I've updated the tables with stats through the 2016-17 season. Some stars of the 2010s like Erik Karlsson, PK Subban, Steven Stamkos, and John Tavares are now included.
I made a minor methodological change to weight all stats by games played instead of seasons played. The reason was to put less weight on the shortened lockout seasons.
Also, here are some stats by decade to show how player usage has changed for forwards and defencemen.
Average stats for all defencemen, minimum of 300 GP in the decade (250 GP for the partial decade of the 2010s and 250 GP for the 1960s shorter seasons)
| EV% | PP% | SH% |
| 1960s | 42% | 26% | 46% |
| 1970s | 39% | 27% | 43% |
| 1980s | 35% | 32% | 37% |
| 1990s | 34% | 32% | 36% |
| 2000s | 33% | 28% | 37% |
| 2010s | 34% | 28% | 37% |
[TBODY]
[/TBODY]
Average stats for all forwards, minimum of 300 GP in the decade (250 GP for the partial decade of the 2010s and 250 GP for the 1960s shorter seasons)
| EV% | PP% | SH% |
| 1960s | 30% | 39% | 17% |
| 1970s | 28% | 35% | 18% |
| 1980s | 27% | 33% | 18% |
| 1990s | 27% | 32% | 18% |
| 2000s | 27% | 35% | 18% |
| 2010s | 27% | 34% | 18% |
[TBODY]
[/TBODY]
You can see the smaller rosters of the 1960s meant a higher EV% for forwards and defencemen. This was also the case for the 1970s, to a lesser degree. The modern roster size was in place for the 1980s and on, and as a result the EV% are directly comparable for those decades.
The 1980s and 1990s were the decades where forwards rarely played the point on the power play, so defencemen had their highest average PP% in those decades and forwards had lower PP%. 4 or more forwards on the power play was most common in the 1960s, less so in the 1970s, least common in the 1980s and 1990s, and made a bit of a comeback in the 2000s and 2010s.
SH% was higher for defencemen on average in the 1960s, when teams played only 4-5 regular defencemen. Similar to EV%, it declined in the 1970s and reached a level in the 1980s that it hasn't left. You can also see the effect in the 1960s of Montreal playing 3 defencemen on the penalty kill as the average SH% for forwards was a point lower.
For defencemen, I would say EV% and SH% numbers pre-1980 shouldn't be directly compared to EV% and SH% post-1980. Keep in mind the lower roster size meant teams did not play 6 regular defencemen in the 1960s and 1970s.
Also, the league-adjusted scoring numbers, whether EV or PP, may be affected by other factors that make it easier or harder for top players to stand out. The most obvious factor is the effect of the 1967 expansion and subsequent expansions on diluting the league strength and lowering the average.
