I don't see what would be so difficult about going back to the old-fashioned way of testing.
1) Establish a dedicated beta-testing team by soliciting users who want to join just like they did with the Xbox version and previous versions of Trainz.
2) Have the users sign an NDA if necessary.
3) The users are given the task of testing and reporting. Have them use specific steps to test specific things, if necessary, as well as free test the new upcoming version.
4) After these beta-testers are satisfied with the quality of the new release, push the release on to public testing.
5) Public testing circulates around to allow various users with various computer systems to test the versions on until everything looks good.
6) Produce an RC version. Let users at it. If all is well...
7) Release the product.
What stops them? M O N E Y. Doing this the right way costs money.
Well, with this release, they have lost money due to having to redo and retest, come up with a hotfix or two, or three, and in the process, they have lost the goodwill of the community and added a nice brown smear on their faces from releasing the crap initially.
If I wasn't a longtime user of the product and purchased this version, I'd probably relegate it to the trash bin just like I've done with other programs in the past. Way too many actually to count.