Daily NHL Game Winner and Player Scoring Predictions

Inside The Rink | DataDrivenHockey Daily NHL Game Predictions
DataDrivenHockey | Inside The Rink

Last Updated: May 10th, 8:27 AM

Welcome to Inside the Rink’s daily NHL probabilities! This page is updated daily with NHL win probabilities for each of the day’s games, as well as player goal-scoring, point-producing, and “Top Threat” probabilities. These probabilities are calculated via the datadrivenhockey game simulation model, GPM1. Information about the model and links to get access to even more model outputs are provided at the end of this article.

Daily Predictions and Probabilities

Win probabilities (and the corresponding implied odds), most likely final scores, and the “Top Threats” for each of today’s games are shown in the image below!

The below image shows the top 10 most likely goal scorers and point producers for tonight’s games. As an added bonus, it also has the best Tim Hortons Hockey Challenge picks (it’s a Canada thing, sorry, Americans/Europeans/Others).

Interested in access to more data?

Even more NHL probabilities are available on the datadrivenhockey Patreon! In addition to the probabilities in this article, you can find:

  • Every players goal, assist, point, and 2+ point probabilities (matchup specific!)
  • Game cards with over/under probabilities and probabilities of overtime for every NHL game
  • Projection cards that show every NHL players: most likely end-of-season goals, assists, and points; the probability that they will hit important goal and point milestones, their probability of producing vs. an average NHL team and the probability they will win the Art Ross or Rocket Richard trophies
  • NHL Fantasy content (in time for next season)

The image below is an example of a game card. This is from game 3 of the Leafs-Lighting first round series.

How does the model work?

For the nerds, here is some information on how the model makes its predictions.

Lineup Predicting: If it is available, the model scrapes lineup and starting goalie information. If not, the most recent saved lineup is used instead (the model saves the most recent lineups, so this lineup is usually quite accurate).

Player Scoring Probabilities per 30 seconds calculation: Each player’s individual goal-scoring, primary, and secondary assist rates are calculated using the @datadrivenhockey point projection model. The point projection model uses four years of data, age curves, potential estimation, and the DDH expected points/goals/assists model to calculate players’ projected scoring rates. Last season the point projection model had an average league-wide error of +/- 9.75 pts for forwards and +/- 6.6 pts for defensemen. The rates from this model are converted into production probabilities per 30 seconds of ice time for both even strength and PP situations.

Correcting for Defensive Skill of Opponents: The expected production rates of each player is adjusted using factors based on their opponent’s skill. These factors are calculated from defensive and goaltending metrics. The skill factors are: goal suppression ability vs. average team at 5v5 and goal suppression ability vs. average team on the penalty kill.

Other Corrections: Rest (time between games) and Home/Away is also considered when adjusting player production rates.

Game Simulation: Each game is simulated 5000+ times. Each game is split into 30 second intervals, and each player is assigned to play in a percentage of the intervals based on their average 5v5 and PP ice time this season. Powerplay opportunities for each team are sampled from a normal distribution, so each iteration of a simulated game has a different number of powerplay intervals available. For each 30 second interval for each player, a random number generator is used to evaluate if a player records a goal, primary assist, or secondary assist (if the random number is less than the probability of the event, we record that as an occurrence of the event). The total number of goals and primary assists a team records are combined with a weighted average to determine the total goals for each simulated game. The probability of winning is simply calculated by comparing the number of times each team “wins” one of the 5000+ simulated games.

Stanley Cup Aspirations Cue’ the Duck Boats Pod

The boys are back after a long break but have plenty to talk about as the playoffs are in full swing. We go series by series, looking at how teams have fared so far and who will come out on top. Thanks for listening! Please rate and review our show on your favorite listening platform. Check out our partner's website at www.insidetherink.com for all your latest hockey news.
  1. Stanley Cup Aspirations
  2. The Final Countdown
  3. Here Come the Playoffs
  4. Home Stretch
  5. Kevy Cooks

Leave a Reply

Your email address will not be published. Required fields are marked *

NHL Playoffs: Just Not Enough; Isles Comeback Bid Falls Short In Game 3 | 4/25/2024

Toward the end of “A Few Good Men,” Col. Nathan Jessup bellowed out, “You want answers,” and followed it up by answering, “You can’t handle the truth!” The truth tonight at UBS Arena, despite perhaps being the Islanders best effort out of the first three games, they just aren’t good enough, as the New York Islanders were once […]

Read More
Game 3 Preview: Nashville Predators vs. Vancouver Canucks 4/26/2024

Game 3 Preview: Nashville Predators vs. Vancouver Canucks 4/26/2024

The series between the Nashville Predators and Vancouver Canucks shifts to Nashville with the Preds looking to take a 2-1 series lead tonight.

Read More
Photo by Mike Carlson/Getty Images

NHL Playoffs: Tampa Bay Lightning vs. Florida Panthers Game 3 Recap | 04/25/2024

Heading home with an 0-2 deficit, the Tampa Bay Lightning knew they had to win to make this a series. Unfortunately, another close loss puts a season that at times looked promising on life support. While this team lacks experience defensively and stumbled through much of the regular season, a March to remember showed they […]

Read More