Thursday, December 7, 2017

NHL Draft Pick Value

What is the expected value of each NHL draft pick? We can measure value by the amount of salary paid to players under an NHL contract (goalies not included), looking at all draft picks from 2004 to 2016 (ergo only players drafted into the salary cap era with streamlined entry level contracts). 

There are some flaws to using only Games Played to measure value, whereas salary does a better job reflecting the player's value to their franchise (with some exceptions). Most entry level contracts are the same (roughly 80% of players on ELC will have an cap hit between $5K - $25K per NHL game). Where salary does a better job of estimating player value is on contracts after entry level.

Salary in this case does not refer to how much money the player put in his bank account, it is what they count against the salary cap (as per, plus a max of $50,000 for those playing AHL games while on a 2-way NHL contract. A lesser sum is awarded for ECHL games, provided the player is on an NHL contract. For example, those who played 1-5 NHL games on an entry level contract and a full AHL schedule average $60,000 in salary. Those playing in the minors on a minor league contract count as $0. 

Age refers to the age of each player on September 15th, such that those born between Sept 15 and Dec 31 count as a year younger. This is necessary because those players are forced to be drafted a year later under current NHL rules. Treating Sept 15th like that's New Year's Eve makes the data analysis more streamlined. An 18-year-old is a player in his first season after his first year of draft eligibility.

There are five different graphs below all examining a different means of measuring value using contract data.

Sum of All Salary per round

The total amount of salary earned by each round of the NHL draft from age 18-27. These are the raw totals only for players who are under NHL contract, or have their rights owned by an NHL team.

1st round picks earned a total of $3.26 billion dollars, which accounts for 61% of all the salary earned by NHL draft picks. The top 10 picks earned $2 billion of those dollars, which reinforces how top heavy the draft pool tends to be. There is an elite tier at the top of round 1, a good tier that extends into round 2 (sometimes into round 3 depending on the depth of the draft class), and everything else is like buying scratch and win lottery tickets. Expected salary really starts to flatten out after the 2nd round. There is not much difference between the 3rd and 4th rounds.

This chart is does an excellent job of showing just how valuable a 1st round pick can be, but doesn't show us rate of decline from one pick to the next.

Expected Career Earnings by pick

This is calculated by taking the expected salary at each age (up to 27) for each pick and then adding them all together (the sum of the average values). The higher picks are more likely to play NHL games as teenagers and thus will earn extra career earnings while other picks have yet to turn pro. If we looked specifically at how good a player will be by a certain age, the difference would be less pronounced than with Total Expected salary.

The 1st overall pick is expected to make almost $60M in future salary (from age 18 to 27) and drops sharply down to around $15M by the 8th pick. The players who go directly to the NHL as 18-year-olds will also get to their second contracts at an earlier age. Using cumulative salary to measure pick value does skew towards the elite tier who plays more games. Over 95% of all picks after the 1st round will max-out their entry level slide.

There is very little difference from the 15th pick to the 30th pick. The 2nd round has a high variance in future salary, then early in the 3rd round we get into the "lottery ticket" zone. The probability of drafting an impact player beyond the 3rd round would be between 5% and 10% depending on the parameters.

Expected NHL Salary at Age 23

Instead of taking what each draft pick earns at each age and then adding them together; this one specifically looks at what a pick is expected to earn at the age when the majority of entry level contracts have expired. At age 22, 87% of draft picks on NHL contracts are still entry level. At age 23, only 37% of draft picks on NHL contracts are entry level.

This chart looks very similar to the one above, except that the values don't drop as sharply from the 7th pick to the end of the 1st round. It does less to punish players for entry level slide, providing a smaller rate of decline in expected value from one pick to the next in the middle of the 1st round. The Expected salary of the 1st overall pick is over $6M, but drops down below $2M by the 12th pick. The decline from the 10th pick to the 50th pick is less severe than with the career earnings graph above.

Probability of NHL Draft Pick Signing Entry Level Contract

What is the probability that any given draft pick will sign an entry level contract with the team that drafted them by the deadline to retain their rights?  Generally, teams have two seasons to sign junior players, four seasons to sign European players, and whenever they graduate to sign college players. This sample does not count players still in the NCAA at age 23 (which is 3% of the draft population).

For this graph the picks are bundled into groups of 10. There is a vertical line at the end of every round, and the dots on the grid-lines represents the last 10 picks of that round.

The probability of a 1st round draft pick signing an Entry Level Contract with their draft team by age 23 (or by the deadline to sign) is 98%. Getting drafted in the 1st round all but guarantees getting offered a basic ELC, 2nd round picks 88%, 3rd round picks 81%, 4th round 65%, 5th round 55%, 6th round 49%, 7th round 38%.

There are some anomalous results. For example, picks 151-160 had a higher % sign entry level contracts than picks 141-150.  Why would the first 10 picks of the 6th round have a higher sign rate than the last 10 picks of the 5th round? My best explanation is randomness. There is a less predictable pattern the later you get in the draft. We get more waves out by the tail. There may be a better explanation hidden in the details, based on the type of player that tends to be taken in those rounds (more European and tier II players).

There was a PhD student who did his thesis project on NHL Draft Pick value, and he also observed the phenomenon of some later picks performing better than some earlier picks. He could not explain it. You can get clusters of good players coming from specific pick ranges later in the draft that defies intuition.

Pick Ownership Percentage (by age)

What percentage of all NHL Draft picks are owned by an NHL team at any given age? These are players either with a contract, or a team simply retains their rights for various reasons (like players heading to Europe before their contract expires).

The steepest rate of ownership decline comes from age 22 to 23, when the majority of entry level contracts expire. There are many players who get an ELC, but never get a 2nd NHL deal.

No comments:

Post a Comment