This is filtered to exclude any year where a player received less than 5% of the available votes (the intention is to avoid overstating their results based on a few random votes). Note also that this data goes back to the end of WWII (so it captures all ten players' entire careers).
While it might not affect it now, it is worth noting that the 5-3-1 ballot could prove restrictive with this type of comparison in some years, as seasons with three strong candidates can leave other great seasons fighting for those scraps and under 5%.
It’s a little weird seeing a chart with Patrick Roy’s 8th place finish in 1996-97 on a 10-7-5-3-1 ballot but not his 4th and 5th place finishes in 1988-89 and 1989-90 underneath Gretzky, Lemieux, Yzerman and Messier, Bourque, Hull, Gretzky. Bourque, himself, is in the same boat in 1990-91.
Had there been a 10-7-5-3-1 ballot, they both likely retain their top-5 positioning while more names are able to meet the threshold because wider ranges of opinions are recorded. Gretzky himself barely met the threshold in 1989-90 because of the 5-3-1 ballot, as the top-3 took up 166 of 189 spots on the 63 ballots - when an expansion of the ballot to reflect the expansion of the league would have capped the top-3 players at 189 of 315 spots.
1989 is probably the most egregious example with Gretzky, Lemieux, and Yzerman holding 185 of 189 spots on the 63 ballots, so there wasn’t even 2% to fight over, while a simple and much-needed ballot tweak 7 years later guaranteed 40% of spots to the field beyond the top-3.
I get the idea of the fear of rogue opinions tainting the sample (Naslund’s surprise 1st place vote in 2001 only got him as far as 11th place), but it’s probably the sort of thing worth an asterisk, because once we get more 1995-96-onward players, it’s much easier to meet the 5% threshold.
Players with 5%
1980: 4
1981: 2
1982: 3
1983: 5
1984: 5
1985: 5
1986: 5
1987: 5
1988: 5
1989: 3
1990: 4
1991: 3
1992: 6
1993: 4
1994: 6
1995: 5
Change in Ballot
1996: 9
1997: 9
1998: 7
1999: 9
2000: 7
2001: 10
2002: 8
2003: 8
2004: 12
Essentially, by having a percentage cutoff but not a ranking cutoff, it will under-report seasons on a 5-3-1 ballot and over-report seasons on a 10-7-5-3-1 ballot. With respect to Mats Sundin, I’m more comfortable with Patrick Roy’s 4th place Hart finish in 1988-89 than Sundin’s 12th place Hart finish in 2003-04, but Sundin would theoretically make the chart.