Just remembering what "screwing around" kind of testing I was doing on 2 vs more cores in approx 2012.. meaning I didn't keep records of what exactly numbers per app/game were or even what particular ones. Anyway it was on Win7 but was on stuff that wasn't all that recent, maybe up to a year or two old at the time, and it was all probably stuff that would run on XP. Where 2012 may have been the cusp of where multicore support was improving for latest new stuff.
What kicked it off was swapping an X2 at 2.6ghz out for a Phenom X3 and being super underwhelmed. In theory I had a whole gigahertz plus extra CPU power but nothing was showing me that, everything was slower, everything. Now I knew that old old stuff would only use one core so no surprise on that, since the cores were 2.2ghz vs 2.6ghz. What I thought should have been decently multithreaded though, was only loading the first 2 cores and getting like 20-25% load on the third. When I got a Q6600 a few months later I tried the same things and "fully loaded" that was going to 100% 100% 25% 5%. So given 3 or 4 cores could only do 20-30% worth of a core's work beyond 2 cores, then if you get a dual core that's 10-15% faster clock than the quad or trip, it will perform about the same in pre-2012ish software. So you're doing better with a 3Ghz C2 or X2 vs a 2.4 Q6600 or 2.4 X4 ... Benchmarks though, were progressively more multicore aware at least 3 years earlier than anything else, so benchmarks will tell you "lies" meaning yeah that's the potential, but it's not what software the same age will actually get. Anyway early tens 2 vs 4 core thing very much like mid noughts single vs dual core thing.
These comments not meant to apply to "power user" heavy multitasking or productivity, movie editing, graphics work, compiling etc etc, which apps tended to be optimised early and different kettle of fish to gaming use.
Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.