I like the concept of a luck-restricted ladder, but I don't like your take on it, because you have distorted the mechanics too much. Other changes seems unnecessary. This is what I think about the Mechanistic changes:
1) Sleep counter should reset when you switch. I don't find the need to change this mechanism, when it is not even luck-oriented. 2 turns are good enough.
2) Phazing - let it be random. I don't think I ever heard people complaining about phazing in a wrong Pokemon. This change seems unnecessary and exploitable.
3) I personally feel that pokemon should be protected from critical hits for the first 10 hits, regardless of defense boosts or no defense boosts (stuff like Chansey or Blissey, Sub-Roost Zapdos, or Jellicent can still stall without defensive boosts), but after the 10th hit, a critical hit (negate boost & double damage) will happen every 10 hits. If the stalling pokemon switches, this count is reset to 0. I prefer this rather than "negating defense boosts after the 20th consecutive hit," because I feel that it doesn't solve the problem - the stalling pokemon would still successfully stall if it could stall for 20+ turns. This can also be exploited in ways that are not conventional to the traditional playstyle, and personally I would like to avoid introducing odd playstyles while minimizing luck.
Ways that this could be exploited:
The stalling pokemon can be simply switched out after the 20th hit and can come back to stall for another 20 turns some other time.
In a CM vs CM war, it will simply be a race of who hits 20 times first, which to me sounds absurd
Usually people complain about critical hits that happens in the first 1 or 2 hits. I think a 10-turn protection from critical hit, and then 1 critical hit / 10 turns to punish stalling Pokemon is the better way to go.
4) I also prefer Thorhammer's suggestion about chances / accuracy. Moves such as Lava Plume and Discharge sacrifice power over a higher chance to cripple the opponent. In the other hand, moves such as Fire Blast and Hydro Pump sacrifices accuracy for power. Rather than making this all or nothing, I prefer Thorhammer's suggestion of the chance of a secondary effect or a miss every certain amount of turns.
The one thing I don't like about Thorhammer's suggestion is that the opponent can anticipate when the status or miss is going to happen, and could exploit this to their advantage. This will severely undermine the reason to use Lava Plume over Flamethrower + Will-O-Wisp.
This is my take:
0-19: Don't expect this side effect to happen for the first 5 uses. However, following those 5 uses, expect it to happen once every 10 uses.
20-29: Protection from side effect for the first 3 uses. However, following the 3 uses, expect the effect to activate once every 5 uses.
30-49: Protection from the side effect for the first use, but expect the effect to activate once every 3 uses.
50-69: Activates once for every 2 uses. This includes Confusion, such as from Outrage.
70-79: The move or effect will happen upon first use, but expect a miss once every 3 uses afterwords.
80-84: The move or effect will happen upon first use, but expect a miss once every 4 uses afterwards.
85-89: The move or effect will happen upon the first 2 uses, but expect a miss once every 4 uses afterwards.
90: The move or effect will happen upon first 2 uses, but expect a miss once every 5 uses.
91-99: Never miss.
This will prevent the first-turn misses / consecutive misses / early activation / consecutive activation that constitute hax. Your ladder has no hax, but it implements extraneous features that makes it too warped from the actual BW OU. And this could hinder the appeal of this otherwise cool project.