Broadly, I agree with what you're saying. Totally just devil's advocate-ing and speculating to provoke thought, so feel free to ignore. I wonder if the enormous number of games available plays into this. I can almost always dig around and find at least one 10/10 game from the last couple of years that I haven't played which is already on sale for cheap. Comparing that to a 7/10 game that just came out at full price... I'd almost certainly enjoy the 7/10 game, but I'd spend less money and likely have more fun with the 10/10. The newness factor may not be enough to bump the 7/10 game to the top of the queue.

With so many great games available an 8/10 might actually feel like a logical minimum for a lot of people, which may influence the scale that reviewers use. If people tend to ignore games with 7- scores and a reviewer feels that a game is good enough that it deserves attention, they may be tempted to bump it up to 8/10 just to get it on radars.

Meanwhile, back in the day there wasn't such a glut of games to choose from. And with better QoL standards, common UX principles, code samples, and tools/engines, games may legitimately just be better on average than they used to be, making it fiddly to try to retrofit review scores onto the same bell curve as older games. To reverse it, I can see how an 8/10 game released in 1995 might be scored significantly worse by modern reviewers for lack of QoL/UX features, controls, presentation style, etc, or even just be scored lower because in modern times it would lack the novelty it had at the time it was released.