Methodology

For this model, I record the percentage of MoneyPuck's Deserve O'Meter for each game. These percentages are then summed over the season to estimate how many games each team deserved to win. I've seen other people using the Deserve O'Meter for the same purpose, but they usually just give 2 points to the team that won that matchup, whether it was a 55%/45% matchup of a 80%/20% matchup. I think my proportional approach is a better indicator of which team "deserve" to make the playoff

Limitations

A key number I tracked is to compare the number of games won by the Deserve O'Meter favorite with the number of games that favorite actually won. I found that the team that "deserve" to win, win on average 9,80% less than expected.

Methodology

For this model, I take the sum of last year's Game Score for each team. I then take out the GS of players who will not play for that team next season, and add the GS of the newly acquired players. For the past two seasons, I've correlated the total Game Score each team gets with their number of points for that season. That gives me a formula that can convert each team's new GS into projected PTS. I've also combined it with my in-season projection, using the MoneyPuck Deserve O'Meter.

Limitations

The model assumes every player will produce exactly like last season. Also, injured player and potentially top prospects waiting to start their NHL career are not counted.

Methodology

I simulated the 2026 NHL Draft 6,400 times, with each team having the ability to swap picks with other teams.

I found 32 draft rankings made by various scouts published online, then randomly assigned one ranking to each NHL team for a simulation. Once the simulation was completed, I reassigned the rankings to different teams and repeated the exercise.

My criteria for selecting rankings were:

• At least 64 players ranked
• No mock drafts
• Published on an established website
• Preference given to lists that ranked goaltenders (although I had to settle for two lists that did not)

For trades, I made a formula that takes into consideration:

• The Perri Pick Value, that you can find on Puckpedia
• Perceived value, which adjusts the base value according to the players remaining on a team's particular list. For example, if a team is trying to move up to pick #8, but the player they intend to select is ranked #5 on their list, they value the trade as if they were acquiring the #5 pick. At the same time, the team holding pick #8 still has their #3-ranked player available, they will perceive their pick as being worth the #3 selection, making them less likely to trade down.
• The number of picks a team owns. Teams with fewer than seven picks are more likely to trade down, while teams with more than seven picks are more likely to trade up.

The average simulation lasted 107 picks. The shortest simulation lasted 75 picks, while the longest lasted 164 picks.

Methodology

Soon

Methodology

Soon

Methodology

Soon