You have entered an alternate universe.
One where hockey seasons have now finished, teams are celebrating their league championships, the Ottawa Senators just drafted their franchise players, and hockey media has turned towards free agency.
And most importantly, the AHL season wasn’t cancelled.
Welcome to Earth 3, a parallel universe to our current reality as well as Dom Luszczyszyn’s Earth 2, where he simulated the outcome of the 2020 Stanley Cup playoffs. Over the course of the next few weeks I’ll be simulating the 2020 Calder Cup playoffs, where at the end we will find out once and for all who will be crowned AHL champions.
I’ll be joined by Belleville Senators writer Spencer Blake for the journey, as the B-Sens were poised to make a deep run at the championship with their young core. Between Josh Norris receiving the award as the league’s Most Outstanding Rookie, Drake Batherson being named to the second All-Star team and Alex Formenton joining the All-Rookie team, their scary offensive unit makes them one of the favourites on Earth 3.
In this article I’ll go over the rules and methods for the simulation — you can already read the results of round one here.
No shenanigans here, unlike the NHL. Here’s the bracket!
The AHL playoff format takes the top four teams from each division and pits them against each other in 1st vs. 4th and 2nd vs 3rd seed matchups. Due to some teams having played more games than others, the seeding was decided by points percentage — fortunately it wouldn’t have been any different even if we went by straight points. It’s not a perfect format as we see the 50.8% Chicago Wolves make the postseason instead of six other teams with a better record, but that’s what the AHL uses.
The league is also unique in that the first round is a best of five series instead of the regular best of seven, which is the case for all subsequent rounds. Home ice advantage works the same way, with the home team hosting games 1-2, the away team for games 3-4, and if necessary it goes back to the home team for the deciding game five.
We’ll be checking in on the simulation once per week for the next four weeks, covering one round each time. By the end, one team will remain as the 2020 Calder Cup Champion!
Quantifying talent level in the AHL is no easy task. The lack of available data for simple things like shots and ice time already takes away most of our ability to make a useful model, but there are workarounds.
The simulation will be based on Win Shares, a stat created by Hockey-Reference.com to compare players across various eras using the data available from each period. Using data from Pick224.com I incorporated goals, assists, on-ice goals and estimated time on ice for skaters into the Win Shares formula. For goalies I used shots against, goals against and ice time from the AHL’s website. The Win Shares will be used as a per-game rate in the simulation to create a level playing field for players who were either injured or spent time in the NHL or ECHL.
There’s a bunch of drawbacks, most notably the inability to split a lot of the data for players who played on multiple teams, given the way Pick224 sets up their data. Nevertheless, it’s a decent estimate for which players provided value this past season, or at least good enough for me to feel comfortable using it in this wacky series.
Here are the top fifteen skaters this past season ranked by Win Shares:
The season’s points-per-game leader Reid Boucher takes first place, MVP winner Gerald Mayhew takes second, Rookie of the Year winner Josh Norris takes third, and a bunch of other players with great seasons fill out the rest. Passes the smell test to me.
And here are the top ten goalies:
Kaapo Kahkonen, winner of the most outstanding goaltender award, places fourth via Win Shares. But it’s also tough to argue having Anthony Stolarz at the top, who posted a 92.2% save percentage in 39 games for the San Diego Gulls.
I’ve also curated the lineups for every team based on each player’s estimated ice time received, plus a couple added opinions from various people who are far more knowledgable about specific teams. Players were excluded if they fall into any of these three categories:
- The player was ineligible to play in the AHL playoffs (i.e. they weren’t on the AHL roster on February 24th)
- The player appeared in less than ten AHL games this past season (to eliminate small sample sizes)
- They were combatting an injury when the season ended (some exceptions will be made for subsequent rounds)
And now for the fun part! For listeners of the Cost Per Pointcast this simulation may seem familiar, as we used a very similar tool to simulate a best-of-seven series after drafting the greatest Sens players of the salary cap era. It was a chaotic episode, and I’m sure we’ll see some of that transferred into this series.
Odds will be assigned to teams for each game based on the cumulative amount of Win Shares on the team’s active roster, then adjusted for home ice advantage. I’ll run the simulation once to generate a random number, determining the winner of that specific game. Repeat for subsequent games until we have a series winner, and eventually Calder Cup champion!
Each game’s simulation will include some extra information too:
- The score, including whether the game went to overtime (no pesky shootouts in the playoffs)
- The goal scorer(s), which will help us determine the most important players of each series
- Injuries, which is the most crucial aspect, as roster changes (decided by myself and Spencer) will be occurring in between games. Players can be injured for one game, two games or for the rest of the series. Rosters will be reset after every series. and keeping in line with playoff tradition, none of the injury specifics will be revealed.
And that’s everything you need to know! Say your prayers to the simulation gods and head over to the results of the first round which you can already find here. Best of luck to the competitors!