A strategy game is supposed to be designed around strategic decisions. Hard limiting anything is the opposite of this, because it takes away decisions. To bring T:L as (only) example of hard-limited game after bragging about the amount of them to exist is kinda ironic, by the way, because everyone bypasses the 60 member limit with ease by creating wings and metas. Your only example doesn't even work, because metas there have way more than the limit would allow. They just are spread out into multiple alliances. Accusing me of having hidden intentions with my con-limit mindset is pretty stupid, I already said I wouldn't be affect at all. We [kingdoms I lead] always stand against a significant higher member count, not having surpassed 45 members in total including confederacies and naps (which mostly didn't exist) for a few years. This is true after union (since unions exist) and during 3-kings-per-ally era. :p
Also I said explicitly, that I don't like huge metas, but that hard-limiting is still bad. One needs intelligent solutions to make the players want to play differently than before, instead of closing both eyes and dumbly throwing a hard-limit in, ignoring everything speaking against it.
The dropping player base "argument", which I read kind of everywhere for everything, is also kind of useless. T:L's playerbase isn't growing either, should I come in and say "all hard-limit's fault" now, like you do? Even if T:L's playerbase was falling apart significantly slower or not at all, it's quite bolt to blame an existing hard limit for that. That's not the only difference between those two games, you know? And not even near the biggest.
Guess you're a little bit upset, because I said your statistics didn't reveal anything non-obvious. This was a bit harsh, I know, and I already said I feel sorry for saying this, but despite being interesting statistics, it's quite true, because the linear dependency between member count and victory points in a peaceful round is indeed obvious if you think about it for a while. Nobody said metas aren't too strong, this sometimes seems to be forgotten.
Like I said, Ammanurt, in a strategy game you should be able to make decisions. You should have all freedom you can possibly get, instead of enforcing one particular way to play. Currently building a meta with huge player counts is a kind of enforced decision, because it's very op and you have to work incredibly hard as smaller kingdom with significant lesser members to beat the huge one, while they can chill on their passive treasure production. This is obviously bad, because again, that's no decision, nothing you have to think about whether you do it or not, what the advantages and disadvantages are. There aren't significant disadvantages right now. But hard-limiting isn't a decision either. Let alone the fact, that, like we (we con-limit guys) pointed out dozens of times, there will just be a wing.
If you want me to stick to your lever-analogy, I'd say it's more like a mixing console (for sound, not sure what the correct english term is), where you can change the strength of certain playstyles with the goal to balance them out. Right now, meta is very strong, now someone needs to change some mechanics to balance it out again, such that metas are weaker (or other playstyles are stronger). But not pulling a lever with an emergency stop sign to ban meta playstyle at all.
The reason I post so much here, by the way, is, because I want metas to be nerfed, aswell. And since we're a lot of people here, who have some experience and similar thoughts on the opness of metas, one could maybe discuss better design options against them, instead of trying to ban them. I don't see, from your side, why you would mind metas being nerfed instead of removed, by the way. You wouldn't see so much of them then aswell, and if you do, you can defeat them more easily.
By the way, a little fun fact for you all, to relief the tension in the atmosphere here: As far as I know, there is indeed a member limit right now. 111 players for non-united kingdoms, 222 players for united ones. Can't tell with 100% certaincy, because this information is a bit older, but if it's up to date and consequently implemented, this should be true.
or 
