2700x is $330 WITH a good HSF. 9900k will most likely be at least $500 and no HSF, so like $200 more for 10% IPC or less.
It's 'good' for stock frequencies and turbo boosts, but also a bit noisy under load. It's actually a very decent stock HSF but not that suitable for overclocking, not that you can overclock a 2700X much even with upgraded cooling.
Also, how many enthusiasts (the type that would buy a 9900K) actually don't own a half decent HSF or AIO? I have about 3 spare heatsinks that I can run on my 8700K...
Im not sure why you adjusted the 9900K price upwards $50 to 'over $500' when it's been rumoured to cost $450.
You also rounded down AMDs IPC deficit to '10% or less' when stilts testing puts it close to 15%, though without AVX included it is around 10% so I'll give you that.
Ultimately, for enthusiasts at least, what sets the 9900K apart from the 2700X is the clockspeed ceiling. The 2700X averages around 4.2GHz overclocked - the 9900K, with a soldered IHS, could conceivably clock a full 1GHz higher... that's probably worth more for overall performance than the IPC advantage.
This is also the reason I struggle to see how AMD can really counter the 9900K with a '2800X' unless they have drastically improved the 12nm process. It is already running close to its limits even on the 2700X.
It's akin to AMD trying to clock an already power hungry Vega 64 even higher to try to compete with the 1080 Ti - at some point, the power curve just makes it unfeasible to keep adding clockspeed. I think the 2700X is already at that point. Is a 150W+ 2800X possible? Technically, it is. AMD can cherry pick the best dies and add an extra 200MHz perhaps , but it still won't really come close to a 9900K, so what is the point, really?