It had been a while since I took Amtrak. My US Amtrak journeys include an over-night long distance trips (Atlanta – Baltimore and back) trip while I was in college (really uncomfortable, about an hour late, and not repeated) in the 1980s, and a few trips on the Metro-liner in the Northeast Corridor (which were fine, though slow) in the 1990s. I also have experience taking my now wife to Amtrak, for trips from Berkeley to Fresno, one of which was delayed 13 hours because of mechanical problems several states away. I have of course taken trains in other countries, which are generally decent rides, some better than others, some hybrids between transit and inter-city, some cleaner, and so on. Most have been highly punctual though, aside from various person-under-train events, which is either very common, or for which I am a magnet.

So after checking it out with two people before I bought tickets, just to make sure, I was pleasantly surprised at the Capitol Corridor, operated by Amtrak between the Bay Area and Sacramento. I was going to UC Davis to give a talk. The train was fairly modern, on-time, and clean. The user experience is pretty good. The ticket itself was “unreserved” so while I didn’t get an assigned seat, I could use it for other train trips, and was not bound to the scheduled ride. The fare was $23 each way ($46 RT).

Now Amtrak is still not a completely modernized system, though far better than it had been on my last Amtrak experience (and of course this is Amtrak California, which is not exactly the same thing as Amtrak elsewhere). They still use paper tickets, and don’t seem to have any way to integrate this line with the standard Bay Area payment systems (Clipper Card). (London is integrating its commuter trains with the Oyster Card, so this is not an impossible task, and seems the likely direction payment technologies should be taking).

BART and Amtrak interface at Richmond station. I was closer to the North Berkeley BART than the Berkeley Amtrak station, so I took BART to Amtrak. Doing it over, I might have just walked the extra distance, since there are no guaranteed transfers, and missing the BART train on the return cost me at least another 10 minutes wait, in addition to BART being slower than Amtrak due to the additional stops.

Going to Davis, I was early at the Richmond Amtrak station (but not so early that I could catch the previous train), a train stopped (not mine) with no information on the Variable Message Sign (VMT), this leading me to wonder whether it was in fact mine (which was listed on the VMS as the next train). The conductor informed me it was not. So human information systems still work even if the electronic systems are incomplete. Other people asked me about the train, so I guess asking is still standard operating procedure at train stations, and I am just anti-social (or new-fangled) for wanting my information machine-intermediated.

On-board, conductors scan your paper ticket and then still print out paper, using their mobile scanning devices to know who is going where. This paper slip is affixed above your seat, so they don’t have to bother asking for proof of payment again. You have to imagine the future is a small electronic indicator above the seats with that information that uses no paper, assuming they are going to continue to rely on proof of payment on conductors checking almost everyone’s ticket rather than gates and tap-in, tap-out. I do acknowledge the simplicity of paper, so this might in fact be more efficient even though it is less advanced. Similarly, proof of checking payment on-board reduces total travel time by minimizing queues at the entrance/exit gates. Still machinery is cheaper than labor, no?

The return trip was equally efficient. The load factor (off-peak time in the off-peak direction) in the morning was about 25%. In the afternoon (peak time in the off-peak direction) it was about 70%.