I feel that with 6 teamslots, it's very difficult to counter every set of every threat but ultimately, the goal of a good metagame should be to
a) Allow a team to counter as many threats as possible
b) Allow for many variations of teams to achieve this
Let's say in a metagame where the top 10 threats each only had 1 counter where none of those counters were shared. The game would basically be pick a team of 6 counters out of 10. This is a bad metagame because not only is there no possibility for team diversity, you can only cover 6/10 of the top threats at any given time. While maybe not as extreme as this, this is where our metagame is at at the moment because some of our top threats have very few counters and checks.
Now if say we were to get rid of these 10 threats. The next 10 top threats to emerge all have 10 or 20 counters/checks each and a lot of those DO overlap. There are infinitely more combinations available for teambuilding and it's up to each individual team builder to find the right combination of 6 pokemon to cover as many threats as possible. A very good team builder might be able to cover 9/10 threats. Another might be able to cover 8 and yet another might be able to cover an entirely different 8 threats. That's the type of metagame we should be aiming for.
I agree with your basic examples' principles, but if the top 10 threats each have 20 counters/checks, and they overlap, then those threats would very quickly become ineffective, and would no longer be the top 10 threats.
Furthermore, if the top 10 threats all share the same 20 counters/checks, it probably means that they are all the same kind of Pokemon - such as SD sweepers, Fighting-types, etc. Having our top 10 be mirror images of each other wouldn't be that great, IMO.
----------
I wish people would stop saying that you need counters for everything. Word choice, people, word choice! Counter is not the word you're looking for.
Here's an example of what I mean:
Jolly LO DDMence has literally 4 counters (5 counters if you factor in Leftovers). They are: Cresselia, Hippowdon, Eviolite Porygon2, Eviolite Dusclops, and Suicune (if you count Leftovers).
Literally everything else can get 2HKOed by either Fire Blast, un-boosted EQ, or un-boosted Outrage (some things don't, like Regirock, but are so bad in the metagame that I didn't count them). However, not every team needs one of these; on the contrary, only two of these are even OU! However, Salamence is not broken by any means. Some teams will force it to use Outrage, then switch in a Steel-type. Some teams will carry a revenge-killer for it. And so on and so forth. A team doesn't have to "counter" each and every threat!!!
Furthermore, why do you have to deal with every threat at all? It would be a total waste of my time to prepare myself for Altaria. If I don't prepare for it, it could hypothetically sweep me with a DD set. But Altaria is one of many Pokemon (almost 600!) that don't see the light of day in OU. My teamslot would be better spent preparing for something else.
Which brings me to my next point. Myrmidon touched on it, and it is overlap. If I prepare for something like DDMence and DDNite, odds are that I'll be safe against that DD Altaria I mentioned. If I prepare against CM Latias and CM Reuniclus, odds are that your CM Gardevoir won't sweep me. If you cover the top threats, then you'll be covering a ton of other things anyway, even though you didn't mean to.
Lastly, this whole "you need to deal with things" mentality is good for teambuilding, but only to an extent. Someone once wrote a grand something on this topic. It went to the effect that it you focus on countering the opponent's Pokemon, they'll be leading you the whole match. They'll be controlling the whole match. You should instead worry about executing your own strategy, to actually win the match. You want to beat your opponent, not counter them.
It's important to be able to handle the top threats in the metagame you'll be playing in, but idea of "countering all the threats" is just not how things work. 6 teamslots is enough for this metagame (good thing, too, because it's all GF gave us!).
@astrohawke:
You asked what kind of metagame mien finds most preferable? In another thread, he said so. He told us that out of all the characteristics of a desirable metagame, the one he holds up highest, values most, is stability. mien doesn't want us to be testing and changing the metagame for the next few years. He wants it to stabilize. At least, that's what he said a few weeks ago.