gkelly will nitpick me to death and perhaps contact the FBI to have me assassinated, but I don't think the illusion entry is any good as a "difficult entry". Many times, it looks more like the skater did the illusion to torque themselves into the spin.
Lock your doors -- the assassins are on their way.
I also don't think the knee slide exits are "difficult". I'm not sure what I'd call 'difficult' exit for a spin, unless it's one of those few times we've seen a 3S done straight off a spin (which would double as an unexpected entry to the sal). A well controlled exit is probably much more difficult to do than some of these 'difficult' exits.
What I'm especially impressed by are spin exits that check out onto a controlled edge and then add other turns or edge work on the same foot. Definitely more difficult than just holding the controlled edge for a second or two before stepping to the next move, and also more skating skill-based than knee slides or flexibility moves on simple edges (which are still more difficult than simple edges alone).
Another risk of knee slides as spin exits at the end of a program is that if the skater has stopped moving any muscles by the time the time limit is reached but the still body is still sliding across the ice, they can get a time deduction.
I've said it before, but I want some features of the spins to be double counted - like "difficult change of position". Change of edge on positions like the Biellmann also deserve to be double counted.
Well, a difficult change of position in which one of the positions also counts as a difficult position earns two features*, as does change of edge in a difficult position.
(*Or three features if both positions before and after the change are difficult, in a combo spin with no change of foot. But skaters rarely plan a CoSp or FCoSp any more since a CCoSp or FCCoSp have higher base values, and the CCoSp is required in the short program.)
Some difficult positions are more difficult than others, and similarly change of edge in some positions is more difficult than in others. And some difficult entrances or exits are more difficult than others. Basically, the IJS says "this counts as a difficult feature" and then realizes that some ways of meeting "this" are not particularly difficult, so they remove that option from the list of features entirely or they specify that the feature only counts in some positions but not others.
Meanwhile, change of edge in simple upright spins and change of direction with a change of foot in simple upright spins no longer count as features at all.
Long ago at the very beginning of IJS, I once saw a skater execute a change of direction in an upright spin all on one foot. It wasn't during competition though. Now that is something I would have loved to count as a double feature! Too difficult even for that skater, so I never saw it again from him or anyone else. If the reward had been commensurate, we probably would have.
With difficult entries/exits, skaters put out options that they can do that are at least slightly harder than not doing them, and then the tech panel community needs to come to consensus about whether those options really are difficult enough to be rewarded. Which can lead to some inconsistency between panels at first. And then changes in the guidelines to specify what does or doesn't count. E.g., illusions no longer count as difficult entries/exits and don't count as difficult variations unless they achieve at least 135-degree split.
It's so different and fun and so them. 
I know there are some of you, I saw you in the PBP thread last night!
I've heard it slammed all season, which is why I was glad to see fellow fans.
