Rank against career NHL games, a matured class
| Pool | Ranked | Matched | Never played | Games ρ | Reading |
|---|---|---|---|---|---|
| North American Skater | 247 | 120 | 54 | +0.61 | Strong agreement |
| International Skater | 142 | 52 | 29 | +0.34 | Moderate agreement |
| North American Goalie | 36 | 12 | 7 | +0.55 | Strong agreement |
| International Goalie | 12 | 4 | 2 | +0.95 | Strong agreement |
Spearman rank correlation, oriented so a positive number means the board agreed with what happened. Matched is smaller than ranked for two separate reasons: a ranked player who went undrafted has no outcome to compare against, and the league publishes no player id, so the join to a pick is made on name and birthdate and declines where that is ambiguous. 188 of 437 ranked players are in these figures. A matched player who never reached the NHL is counted as zero games, never dropped.
1963–2019, 57 classes with at least 7 seasons played out
Career NHL games a club's picks actually played, against what those same slots returned league-wide. Clubs are as they were on draft day— Atlanta's picks stay Atlanta's rather than becoming Winnipeg's, because rolling a record through a relocation is a judgment about who owns it rather than something in the data. Clubs with fewer than 20 counted picks are left out, since a leaderboard whose leader had four picks is not a finding.
1963–2019, 57 classes with at least 7 seasons played out
| Overall pick | Career games, average | Players |
|---|---|---|
| #1 | 742 | 285 |
| #33 | 229 | 459 |
| #65 | 150 | 459 |
| #97 | 90 | 450 |
| #129 | 79 | 425 |
| #161 | 67 | 405 |
| #193 | 43 | 387 |
| #225 | 54 | 234 |
| #257 | 37 | 96 |
| #289 | 93 | 40 |
The mean of the actual record, not a fitted curve — each slot pools players taken within 4 picks either side of it, across every class counted, so one slot is not left to ~57 samples. Players who never reached the NHL are in these averages as zeros. Rows sample the curve one pick per round; the endpoint carries every slot.