We need to reevaluate 6v6 class limits.
The TF2 6v6 metagame has been originally molded by trial and errors from the O.G. teams to find the optimal compositions, finally settling on maximizing speed and versatility. Class limits were imposed to restrict domineering overpowered styles, beginning by limiting the Medic (Uber is OP) and the Demoman class (then ultra OP).
Every other composition that deviated from the cookie cutter was automatically inferior by very design, but class limits began to be imposed not out of a fear of overpowered styles but because of their situational gimmicky usefulness. Engie, Heavy, Sniper and Pyro were all restricted to 1 despite them having rarely been at all viable outside of specific situations. The Sniper was limited to 1 for example because Epsilon whined that Broder's 2 sniper defense on Process last was too OP, even though it was the only time when running double sniper was ever viable.
The nerfing of the Demoman and the accelerated dominance of the Scout class, together with the uselessness of other classes, should leave us to naturally ask whether or not it is time to reevaluate class limits.
Scout is now the strongest class in 6v6 and yet it has a higher class limit than Heavy, Pyro, Sniper and Engineer. The only non-standard class that has survived the chop has been the Spy, but only because no one has yet figured out how to amount to anything more than complete and utter shit with it.
The current design of class limits is strongly biased on tradition and anything that threatens it gets the chop. This is antithetical to the competitive spirit. CSGO Majors do not set restrictions on the amount of AWPs or Auto-Snipers their players can use, nor are the organizations conditioning the players to a specific playstyle. This is all chosen by the professionals themselves. In much of the same spirit we should not so viciously ban, ban & ban just because we are too lazy or scared to adapt to new strategies.
Let me run double Heavy to mid and let me defend last with two Snipers. It's none of your business to ban them if it proves effective in a niche situation and it's neither none of your business to advocate banning it "because it's dumb". If it works then it's not dumb. If you're convinced that it will fail then allow it at least the chance to fail all of its own.