*sighs*
Ok, I'm putting my thinking cap on to sort through a service option (let's face it, I need the distraction right now) using nothing over a 125 MPH non-electric (i.e. diesel) locomotive. A dual-mode would be acceptable as well if one were available, but the added loco cost would probably be, well, loco.
As is often the case, there are two options for such a service. One option is west over either SCRRA Cajon Pass (where I-15 and BNSF go through the mountains). The other is via the Antelope Valley line, heading north from Los Angeles to Palmdale. From Palmdale (or another location along the route) you would proceed east to Victorville or thereabouts, at which point the two routes would rejoin one another.
So, at first let's consider the costs:
-The Cajon Pass route would require significant work within the pass (which is already overloaded with freight traffic), and for a moderate number of daily trains (say, 5-8x daily round-trips/10-16 daily movements) you'd almost need an additional track there. West of there you would need some additional trackage (i.e. double/triple/quad tracking and/or passing sidings), though doing so would probably be less of an issue (you might even be able to "buy off" some Metrolink slots, depending on which routing you chose).
-The Antelope Valley route would require additional trackwork along the existing Metrolink route (double/triple tracking and/or sidings), though it stands to note that such work could, in the short term, also serve any through trains heading on to Bakersfield that might be using HSR tracks from there north (i.e. assuming non-electric ops along that route). You would then need to build about 90 miles of track westward to Victorville.
-From Victorville to Las Vegas you would be building new track, either near I-15 or near existing UP tracks. Again, not going with an HSR plan or a massive number of frequencies you'd probably need somewhere around 1/3 of the distance to be made up with double-track (either in the form of long passing sidings that would allow passing "at speed" or in the form of outright double-track).
Therefore, for the Antelope Valley line...starting with the $6.4bn from XpressWest's estimates, I figure you can knock about $1bn off the front end because you're not electrifying the line (IIRC electrification tends to run about $5m/mile). I'm not sure how much you'd knock off by "losing" some double-track (and it might be worthwhile to keep some of that; if nothing else, at night you could probably work some sort of deal to let UP dump some freight ops on this line in exchange for some cooperation on slots heading into LA). My best guess is that the cost/savings there is about $4-5m/mile (per a Cambridge Systematics report which pegged the cost of adding a mile of CTC-TCS track at $3.8-4.4m/mile).
Taking all of this into account, you would probably have a cost somewhere between $2.5-3.5bn on the Antelope Valley line. [1] Moving over to the Cajon Pass line...I'm guessing that the costs are about the same: You save 50 miles of track by going through the San Bernadino Valley, yes, but you also lose any savings because Cajon is likely more trouble to work with as it stands and you're going to be using a lot more dynamite on mountains to make that happen. [2]
[1] Backing this out, at $4.4m/mile the 185-mile Victorville-Las Vegas line, all double-track with no electrification, would probably run about $1.63bn. It might even come in somewhat lower because of the land situation, though engineering near Las Vegas would be an issue. The Palmdale extension would throw another $440m onto the project (50 miles at $8.8/mile double-track). Palmdale-LAX is anyone's guess, but I'm going to guess $500m in additional trackwork, giving you a total of $2,570m for trackwork, plus stations and equipment.
[2] Basically save $440m on the Palmdale line and probably $100-200m on trackwork since the existing ROW is probably more workable in Southern CA (there's simply more track to play with in Southern CA as well as more old ROW that can be re-tracked), but put that right back into Cajon Pass.
Speed-wise, Antelope Valley would probably offer about a 90 MPH average speed from Palmdale to Las Vegas (235 miles, which would be covered in about 2:40). I'm basing this generally on All Aboard Florida's expectations from West Palm to Orlando; 90 MPH should be achievable on a line that should have at most three intermediate stops (Victorville, Barstow, and suburban Las Vegas). From Palmdale it would almost definitely be faster than Metrolink; the Metrolink express train takes about 1:30-1:35 on this line to get to LAX, so total travel time should max out at 4:10 or so. Assuming some improvements on the Antelope Valley line (curve straightening and the like) you could probably jam that down to something around 3:45 or so. Compared to driving, this would probably be a winner from just about anywhere in the western LA area (on the eastern end of the San Bernadino Valley the proposition is more questionable since a lot of the bad traffic is in the valley...though I-15 has much to be spoken ill of on this front as well).
The Cajon Pass line is probably more problematic: The SW Chief takes 2:55 Los Angeles-Victorville (averaging a whopping 40.8 MPH). Even assuming 90 MPH Victorville-Las Vegas (approx. 2:00) this gets you to about 5:00 in travel time Los Angeles-Las Vegas: This is drive-time competitive in bad traffic (though San Bernadino/Victorville-Las Vegas would be more competitive). However, going for a San Bernadino Line timing (and going with that line's express service times as a guide), you'd clock 1:05 Los Angeles-San Bernadino. Add another 1:10 San Bernadino-Victorville (for 2:15 total) and the grand total would probably be 4:15-4:20. This is on par with the Antelope Valley option.
One other aspect matters on this front: Ridership. Between the two options, the Cajon Pass line would be the obvious winner in terms of picking up passengers along the way (as it would likely operate along an existing commuter route in the LA Basin), though premium fares could be used to keep the line from being swamped by local traffic (with lower fares used to permit it to offset traffic boarding later on, allowing the seat to be "sold twice").
I guess the question at the end of all of this is whether a potentially $3-3.5bn conventional rail project (maybe more, maybe less) would be worth the effort? You could vary frequencies from my initially proposed 5-8x daily round trips (assuming capacity in the passes you can easily bump that in the 10-16x daily range, if not higher, if demand allows...and doing so would probably be of substantial benefit to Metrolink if you allow "local" space to be sold, subject to availability controls, within the LA area: 16x daily trains would double service along the Antelope Valley and nearly do so along the San Bernadino line), and doing so might not add too much to costs (the line from Palmdale or Victorville is largely a fixed item, for example). There's also a lot that could be saved if you were willing to drop to single-track (since each train would only need to handle a handful of meets along the run outside of Metrolink territory).
