I tend to agree with all of that part. (Amtrak did try running them on alternate days in 1978 and 1979. It would be interesting to see how much the Builder's ridership was hurt by that, vs. the daily Builder and triweekly NCH before that.) A 4x/wk California Zephyr and 3x/wk Pioneer was certainly ugly.
In the 1995-1997 period when the Empire Builder was reduced from daily to four times per week (opposite the tri-weekly Pioneer), ridership did indeed fall by about 40%, the amount of the frequency reduction. The theory was that people would simply change their travel plans to coincide when the train operated. That didn't happen, and part of the reason was that in 1994, the Empire Builder was the single most-ridden long distance train and was sold out a lot during hte summer, so the less-than-daily operation simply turned people away.
This is only a problem if you hold the Chicago departure time fixed. Note that during much of the time that both trains ran, the NCH left Chicago 3 or 4 hours before the Builder and got to Spokane 1 or 2 hours before. I would hope for a new service that either did the same thing, or ran one train overnight from Chicago to Minneapolis.
Well, that's the point. Something's got to give with longer running time. Either it takes extra equipment because the consist can't turn on the West Coast the same day (and likely no Coast Starlight connection) or it leaves Chicago much earlier and severs connections, which naturally reduces ridership. The salient point is that the current (when daily) Empire Builder schedule maximizes both connections on each end and equipment utilization. Anything longer sacrifices one or more of these attributes.
There is much fuss made of the NCH route being slower -- but Fargo to Spokane was only between 1 and 1½ hours slower, both pre-Amtrak and in 1979, than the Builder. A lot depends on the quality of track in North Dakota. MRL has maintained the ex-NP in Montana to a very high standard and runs freight at 60 in many places - can probably run passenger at 79 without doing much besides putting up the speed limit signs.
Well, the pre-Amtrak and early Amtrak schedules really don't mean a lot. Pre-Amtrak, the Empire Builder always had a lot of fat in the schedule, and prior to 1966 or so, the Empire Builder was about 2 hours faster than the North Coast Limited Fargo to Spokane and 3 hours Fargo to Seattle. Both ways, but especially eastbound, the Empire Builder was scheduled for optimum arrival times at Spokane, Whitefish, and Minneapolis rather than operating on the fastest schedule. In the later 1960s up until Amtrak, the Empire Builder was bogged down by its "slowest common denominator" which was the North Coast Limited, as the trains were consolidated between Chicago and St. Paul and Pasco and Portland. GN added time to the Empire Builder at intermediate stations as to not cause long station well at key stations a pre-700 AM arrival in Seattle. The last BN Empire Builder departed Seattle 1 hour 15 minutes after the North Coast Limited, but left Spokane 4 minutes ahead of it and arrived in Fargo 2 hours, 18 minutes before, then set out across Minnesota on a schedule with a 39 MPH average running time because it had to wait for the North Coast Limited in St. Paul. But even this schedule (west of Fargo) had fat in it: 5 hours 45 minutes Whitefish to Havre on a run that could be done in under 5 hours (and was reflected as such in Amtrak schedules). As an Amtrak train, the Empire Builder operated via Grand Forks on a route with less than stellar track and included a backup move out of the station. Today, with no backup at Grand Forks and an upgraded railroad courtesy of the Bakken Oil Boom, the eastbound Empire Builder is about 50 minutes faster from Minot to Fargo as the first Amtrak Empire Builder in 1971.
The "quality of track" is not an issue, but it is illogical to presume that a route without passenger trains for 41 years could operate on a schedule when passenger trains did run. It all depends on money: If you have enough of it to invest in track, signaling, signage, and recalibrating grade crossing warning devices, then the schedule can be as fast as the money dictates. If your budget has limitations (very likely), the schedule will reflect. Primary case in point: Missoula to Paradise. Passenger trains operated over Evaro Hill rather than along the river via St. Regis. Evaro is faster, but is now dark territory. St. Regis has block signal protection but takes longer. Either way, the train would take longer than in Amtrak days, but the choice would have to be between the even longer route via St. Regis or adding signals (a significant expense) on the route via Evaro. West of Spokane, there would likely be a local desire to tie in this service to the existing push for east-west service via Yakima (and BNSF would likely not support a second passenger train on the Cascade Tunnel route), which would add additional time (and on a route without block signals) and expense. And yes, the Southern Montana train could be consolidated with the Empire Builder at Spokane, but again that would be time-consuming and potentially risky (if one train or the other was late) and likely could put Empire Builder connections in Chicago and Portland and West Coast turnaround time in jeopardy.
Given the challenges on MRL's three steep grades (Bozeman Pass, Winston Hill, and Mullan Pass) and the what would be necessary to sufficiently increase track speed on curves, my guess would be a St. Paul-Seattle running time of 4 to 6 hours longer than the Empire Builder unless a buttload of money was available.
