I cannot really see putting Lindros (or Clarke, either, but at least there's a case there) over Mark Messier. How is that defensible?
You're saying his peak or "best day version" is higher...? But is it?
Scoring finishes:
Lindros: 1, 6, 7
Messier: 2, 3, 5, 5, 7, 10 -- his "2" is behind Gretzky, and Lindros' "1" is a tie (with Jagr) while Mario was injured.
I say Messier wins that comparison. But okay, Lindros was injured a lot, so let's see...
PPG scoring:
Lindros: 1, 2, 3, 3, 4, 6, 9, 9
Messier: 3, 3, 7, 7, 7, 8, 10, 10, 10
This is probably the best (only?) argument for Lindros > Messier... but then again, Mess's two "3's" move up to "1" if taking out prime-era Gretzky and Lemieux (Lindros also jumps to another "1" in 1997 by the same system). But Lindros has a lot of games missed for many of those finishes beyond his two biggest in '95 and '97. It's the Mario-Lemieux-factor, where I'm skeptical of whether he could have maintained those paces for a full season.
Peak season scoring:
Lindros: 115 PTS in 73 games / 97 PTS in 65 games / 70 PTS in 46 games / 79 PTS in 52 games
Messier: 129 PTS in 79 games / 111 PTS in 77 games / 107 PTS in 77 games / 107 PTS in 79 games
I guess these are pretty close overall, and (again) you can make a per-game peak argument for Lindros... BUT in reality Lindros had one 100+ point season and Messier had six 100+ point seasons (and another at 99).
Playoff peaks:
Messier wipes the floor with Lindros here. Never mind six Stanley Cups to zero in Cup-counting, but just in per-game stats: Lindros had a great playoff in '97 (led in points), despite being partly shut-down in the Finals. The same applies to Messier in, say, 1990... but Mess won the Cup as team leader and captain.
And then, you add on Mess's 1983 (amazing), 1984 (Conn Smythe), 1985 (solid), 1987 (amazing), 1988 (amazing), 1994 (legendary), and Lindros is left crying like a baby.
Major individual Hardware / All-Stars:
Lindros: 1 Hart, 1 Pearson, one 1st-team All Star, one 2nd-team All Star (three of those four apply to one season)
Messier: 2 Hart, 2 Pearson, 1 Conn Smythe, two 1st-team All Star (center), one 2nd-team All Star (center), + two 1st-team All Star (left wing)
Needless to say, Messier completely trumps Lindros in career value, but I understand that isn't the argument you're making.
So, in sum, I can maybe see the argument for Lindros over Messier (by degrees) in peak value, if by "peak" we're going to look at about two isolated (and non-consecutive) seasons... although even then, it's debatable. But be aware that if you're using that criteria, it should also affect the way you rank everybody else on your list (so, now, Mario Lemieux should be ahead of Gordie Howe, for example).
If we're looking at a full season + playoffs, did Lindros have even one season as good as Messier's in 1990? Obviously not, because Lindros was never 2nd in scoring, a Hart winner, playoff scoring leader, and Cup winner in the same season.
I think, when the peaks are quite close (as Lindros and Messier are), and the career value / winning / longevity / hardware of one guy (Messier) completely wipes the floor with the other, it's a bit silly to rank the first guy higher based on a sketchy "peak" argument.