Outline
- Intro
- Data
- Results and Discussion
- Code
- Connect
Intro
All the excitement surrounding the 2022 NHL draft called into question the risks and rewards of trading current NHL players for draft picks. For example, how often do draft picks actually develop into serviceable NHL players? In this article, we will investigate this as well as look at the associations between overall draft position and draft round with basic metrics like games played, seasons played, and points.
Data
5096 records from 2000–2021 were downloaded from hockey-reference.com. Any analyses involving points excluded goalies. Most analyses excluded the 2018–2021 drafts to allow for sufficient time for most draft picks to make the NHL (Figure 1). Only the first 7 rounds of drafts were kept for most analyses as the more recent drafts only have 7 rounds.
Results and Discussion
Most draft picks never play in the NHL
Below, Figure 1 shows that only 3% of draft picks from the 2021 draft played at least one NHL game in the same 2021–2022 season. This percentage jumped to 12%, then 27% for the 2020 and 2019 draft classes, one and two years removed since their respective drafts. Four years removed since the draft year, 43% of the picks from the 2017 class had played at least one game. These percentages hovered around this high 40s — low 50s for the classes that follow, indicating that it takes a few years to reach this plateau and that often the majority of draft picks never make the NHL.
Expectedly, this was skewed towards lower draft picks (Figure 2). 94% of first round draft picks (between 2000–2017) played at least one game, 79% of which played at least 82 games (season length). 68% of second round picks played at least one game, while only 41% of which played 82+ games. On the other extreme, only 25% of 7th round picks played one NHL game or more, 11% of which played a full season or more. The Pearson correlation coefficient between draft round vs 1+ games played was -0.95, indicating that the later the round a player is selected, the less likely the player is to ever play an NHL game.
This strong negative correlation remains when looking only at the top 100 draft selections (Figure 3). 100% of first overall picks play at least one NHL season, but this percentage quickly dropped off moving down the list.
Earlier draft picks tend to have higher points per game
While earlier draft picks tend to have higher points per game, there remains a high level of variability (Figure 4). Notable players include Nikita Kucherov (58th), Brad Marchand (71st), Jake Guentzel (77th), Brayden Point (79th), and Kirill Kaprizov (135th), who all have some of the highest scaled points per game values but were selected well outside of the first round. Still, at the very top sits Connor McDavid (1st overall, 2015), followed by Sidney Crosby (1st overall, 2005) and Evgeni Malkin (2nd overall, 2004).
The scaled points per game values were calculated by first calculating each skater’s point per game, then subtracting said values from mean of their respective draft classes. The purpose of this step was to reduce the effect of years in the NHL on points and create a class-by-class comparison. Of note, this analysis included both forwards and defensemen. Also, only points per game was analyzed, which by no means is the only metric to compare players.
In summary, there is substantial risk in trading current players for draft picks, especially as selections are lower down the draft board. Of course, there remains the enticing, massive plus side if a team is fortunate to draft a Kirill Kaprizov or Brad Marchand.
Code
Available here: https://github.com/Christianlee19/hockey-stats/tree/main/bin/article_scripts/07-2022_draft.
Connect
Feel free to connect with me at my new twitter account!
