The point about factories and roads is interesting, but less significant (at least in Pak128.Britain-Ex) than you suggest because, in that pakset, most industries (including all early industries) are set to be built in towns where there is already a road network, and later industries that are built out of towns are mostly the heavy sort of industries to which one would normally build a rail rather than road connexion. This arrangement is more realistic, I think, as early industries were almost invariably built in and around settlements in order to have enough workers.
However, some consideration may need to be given to restricting out of town industries to being built next to an existing road to prevent the problems that you discuss. I should be interested in people's views generally on that subject.
As to the difficulty of your pakset, I see that you have put in place compensating mechanisms to prevent it from being too easy, so I was perhaps hasty when I suggested that it would be easy. I think that it is probably just a matter of having a different taste in what sort of challenge to have: the ideal with Experimental is to have as realistic a challenge as possible: the things that are interesting and challenging about real life transport planning should be, in so far as possible in this sort of game, the same things as are interesting and challenging about playing the game. The idea is that one should solve challenges in the game in the same sort of way as one would solve the equivalent challenges in reality, so that players can imagine that they are in charge of a real transport network and make the sorts of decisions in the game that they would make if they were in charge of a real transport company.
I fully accept that this sort of realism is not to everyone's taste (which is why we have the separate Experimental and Standard branches in the first place), but there are a goodly smattering of people, including me, who do like the realism, I think, so it is all worthwhile in the end. Your experiences in finding what aspects of Standard do not permit realistic balancing are most useful in understanding how to develop Experimental to allow a realistic balance, and I am most grateful to you for sharing them. If you have any more, I should be very interested to know about them.