m profit per customer is too optimistic. $5m profit w/$20m cost is much more reasonable so 4.5 years to pay for the mission (unless we put more ships in orbit. They're cheap.) We don't send you to Russia for six months training either. It's a half hour training film instead. Professional crew will keep you out of trouble on the tour. Over time, the Falcon Heavy could lower the cost per passenger from $20m to $3m meaning a price reduction from $25m to $8m. As costs go down more will take advantage of the opportunity to experience space.
Update 2: One FH should provide fuel for an around the moon tour so... $25m for the basic one month LEO stay. $5m for each additional month. $10m more for the Apollo 8 experience. Visits to the BA330 zero g racquetball court, no extra charge.
Last thought? Another risk mitigation would be to take a four passenger lander with them in case they can't rendezvous with two of the six landers waiting in orbit. They would supply it with the free return supplies.