After posting that the best way to reduce Fort Collins' car dependency was to introduce a rail service to Denver, I decided that I should actually work out its practicality. I measured the interesting stops between Fort Collins and Denver along the existing BNSF lines: Fort Collins (0), Loveland (14), Longmont (30), Boulder (43), Broomfield(54), Denver(66).
Let's aim for exactly 1 hour to Denver. Does 1 minute stop at each station sound reasonable? If so, and assuming a 0.1G acceleration leaving and entering a station, a running speed of 80mph would meet this target nicely. We would double track the line and divert existing freight traffic to off peak times and to the parallel Union Pacific line. The high density of track and reserves in the corridor means that there is plenty of opportunity for branch lines for future capacity. A link to Greeley is the most obvious extension, along with an alternate route through Eastlake and Commerce City.
It would run every say 10 minutes, and it turns out that 10 min @ 80mph is 13 miles - approximately the inter-station distance.
So there would basically be a train at every station in both directions. That means that with a train idling/refueling for 10 mins at each end you'd need 14 trains. Each double decker DMU train can hold 218 passengers which means a capacity in a single direction of 218 * 6 = 1308 passengers an hour, about a quarter the maximum capacity of the I25 (The main competitor). Of course the nice thing is that you can simply add more coaches as demand increases. To match the peak capacity of the I25's 1800 cars / hour lane = 5400 cars / hour (i.e about 5400 people), which must be reached most week days, requires just 24 coaches an hour. Grouping in 2s means 5 mins between trains.
I believe that it is better to increase frequency than train length. Above a certain frequency you don't need a timetable, you simply guarantee a train within say 10 mins, and as demand increases you just fit more in the gaps and add coaches.
The service could easily be extended to Cheyenne. Skipping Boulder and traveling directly from Longmont to Broomfield shaves off 10 miles and about 10 minutes from the trip time. There is little time benefit from skipping any other station.
One might be tempted to add more stops in to pick up more passengers, I think this would be unwise, as more stops eats into the 80mph operating speed quite quickly. A better option is to provide a feeder service using street cars or buses. Fort Collins would do well to extend its street car through town along the old route (the rails are still mostly there) past the library. Good bike facilities (perhaps including a bike car) are a must.
Similarly, it is common for such services to try and start out as a commuter express - only traveling to the city till 9, and only traveling back at 5. This is a mistake. For example, the CalTrain runs infrequently off peak, and despite the far greater market, carries less passenger miles than the VTA light rail, which runs at least every half hour, and up to every 7.5 minutes, to places that nobody wants to go to (compare San Francisco to 'Great America', which one have you heard of :)
Instead, run the train from say 6am to midnight, every day. People don't like uncertainty, and they don't like to be stranded.
Update: Turns out that RTD is already part of the way there: FasTracks
Interestingly enough, they are implementing BRT, LRT, diesel and electric commuter rail. The electric commuter rail will cost about $16M/mile, compared to the bargain BRT at $13M/mile.