If you build a rocket with 1900kg of mass,with a Terrier engine,the TWR at Kerbin surface is the one I calculate in the picture below,and its the same one that the game gives you (bottom right of the picture).
But,if you calculate the same in vacum,my TWR and the game are the same if g is also 9,81,but in vacum it should be less.
So if you calculate the same in an orbit of 90km,the game gives you a TWR of 3'21,and it should be more,4'26 as I demonstrate.
So yes,i'm sure that KSP1 physics use g as a constant of 9,81.I imagine that for diferent bodies,its diferent,but it doesnt change with the distance between body and rocket.And thats my suggestion for KSP2,that g changes like in real life with the distance.