[-1] Beast Change
[-2] Hulking Size and Claws modified by Human Form
[-4] Supernatural Strength
[-1] Nimble  (no defense penalty by Hulking Size)
[-1] Supernatural Toughness (catch is poison)

Man, I got to ask, but what the hell is "poison"?  Like, I know the definition of the word, but what the hell does it mean in the context of a catch.  So what, substances that harm the character will harm the character?  Does a catch of "weaknesses" work the same way? 

But outside of the slight different in movement distance none of that has anything to do with the character and Sloth Man could pull of the same trick if his player decided to describe it as such.

Alright, question:  What prevents sloth man also describing his attacks as laser beams coming out of his eyes (assuming he doesn't have such a power, and that the description has the same mechanical effect as something he can actually do in the situation)?  I'm not being sarcastic or rhetorical.  It's really important to know how you'd answer this question to address your concerns.

Denarians:  I'd give everyone that touched the coin a shadow.  As a GM, I'd have said to the players something along the lines of "are you sure you want to have your character handle that?", just to hedge against a group splitting misunderstanding, but otherwise, let the Denarian started sewing the seeds for all kinds of plans within each person.  Of course, only one person can take up the coin, and doing so would give them access to much greater power, along with a lot more temptation than Harry had to deal with.  So who ends up with the coin at the end could be an interesting story in itself.

The temptation plan looks sound.  If the denarian grants a character use of powers, such as hellfire, then of course that costs refresh.  If a character actually takes up the coin, you might push the Demonic Copilot power on the character.  You should also heavily, heavily encourage (perhaps make it a requirement) a coin holder to take an aspect related to the Denarian.  This aspect is what you compel to have the demon offer its temptations.

Glamours:  As far as I'm concerned, you allowed the player to do a lot more with glamours than is strictly by the book.  Glamours grants veils, personal disguises, and the ability to create small fake objects, like a class ring or an I.D. badge.  If the player wants to be able to use the power as a substitute for a variety of combat actions, I'd suggest that they instead take channeling (spirit), which is more closely balanced for that sort of thing.  If it has to be a fae thing, and it has to give the functions that spirit channeling would, there is nothing preventing the player from taking the channeling power and calling it glamours.

Consequences:  Well, they grant one free tag, so remember that.  And let players pay a Fate point to the opponent to compel the consequence if you think its a good compel.  But otherwise, yeah, consequences are a good way to extend a fight without reducing fighting capacity.  Varying the amount of consequences a thing will take is a good way to model how important and serious a fight is supposed to be.

Resources:  You can only use the Buying Things trapping of resources once per session.  Thats a pretty effective way to stop it from dominating.  But if a player is making that one of their highest skills, then its supposed to be powerful.  So don't completely shut it down by doing things like "your bank accounts gone".  Thats just punishing the player.  Oh, and call for a Lore roll when he tries to buy something that would rely on knowledge of what hurts supernatural things.  And call for a contacts roll when trying to buy something illegal.

Target Numbers:  If you just don't know what the difficulty is, make it 3.  This lets characters that are experts at the thing usually succeed, and unskilled characters usually have to pay Fate points to succeed.  Don't make them roll for something just because there would be a logical chance of failure.  Only make them roll if you would be interested in seeing that failure occur.

Of course the way I play it is that both players continue to spend fate points until they are both satisfied with the result (or are out of fate points). It works like your method without the unnecessary procedure.

The procedures are identical.  Mouse just provided a more analytical description of what is happening.

About two compels per session seems about right, but an average of 4 might be more reasonable if you were particularly active about getting into trouble.  Personally, I find that if I haven't spent my fate points by the end of the night, I feel like I didn't get enough done.

Here's a nice parallel:  Players are in control of whether their character lives or dies.  The system is designed so that the character only dies if the player wants them to. 

Scenario:  Player has character attack a monster, with the GM forecasting the monsters intent of killing the player.  The monster hits him with 12 shifts of damage on the first attack, and the player doesn't have consequences to absorb the attack.  The GM decides that the character dies.   Does this invalidate the idea that players decide if their character lives or dies?  No, because he had options available to him prior to rolling the dice against the monster that would have guaranteed (and that the player knew would guarantee) the characters living.  The player certainly didn't fight the monster with the intent of dying, but he accepted the risk of it happening anyways. 

I see the "players always have the choice to kill a character or not" as roughly falling under the same distinction.  You can always describe your victory as killing the other character, or letting them live, but the way you go about securing that victory puts limits on how you describe the victory.  Using weapon: 5 attacks is taking a risk by saying, "I may end up killing him, but its important enough to take him down that I'm willing to sacrifice some control over the outcome".  There is a silly place that I see being defended where the character loses control over the outcome, but that loss of control never translates into a real consequence for the character because the player always gets to negate that loss of control through pure metagame.

(bolding added)

Like what?

Like if the hypothetical player has the aspect "Always lets the bad guy get away".  Thats a blunt scenario, but if you refine the situation enough, you can easily get circumstances where less obvious aspects can be compelled towards the effect of the bad guy getting away.