There is one bit of Ovechkin's legacy that I've been wanting to throw in a mathematical thread somewhere but there hasn't been a good spot for it. It involves sample sizes and what we think we know about ranges of outcomes.
Years | Games | Goals | Assists | Points | Shots | Sh% | SPG |
05-06/17-18 | 1003 | 607 | 515 | 1122 | 4896 | 0.124 | 4.88 |
That's Ovechkin after 13 years and 1000 games in the NHL, averaging 4.88 shots per game, and shooting 12.4% for his career. We have 13 individual season-samples, and we have 2 peaks at 14.6/14.5%, the first his peak year in 07-08, the other in the shortened season of 12-13, and a valley at 8.7% in 10-11. One would think that this is a large enough sample size to have an accurate grasp of Ovechkin's potential range of outcomes in terms of shooting, but let's look at his next 5 years.
Years | Games | Goals | Assists | Points | Shots | Sh% | SPG |
18-19/22-23 | 344 | 215 | 148 | 363 | 1459 | 0.147 | 4.24 |
He loses about 15% of his shot volume, but it seems those were all the ones that weren't going in anyway, as his cumulative shooting percentage for these 344 games was higher than his peak shooting percentage seasons in his first 1000 games. He had 3 separate years over 15% (though none higher than 15.4%), all of them full seasons (however the 19-20 season of 68 games is theoretically partial).
One would think that having 1000 games and nearly 5000 shots on goal would be a big enough sample to have a settled range, but apparently not. I've always wanted to use this in response to someone putting too much emphasis on a small sample size, showing that even in the biggest of samples, it's all still randomness in the end.
Also, I did a breakdown by EV/PP, to see if the power play had any effect, and it really did not.
EV | Games | Goals | Assists | Points | Shots | Sh% | SPG |
07-08/12-13 | 438 | 179 | 160 | 339 | 1551 | 0.115 | 3.54 |
13-14/17-18 | 402 | 132 | 97 | 229 | 1210 | 0.109 | 3.01 |
18-19/22-23 | 344 | 144 | 101 | 245 | 961 | 0.150 | 2.79 |
| | | | | | | |
PP | Games | Goals | Assists | Points | Shots | Sh% | SPG |
07-08/12-13 | 438 | 90 | 103 | 193 | 662 | 0.136 | 1.51 |
13-14/17-18 | 402 | 104 | 53 | 157 | 632 | 0.165 | 1.57 |
18-19/22-23 | 344 | 70 | 47 | 117 | 494 | 0.142 | 1.44 |
It is remarkable that we have ~13600 minutes of Ovechkin at even strength, and all the evidence shows he's an 11% shooter at even strength, and then in his next ~5400 even strength minutes he shoots 15%.
[PS There's a 3 goal 2 assist 15 shot discrepancy between hockey-reference and NHL that I'm not going to chase down for the purposes of this post.]