I don't want to turn this thread into a programming debate, which is one of the reasons that I rarely post code
I hope that none of the other readers of this thread have been put out by our monopolization of the topic of discussion
Are you suggesting that instead of a new list, I use the LINQ without assigning to a new list as the target in the foreach loop?
If so, I prefer to avoid that. It's harder to read and harder to debug and those are more important to me than straight performance.
Yes, both the CurrentSystems and PossibleSystems variables would be IEnumerable<KnownSystem> instead of List<KnownSystem>. I can’t really see how the code would be more difficult to read, since everything else would be identical except that you would just not call ToList().
However, I suppose I know what you mean about debugging. You want to stop at a breakpoint and be able to view the lists in the debugger. I tend to forget that not everybody can use omniscient debugging yet (I use a debugger called Pernosco a lot, but it is for Linux only.) If I were using a regular debugger and I wanted to verify which systems had been in the final enumeration, I would add a conditional breakpoint inside the loop to print out some information about ks (the system name and number, for instance) and then continue execution. (In Pernosco you can do the same thing but you get all the information that would have been printed without having to actually run the program again.)
As for performance, if this code is just called once any time a fleet transits an unexplored jump point then it is probably irrelevant. Even though you make two whole copies of the list of systems (and you did just make that list significantly larger), that probably takes so little time that nobody will notice. But you might take a look to see if you have made this mistake in more important parts of the code. What is a very minor mistake here would be much more regrettable in the code that implements ship detection, for instance.