OK, let me try this another way because I sometimes struggle translating my brain through my fingers and we've gotten sidetracked from the point I was evidently failing to make.
The entire reason I've gone on this tier vs core count thing is because of the notion suggested elsewhere on this thread that if AMD have an 8c/16t Ryzen 3000 that matches or beats the 9900K then AMD should and will charge 9900K money for it. I've argued that doing so would be counter productive to AMD's sales because that would see AMD significantly increase their CPU prices beyond the point which would make sense to the general consumer, especially in an era where perception and mindshare is still skewed heavily against AMD.
Price/performance ratio in and of itself is a nebulous thing without any context. What I could consider poor price/performance ratio could be great for somebody else. And by extension then, if price/performance ratio is driving your purchasing decision, surely then you need to consider the price/performance ratio of other products and of competitors to make an informed choice.
Then this "undercut Intel by 20%" thing came along and it's confused the issue.
What I am trying to get at is this...
Let us ignore Intel even exist for a moment. AMD's Ryzen product stack consists of 3 performance tiers: entry level 3, mid-range 5 and top-end 7. Each one of those tiers has a price range applicable to the market segment they are targeted at. My argument is that the performance of the Ryzen 3000 series relative to Intel is largely inconsequential to the price brackets AMD will put to market. It doesn't matter if the 3300 SKUs get released with 6 cores and start trading blows with a 9600K, AMD could not justify in the eyes of the general consumer why their entry level range has suddenly doubled in price by price-matching their competitor's closest-performing part.
Now yes, AMD will have to be competitive with Intel in both performance and price, but AMD simply couldn't set a price band based on price/performance ratio to their competitor, otherwise the prices go through the roof and nobody would buy anything.
The only reason Cinebench is even mentioned is because that's the only tangible thing we've seen which gives any indication of Zen 2's performance. And in that carefully-controlled demonstration, AMD have a product that can match Intel's current top part.
So instead of asking the question again, let me try and do numbers. Hopefully that might clarify what I'm trying to get at. For purely argument's sake, let's say we have performance parity between a Zen 2 core and a Coffee Lake core, and AMD increase core counts in each product tier for Ryzen 3000.
AMD could take a 6c/12t Ryzen 3 and price it favourably against a Coffee Lake i3 because they are analogous to each other as entry level CPUs. That would see a Ryzen 3 price not much higher than its previous generation because that is the general expectation of how much an entry-level CPU costs. The absolute performance of that Ryzen 3 in isolation, and the relative performance compared to the i3 is then irrelevant. There also wouldn't be a need to undercut Intel by a whole lot, if any, because AMD will have the superior product performance at a good price.
or
AMD could take a 6c/12t Ryzen 3 and price it favourable against a Coffee Lake i7 because they are analogous to each as similarly-performing CPUs. That would see a Ryzen 3 price almost double its previous generation, even with a 20% undercut on Intel prices. And if AMD's lowest performance tier suddenly costs twice it did, how crazy expensive are the 5s, 7s and possibly 9s cost?
To me, the latter is utterly ludicrous and simply not realistic. Intel were lambasted for the price of the 9900K, do you really think AMD would commit the same PR suicide with a new, potentially market-disrupting, product?
