ORIGINAL: Don Bowen
First, let me say that the following is all that I feel able to give by way of explanation on the methods of TF routing. The internal workings are, of course, proprietary anyway. So please do not bombard the forum with requests for additional information or specific "what ifs". This is it.
Second, I am away from my home computer tonight and have only my failing memory to rely on. And, as they say, memory is the second thing to go. But it gives me more time that I normally have, so:
Calculation of TF routing path involves examination of possible routes between source and destination, and assignment of a movement cost for each hex in the path and, in summary, total path. The lowest cost path that successfully routes is selected. If no acceptable path is available, the dreaded 9999 range is returned. The "cost" of each hex is determined by topography and enemy threat. Each are separate items, and are additive.
Some hexes are blocked - land, reef, etc. Others are just undesirable - shallow, coastal, bad spearfishing.
Enemy threat decreases in effect as range from the point of threat increases.
To answer the specific question, topographical considerations are still in force if no enemy threat is present. However, with no enemy threat increasing the cost, a TF may be better able to tolerate highest topographical costs.
Why then if I give the same TF the same destination multiple times, it picks different routes?
When I see a strange route, sometimes I cancel the TF's destination, and give it the same destination again and it will sometimes pick a different route!