NHL Draft Pick Probabilities for Fantasy Hockey

Jokke Nevalainen



This is the fourth article where I analyze the results of NHL Drafts between the years 2000 and 2009. This time, I’m looking at probabilities for fantasy hockey purposes. That means, skaters who can put up points and goalies who play a lot of games with good save percentage.


The first three parts can be found from here:


Quality of players


To quantify the quality of the player for fantasy hockey purposes, I split all drafted players into six categories based on their career stats. In the table below, you’ll find the minimum requirements to reach the player quality category. I have also assigned a value to each category – more on that later.



The idea behind the categories is that elite players are rare and difficult to find. They are the best of the best, franchise cornerstones. Great players are also valuable pieces but in shallow leagues, these ones are not that difficult to find. In shallow leagues, you don’t need to look any further than these two categories.


The good players are pretty easily replaceable even in leagues that are a bit deeper. For decent players to be of interest, you’d have to play in a very deep league. And depth players are only for the deepest of the deep leagues.


After setting the player quality standards, I checked to see how many players qualify for each category. I broke down the results based on player’s position.



Remember that this is a 10-year sample, so on average, one draft would produce three elite players, nine great players, eight good players, and six decent players. That means a total of 12 players per year for shallow leagues, a total of 20 players for deep leagues, and a total of 26 players for super-deep leagues. I’m pretty happy with those numbers, so I think the player quality categories were set pretty accurately.


Realized value


Next, let’s break down the results by draft round. For this, I’ll use the “Value” mentioned in the first table. The value of busts is zero, so they will be ignored here.