You take things to such extremes. Either that are only know perfect people and only play with perfect people. The examples I mentioned are one aspect of an other wise fun person to play with. So instead of saying get out of my game because you do one thing that bothers me I take precautions to keep the game fun for everyone at the table including them. The guy that cheated on his dice rolls moved away and I get emails asking me to run a game on roll20 because he misses playing with me as a DM. I would still play with him and was sad to see him move away because other than the need to cheat on dice rolls he was an engaged player who made the game fun for everyone at the table.
The powergamer who I never say yes to right away is also a great role player. I love having him at my table. He gets very involved in the world he writes game journals and when I say no or when I make modifications to what he wants he accepts it.
Why would I not play with him when I have a perfect solution that works.
Well, part of that is the nature of the Internet. I only know your group based on what you say, so, based on that small slice of knowledge, I can only offer limited advice.
You have described four pretty dysfunctional gamers - 1 that flat out cheats, 2 whose play style conflicts so badly with your own that it makes the game unplayable and one who will sabotage any game where he doesn't get to play what he wants. Now, with further information, maybe these are only minor issues. Fair enough.
But, that also makes your advice really only apply to your group. You didn't say, "Limit the options of power gamers when the power gamers also bring their A game to the table". You just said, "Limit options to force power gamers to not power game". Why not give the advice, "Hey DM's, learn the system so that the power gamers in your group can power game without breaking the game"? Why is that not just as useful advice as, "Well, we have to force restrictions on the players so they don't get advantage over the game"?
My entire problem with this whole thing is that it's presented as, "Players must always compromise with the DM but the DM only has to compromise if he or she feels like it. If compromising would bring something into the game that the DM doesn't like, then, well, tough noogies, don't compromise".
I think that's very toxic advice to many gaming tables. I believe that unless the DM can convince the players with well reasoned arguments, then what the player wants is probably fine for the game. And no, I don't believe that "Well, I don't like it" is a valid argument for a DM. If the player really wants a tiefling and the only issue is an aesthetic one? Suck it up and let the player play what he wants. It's not your character, it's his. He's the one who has to fill that roll for dozens if not hundreds of hours. I dislike strongly any advice that says that the DM is entitled to enforce his personal tastes on the group.
....
[MENTION=957]BryonD[/MENTION], sorry, no, I don't see the issues as separate. A DM who is that laser beam focused on his campaign that he cannot adjust it before the campaign starts to accommodate fairly minor changes - adding in a one shot race character isn't that hard after all, we do have Plane Shift spells in the game - IME will not suddenly become completely free and easy after the campaign starts. If he won't allow you to take something he doesn't like, not because it will break his game or invalidate the campaign in some way, but solely because of his own personal tastes, what makes you think he's suddenly let you do something that doesn't fit with his personal tastes once the game starts.
I'm sorry, but, I've seen it far, far too many times to believe that it's an isolated thing. Granted, it's entirely a taste thing. The second I sit down at a table and the DM starts getting very controlling about exactly what gets created at chargen, alarm bells start going off in my head. And, again, IME, after the game starts, that level of control over the game won't stop.
Now, it may be that you and others aren't like that. Fair enough. But, I'll stand by this prediction. The presence of exceptions doesn't mean that I'm wrong.