In Best very early money, and implications, I showed that early trains do not benefit from going farther - in fact, it's a detriment. At that point of my test game, I hadn't gotten to the later, faster engines. Now I have...
* A train begins to have a speed/distance bonus when it moves at 60 mph or better. This is for the whole train (i.e., with its full load), not the "max mph" seen in the Roundhouse.
* 50 mph is a wash. Sometimes it will get a little more going a far distance than a short distance; other times it won't.
* In the best case - GP-Series over a long distance - A train full of passengers and mail (P&M, picking everything up at both ends) made:
NYC <-> WDC, 50.5 game miles: $3,283,500
NYC <-> Aberdeen, 34.3 game miles: $2,536,500
NYC <-> Trenton, 9.9 game miles: $1,540,500
* These were the values for a fixed period of time (Jan 1952 to Aug 1958), so they could easily be turned into a "rate" for the train (dollars per year, or whatever). It was a Named train, so your dollar figures will be less, if not Named. All three cities were long-established Metropolises. It was a Northeast scenario (Financier diff.), but I'd bet all scenarios use the same algorithm.
* Be careful of looking at numbers for one delivery; you have to take into consideration the time it takes to move to the various places. A.k.a., it doesn't matter if one load made twice as much money, if it took three times as long to get there... it's actually making less money over time.
* As you can see from the NYC to WDC numbers, over twice as much can be made for going a long distance rather than a short one. But this was for trains on dedicated tracks, without any interruptions... if there are any waits, the advantage may be greatly reduced, or lost entirely.
* For comparison, here's a Mikado doing the same route. It's a strong engine that can pull all available P&M, and still make max 60 mph:
NYC <-> WDC, 50.5 game miles: $1,692,000
NYC <-> Aberdeen, 34.3 game miles: $1,405,500
NYC <-> Trenton, 9.9 game miles: $1,108,500
This may look like a big difference, but remember that that's the total income for a 6.6-year period (Jan 1952 to Aug 1958). The WDC run is actually only making $88k more per year than the Trenton route. Still, a 50% increase in Passenger and Mail - the top two revenue earners in my Northeast game - is nothing to sneeze at.
* Be careful when using the MPH figures shown in the Roundhouse. Ideally, you want your train to be able to carry all the P&M that has accumulated in the two cities on your route. For example, the Pacific engine looks tempting, with its max mph of 80... but the graph shows it starts slowing considerably at around half full of P&M.
* You might think to e.g. run the Pacific with only enough cars to keep top speed. But unless you then do something with any P&M not picked up, you may ultimately be losing money. Adding extra short-run trains to pick up the excess P&M may make your long-run train Wait and/or otherwise cause congestion, and ultimately lose money overall, too. In an actual test of the Pacific, it failed at long runs. It gets slowed too much if trying to pick everything up.
* You might also consider only hauling Passengers (instead of both P&M)... they pay considerably more than Mail. But even then, the Pacific did significantly better on the NYC-Trenton route, vs. Aberdeen or WDC. It just doesn't have the pulling power to pick up everybody that accumulates over long runs (and leaving some passengers behind makes other complications).
* One way to estimate top speed is to choose an engine and add cars to a route, and watch the speed read-out on the Route planner. In my hands, though, it usually under-estimated the speed for fast trains.
* Some interesting complications can arise. Example: The F-Series could reach mid-70s (max rated mph = 80) bouncing from NYC to Aberdeen when picking up all available P&M. If it did the same for NYC to WDC, however, it picked up so much P&M that it slowed - and the NYC to WDC route made less money than NYC to Aberdeen, over the time period.
* If faced with limitations such as the above, then, an alternative is to just haul Passengers, instead of both P&M. And try real hard to use a Named engine!