I disagree. I think the thing it affects the most is his ability to get out of combos, which is already good since he's so floaty. While it does assist in his combo potential, he certainly has a strong combo game without it, particularly on fast-fallers. He doesn't have an amazing approach without it, but that's not to say he doesn't have a good approach. All of his tail moves have very good reach, so he can get in, just not for free. As for it being a recovery buff, it totally is. Even if the opponent can guess the direction that he'll go, he can still cover himself by doing an aerial (bair), which is pretty ridiculous.
Your second point is definitely part of what makes it good, but I do think the duration is still an issue. With float, you can pressure a shield eventually get a poke, even if your opponent clearly sees it coming. Also, I hate move that are super multi-hit and last long (like Ivy's nair). They slow things down, and they require a much more lenient accuracy/timing threshold. But yeah, your suggestion would probably work.
I find fair a far better option to escape combos, but maybe that's just a preference of mine. I used to play Melee Mewtwo, back before I picked up PM Mewtwo, and to me it's not so much an advantage for recovery or combo escaping. In Melee you could teleport to the ground anyway. Teleport does have ending lag, just not very much. It is very possible to hit Mewtwo before he can attack out of teleport, just difficult. Also, if you can predict where he's going, and know that he'll bair out of teleport, just wait for it and punish the bair. But I really don't see a good recovery as much of an advantage to begin with. Look at Melee. Falco is the second-best character in the game, yet has one of the worst recoveries. Mewtwo has the hands-down best recovery yet is still one of the worst characters in the game. Look at Brawl. Olimar has one of the worst recoveries yet is still number 3. Samus is one of the worst in the game yet has one of the best recoveries. Clearly onstage game is far more important than recovery.
Also, regarding the suggested change to teleport in the OP of making it put him in a helpless state if he's used his second jump, that
would absolutely affect his combo ability. For those of you who aren't Mewtwo mains, DJC (double jump cancelling) is a very important part of Mewtwo's combo game, allowing for strings of 3 uairs on most characters with uair -> DJC uair -> teleport -> uair. The problem with this change is that you couldn't attack out of the teleport to get the last uair, because you used your second jump on the DJC, giving him combo ability much more like that of Melee Mewtwo (aka bad combo ability).
Nair though... I watched the VODs of EmuKiller at SKTAR 3 and actually paid attention. The shield pressure is definitely over the top. I said the Melee top 4 were where I thought the PM characters should be. Melee Fox and Falco have insane shield pressure, but PM Mewtwo's is waaay better. Having paid attention to it, I think a change is necessary:
First off, nair has four separate hitboxes iirc, one in each corner. I forget which is which, but it's either that the top ones do 1% and the bottom ones do 2% or vice versa, with the last hitbox (the one that launches people) doing like 4% everywhere. For a hitbox to shieldpoke, it needs to hit the player's hurtbox without touching the shield hurtbox. Having four smaller separate hitboxes makes this much easier to accomplish.
On top of this, since the last hit deals 4% and is an electric attack, it has essentially the same shieldstun as Fox's Shine. Since there are only 4 frames from the 4% hitbox to when Mewtwo can input another attack, and fair comes out on frame 5, you have 8 frames minus shieldstun to have a hitbox come out before fair hits. I don't know how many frames the shieldstun is, but let's guess (it's really a guess. If someone actually knows, that would be nice) 5 frames. Now you have 3 frames to use an OoS option. That's not even enough time to roll.
So, one of my new propositions for a change to Mewtwo's nair is to make it a single hitbox that alternates between 1% and 2% every hit and is transcendent, and also increase the ending lag (but not landing lag) of it just a little (only a handful of frames), to allow more OoS options to hit after shieldstun ends and before Mewtwo can hit a fair/bair.
Alternatively, another possible change would be to keep everything the same on it, except make all the 1% and 2% hitboxes transcendent and reduce their shield damage to 0. Keep the last hitbox as is. So it wouldn't shieldpoke, but Mewtwo can still get shield advantage off a hover-cancelled fair/bair at the end.
These don't change its combo/punish utility at all, only shield pressure and approaching. I really don't think the duration is the problem. I, being a Mewtwo main, of course prefer the second, because it still lets you pressure the shield for the full duration of nair and get some frame advantage on a stationary target with another aerial. The first would require you either to read your opponent's roll (and tech chasing isn't Mewtwo's strong suit, given his lackluster dashing speed), or to back off before the nair ends if they have a nice up B OoS option.
I still say teleport is fine: escaping combos upwards rather than down to the stage doesn't really help; punishing his recovery is possible, just technically difficult; and it would affect his combo game too much and remove one of his best approach options.
Sorry long post.
