I'd give it to Malkin for 2009.
He had an Art Ross winning season followed by arguably the best Conn Smythe run of this century to date. Peak Malkin was also better defensively than peak Ovechkin (controller disconnected) as well.
You're right, Malkin did lead the league in takeaways in 2009, and was pretty good in all 3 zones, as opposed to Ovechkin who really only put in the effort on one end of the ice.
Let's take a closer look at this: 1.39 for Ovechkin and 1.38 for Malkin, which is equivalent to less than 1 point per 82 game season.
Lemieux often gets docked for his (post-prime) 1995-96 regular season because he sat out games where he was healthy (eg. back to backs), which gave him an artificially elevated PPG average on the year. The same can be said for Ovechkin in 2008-09, who missed several games while healthy for personal reasons. Ovechkin started out the year poorly, with only 5 points in his first 8 games (Malkin had 13 points in his first 8 games). Ovechkin then got to rest for several games while healthy, and no surprise, puts up a hot streak upon returning of 17 points in his next 8 games. It is highly doubtful Ovechkin does this if he was not nicely rested at that time.
Also, with a less than 1 point difference on the year according to PPG, score effects come into play. With just 4 games left to play in the regular season, Malkin is comfortably ahead of Ovechkin by 4 points, and both players have nearly identical TOI numbers per game at thus stage. Ovechkin is chasing Malkin, and still has some hope of catching him, and receives the following icetimes in the final 4 games of the season: 25:21, 20:08, 23:15, 23:19. He puts up 6 points. Malkin meanwhile is on cruise control in the lead, and his icetimes were: 22:19, 18:20, 16:53, and 19:37, well below his final season average of 22:31. He only puts up 5 points in his last 4 games as a result.
In sum, Malkin was better in terms of offensive production than Ovechkin both at their respective peaks, and throughout their careers as a whole. He most likely outproduces Ovechkin in 2008-09 again if Ovechkin plays all 82 games for two reasons: first, Ovechkin does not get the benefit of resting during the season while taking games off when he is healthy (ala Mario in 1995-96), which would lower Ovechkin's PPG; and second, because instead of decreasing Malkin's icetime at season's end, Malkin would be getting his usual TOI or an elevated TOI (like Ovechkin received) if he was behind (or tied) in the Art Ross race, which would likely have been more than enough for a player like peak Malkin to overcome a 0.01 PPG difference.