Based on discussion I've seen about the Rage mechanic, I'm starting to strongly feel that 1-stock is the only right choice for this edition of smash. Let me explain.
Rage has a weird effect on respawning players, which I didn't realize until other posters pointed it out. If two players are at 130%, and one gets a kill first, one guy loses a stock and respawns. The killer got the right read first and deserves a slight edge for that. But instead the killer gets a huge edge, because he's now no longer at kill % himself because his opponent's moves do less knockback at 0%.
Since comeback mechanics are anticompetitive, you might assume stay-ahead mechanics are opposite and therefore competitive, but both are actually anticompetitive. Both comeback mechanics and stay-ahead mechanics make tests of skill less important, with stay-ahead only the first tests are most important, with comeback only the last tests are important. It's a weighted average of skill tests instead of a true average of skill tests. The former fluctuates more because the additional tests don't counterweigh quite right.
Rage actually functions quite fine in time matches, the default for the game. And it also functions quite fine in one stock games. It's not a stay-ahead mechanic and one stock games, and is somewhere between neutral and slight comeback mechanic in one stock games depending on how you view it. It's kinda like since a tenth of your damage buffs your opponent's knockback, all damage values are just nine tenths as useful across the board.
A nifty side effect of one stock games is they don't have the respawn invincibility mechanic at all, which I won't assert as being bad for competition but I will assert as being unfun for me personally and an anecdotal handful of people.
Ask yourself, if nintendo patched the game to add the KO healer effect to every character, would you want one stock games? Because really that's what's happening under the hood, it's just less obvious. It's something that's going to become even more evident as players master this game at a higher level.
_______________________________________________________________________________________________________
An unrelated reason 1 stock format is good for this game is that it has so much gimping and SDs. The new ledge mechanic encourages some really aggressive offstage, and some of the spikes and SDs can have the feel like a home-run in a 3 inning game of baseball in a Bo3 set. 1 stock format obviously frees up time to go beyond Bo3, and using additional games is a better way of adding sample size on some of the high impact home run plays like Mario fair. If you land a Mario fair spike in 2 stock or even 3 stock Bo3, you just have to trade damage for the rest of the game to win, and then you only need to win 1 of the other two games while your opponent needs to win 2. If you land a Mario fair spike in a 1 stock Bo5 (or Bo7, Bo9..), you don't need to trade damage to finish out the win, but trading damage and breaking mostly even isn't hard anyhow. But now your advantage is reduced to needing to win 2 games while your opponent needs to win 3. That means that now you need to show you can pull off that Mario fair spike consistently and prove you deserve the win!
_________________________________________________________________________________________________________
The drawbacks to 1 stock smash is:
1: There's a new tempo to get used to, which is a little bleh.
2: The competition process itself is slowed down by the loading screen.
3: The stage selection rules have to be restructured
1 is an unavoidable cost, but one that hopefully evaporates after an acclimation phase.
2 is a cost we don't know the exact nature of yet for wii U. I have noticed that games load faster in Smash 3DS if players don't change characters though, so that's something to keep in mind. For speed purposes, the character counterpicking freedom should perhaps be reduced to keep the loading screen swift. Reducing character counterpicking would really just bring the format back to the same feel as 2-3 stock counterpicking, since you can't counterpick between stocks.
3 is an initial cost but not really a longterm one. Some games should probably just be played on a repeat of the same stage, for speed of play and possibly loading time effects too. The current stage selection method adds more time spent stage selecting the more games a set contains, but my favored stage selection method takes the same amount of time regardless of how many games are in the set (FLSS, game one on unstruck stage, game N+1 on the new stage that the winner of game N struck latest during stage striking.)
Well, that's my .02$