Ok, so 18% would be a weighted average increase across all cores. Technically, the entire multicore performance boost
could come from the efficiency cores alone. With the M1, the efficiency cores added
20-33% to the multicore score. Assuming the best case 33% number, the efficiency cores would need to be 54.5% faster to achieve that 18% multi-core increase. If the efficiency cores were 54.5% faster, that would make the M2 a dramatically different chip from the M1...
From your original post, the only thing that Apple "missed" was 4nm and that was from Digitimes, which is no more reliable than any random poster on this forum. A new chip running on the same process node doesn't mean they they are the same chips. There's no epic fail here (except for me wasting my time), plain and simple.