please excuse the double posting, but i don’t want this to get lost in an edit:
i had a mistake in my calculation, now i presume the gradients to be around 4.5 % and 2.25 %.
here is how i came up with it:
we know that the contour lines are spaced at every 10 height metres. i then built a piece of track at the maximum inclination between two of these lines with the building cost set to 1 $ per metre. i of course bulldozed this piece of track and rebuild it in order to eradicate terraforming costs. it then costs 215-220 $ to build this piece of track, which should indicate its length. then i used pythagorean math to get the angle of the gradient, which turns out to be 2.6°. this then translates to around 4.5 % for the maximum gradient.
now i don’t know for sure whether there is still an error in there somewhere, but it seems that the trains definately can pull more (at a speed of well over 3 kph) weight than the pulling force figures suggest. it appears that the supplied pulling force might be the force that a locomotive can pull at around half its top speed. at least that about what they did when i tested it with the borsig,g3, ae 3/5 and the nohab.