Mike and I went over the secondary/goal scorer thing a couple years ago and the preliminary conclusion is that its about 5-10%, across everybody, and subject to a ton of variance. It obviously affects someone who scores 40 goals more than someone who scores 20 goals, but the percentage of goals it affects is in the same ballpark.
[I went searching for the post and found it here:
Top-200 Hockey Players of All-Time - Preliminary Discussion Thread. I thought there was another follow-up from Mike where he broke down his video assessment of some goals but couldn't find it.]
As someone who in the end is mostly 'Points Uber Alles' in regards to forward impacts, I still understand that there's a shit ton of randomness involving points scored. In the end, it all sort of washes out, the positive and negative variance, and true talent mostly shines through over a larger sample.
There is however one points-related Ovechkin stat that I do want to point out that does affect him more than others. Nobody would disagree that Ovechkin is one of the most impactful players on the Power Play in NHL history, and yet when you look at his IPP (individual points percentage) compared to a whole host of other players, his total is substantially lower than just about everybody else. Thus, even though he's been on for almost 175 more power play goals than Crosby, he only has around 40 more power play points [in the 07-08 to 23-24 period]. If you look at the averages of a whole bunch of players, the superstars are around 70%, the stars are around 65%, and Ovechkin's at 60%. The forwards around him (or lower) include Pavelski, Marchand, Benn, Toews, E. Staal, Perry, Duchene, and Voracek, mostly secondary stars. [Even the players known more for their one-timers/shooting are still much higher - Stamkos 68.6, Pastrnak 69.9, Matthews 66.1. I went through 57 forwards and 28 defensemen, the highest scorers since 07-08 and summed their PPP and oiGF according to hockey-reference's extra stats.]
If you assume Ovechkin's power play impact is equivalent to the superstar players that are his normal peers, that would mean boosting his IPP to that 70% range, which would lead to an extra 84 points, or about 5 points per season (84/17, since we're missing the data from 05-06 and 06-07).
But again, I refer you back to my initial statement that there's a ton of randomness involved in points, and positive and negative variance wash out, and true talent shines through in a larger sample. We certainly have a large enough sample to state that Ovechkin's true talent on the power play is a 60% IPP. Whether that's an accurate measure of his power play talent is more subjective.
