The league varies in strength and in performances of top players from year to year. You have to take that into account, somehow.
Well, the one thing I would always keep in mind is that one season, alone, is basically meaningless.
On the other hand, as
@jigglysquishy points out, if you compare samples of more than five years or so, you're getting into sketchy comparisons because of how the game as a whole---notably scoring environments, etc.---can change.
So, to do statistical comparisons of "peaks" (as per this thread), I think the safest way is to compare periods of about three seasons. Two is optional, four is probably better. But I can live with three.
I find the past ~4-5 years to be extremely competitive in regards to high end performances.
Well, since 2019 (I think), we've seen scoring go noticeably up and we've had Art Ross winners with 128 points, 154 points pace, and 153 points. The same goes for defencemen's scoring peaks.
It's pretty clear that McDavid (and also Kucherov, MacKinnon, Makar) are all-time elite players by peak level. So, there's that. But then we always have to consider how the scoring environment / strength of competition affects this. This is where it gets tricky. Generally (but
not always), when scoring peaks go up, it's in part because strength of competition goes down---and indeed, overall scoring has gone up since about 2019. Then, the past few seasons have had some of the lowest NHL parity since the early 1980s, with some truly awful teams. Just last year, we saw Erik Karlsson and Steven Stamkos have their biggest productive seasons ever, and those two were previously thought to have peaked around 2013 to 2015.
The thing is, how do we objectively determine when elite-scoring is
more competitive or
less competitive? It's not a measureable thing, which means we're all generally bound by our subjective biases.
As an example:
Crosby's best years? Jagrs? Howe's? A bit less so. There as a bit of a lull in the late 90s/early 2000s, and same in the early 2010s. And the early 50s as well, definitely.
Gretzky/Lemieux is a similar effect. Gretzky's best years there's a bit of a lull at the top, so his %'s over competition is insanely high. Lemieux? His best years in late 80s/early 90s, competition is a lot higher, and so his %'s don't stand out as much.
We've seen it often said by the "Howe-wasn't-that-great" crowd that the early-1950s wasn't a strong period for elite talent / scoring. But, I mean... how can you determine that? Basically, unless you were watching a lot of games then, of all the different teams, you can't really say. And even if you were watching all the games, it's still going to be a totally subjective opinion.
Also, how was there a "lull at the top" during Gretzky's best years? Are Dionne, Bossy, Stastny, Nilsson, Kurri, Coffey, Lemieux, Messier, Savard, Hawerchuk, a "lull" in high end talent?
Lemieux's late-80s / early-90s competition (besides Gretzky... who outscored him from mid-1989 to mid-1991) was Yzerman, Lafontaine, Messier (same guy), Oates, Robitaille, Nicholls. Is this demonstrably "harder" high-end talent than Gretzky faced?
So, I dunno. A lot of this is just viewer-subjective bias.
I think, for scoring peaks, about three seasons (or four) is the safest way to compare.