You're not giving Brad Richards nearly enough love. He had one of the most clutch postseasons of all time. NHL record 7 game winning goals in the postseason. To put that in perspective, the NHL record for an entire regular season is 16, he managed nearly half that in 23 playoff games. Your argument that a couple of the GWG weren't important because they were early in the game is pretty silly, scoring the only goal in a 1-0 Finals game is as important as it gets.
During the 2003-04 season, the Lightning were 31-0-2 when Richards scored a goal, 9-0 in the postseason. It was just one of those magic years when everything fell into place for him.
And talk about clutch, with the Lightning facing elimination in Game 6 of the finals, he scores the only two goals in regulation for the Lightning, bringing it to overtime before Marty St. Louis wins it (assisted by Richards of course).
Give Richards some credit, he might not have been able to sustain it his whole career, but for this one year everything he touched was gold.
I had Richards at the high end of the 3rd tier (along with Yzerman). Definitely a reasonable argument could be made that he's in the lower end of the 2nd tier.
The only part I disagree with (which I talked about in the initial post) is that the GWG statistic isn't always meaningful.
Two of his game winners (April 12th vs Islanders and May 31st vs Flames) were scored in the first five minutes of the first period, and Khabibulin had a shutout. Is the first goal important? Of course, but that's not really what I think about when talking about game-winners (usually you think of an overtime goal, something late in the third period, etc). Or May 18th vs Flyers is another example - he scores to make it 3-0 in the second period, the Flyers mount of comeback that falls short, and that goal (which nobody would have considered clutch at the time) retroactively becomes the game winner.
That's not to say he didn't score important goals (two OT winners, game 6 in the finals as you mentioned - as did in my first post) - just that the argument that he has the all-time record for game-winners doesn't accurately represent what happened on the ice.
In contrast, in 1999 when Nieuwendyk had six game winners, that stat really did show that he scored big goals consistently. (Richards was much better than him overall, though).
One argument in is favour - Richards seemed to be the obvious choice for the Smythe (maybe there were arguments for Khabibulin, St. Louis, or even Iginla, but they'd be fairly weak). Maybe he does belong at the bottom of the 2nd tier after all.