Bug? distance_per_tile setting effect on vehicle speed

[megasycophant]

I've been trying to figure out how vehicle speed in km/hour translates to actual tiles per month (please give me a clue if you know) so I decided to set up some simple test cases to measure the effects of different game settings on # tiles traveled. I'm running Simutrans 102.3  Experimental 8.2 (built 2010-06-28) on 32 bit Linux, using pak British 128 Experimental v0.6.

Scenario: 1830 timeline, SDR "Locomotion" 0-4-0 engine (top speed 25 km/h) running out of the depot on a long straight, flat stretch of track, starting out of the depot on Jan 01 1830 0:00 (at game start), tiles traveled measured around Jan 15 1830 0:00 (14 days). Note that this does not account for time to come up to top speed, but the intent is just to gauge whether distance_per_tile setting has an effect.

First test case: default 4 tiles/km (distance_per_tile = 25)
Engine traveled 35 tiles (8.75 km).

Second test case: 1 tile/km (distance_per_tile = 100)
Engine traveled 72 tiles (72 km).

So it seems that vehicles' top speeds vary inversely (geometrically even?) with distance_per_tile? What gives?

[jamespetts]

As you know, there are three relevant factors in all of this: speed, distance and time. Speed must remain constant, as the speeds of vehicles in km/h are fixed so as to reflect their real-life speeds, and their speeds in game tiles per real-life second are fixed so as to be as user friendly as possible (and are difficult to change). So, if speed is to remain constant, when changing distance, time also has to change. The way in which time is measured therefore fluctuates depending on the distance per tile setting.

As should be apparent, there are two different measures of time in the game which, at any given distance_per_tile setting, have a fixed relationship to each other: (1) the historical time (days, months, years); and (2) the travelling time, which is based on the reported speed of the vehicles and the number of kilometres or parts thereof travelled per tile traversed. It is only the time of type (2) that changes when the distance_per_tile setting changes.

In any event, the point is, given a fixed speed, when the distance scale changes, the time scale must also change. I'm sure that Einstein would have approved.