Model Information

How ShiftEdge works.

This is the full reference for ShiftEdge: how the NHL win-probability model works, what the numbers mean, how predictions change, how the model was historically tested, and how live 2026–27 results are tracked.

View Game Predictions View Model History View Betting History

What the model is designed to do

The model is not built simply to pick the team most likely to win every game. Its primary purpose is to estimate win probabilities and compare those probabilities with available market prices. That difference is used to identify potential edge and expected value (EV). A team can be less likely to win outright and still represent better value if the market price is favorable enough.

Research & gambling disclaimer

This site and its model are provided for research, analytical, educational, and informational purposes only. Model probabilities, edges, expected values, picks, and historical results do not guarantee future outcomes and should not be treated as financial or gambling advice. Gambling involves risk and can result in financial loss. Any decision to place a wager is made solely by the user, who is responsible for evaluating the information, complying with applicable laws and age requirements, and deciding whether gambling is appropriate for them.

How it works

The live prediction is built from the expected players and goalie for each team plus schedule context, then translated into a win probability.

1. Historical dataPlayer game history supplies the model's rolling inputs.

2. Current lineupThe most up-to-date projected lineup is used before puck drop.

3. Team contextRoster continuity, goalie history, rest and back-to-back context are calculated.

4. ModelThe 14-feature model converts those inputs into a win probability.

5. Market comparisonModel probability is compared with available market prices to calculate edge and EV.

Model information

A compact description of what is currently running behind the site.

Current model

nhl-14-feature-v1-2026-27. The model uses 14 engineered features covering expected roster continuity, player history, projected goalie context, rest and back-to-back status.

Lineup policy

Future games use the latest verified lineup information available for that team. The model is recalculated as lineup information changes, then the prediction used for official history is locked at puck drop.

Player and goalie impact

Impact displays estimate how the players currently expected in the lineup affect the matchup-level model output. Projected goalie impact is shown separately underneath the top skaters.

Live refreshes

The prediction service performs a full scheduled refresh every 30 minutes. Market prices can refresh independently, so odds and model timestamps should be read separately.

How to read a prediction

The most important comparison is not simply which team has the higher win probability — it is how the model's probability compares with the market price.

Example: Team AIllustrative example only

58%Model Win Probability

50%Market Probability

+8.0%Edge

+16%Expected Value

How to interpret it: the model estimates Team A wins about 58% of the time, while the available price values the team closer to 50%. That gap creates an 8-point model edge. If the offered odds produce positive EV using the model's 58% estimate, the price may represent value. This does not mean Team A is guaranteed to win the individual game.

What can change a prediction?

Pregame probabilities can move as the information available to the model changes. The site continues updating future games until the official prediction is locked at puck drop.

Lineup changes

Players entering or leaving the expected lineup can alter roster continuity and the player inputs used for the matchup.

Goalie confirmation

A change in the projected or confirmed starter can affect the goalie-related inputs and move the team's projected win probability.

New player data

Completed games add new player history to the system, which can change the rolling values used in future predictions.

Rest & schedule context

Days of rest, back-to-back status and the timing of the matchup are incorporated into the model and can differ from game to game.

Historical testing

These are the stored out-of-sample holdout metrics for the current base model. Lower Brier Score and Log Loss are better; higher AUC is better.

58.73%Holdout Accuracy

0.636Holdout ROC-AUC

0.657Holdout Log Loss

0.233Holdout Brier Score

2025–26Latest completed holdout season

Historical testing measures how the model performed on data kept outside the fitting process. The 2026–27 Model History is separate prospective validation: predictions are locked at puck drop and then graded against the actual game result.

History & transparency

Model performance and betting performance are intentionally tracked separately.

Model History

Every completed 2026–27 regular-season game with the final locked model pick and whether that pick was correct. This measures the predictive model itself.

Open Model History

Betting History

Only the wagers that qualify for the site's official 2026–27 betting record. This measures the results of the betting rules, not every model prediction.

Open Betting History

Definitions

Reference definitions for the main terms used throughout Predictions, Matchup Details, Top Picks, History and model testing.

Win Probability

The model's estimated probability that a team wins the game.

Market Probability

The probability implied by the listed market price. Sportsbook probabilities are normalized to remove bookmaker margin where applicable.

Edge

Model win probability minus market probability, shown in percentage points. Example: 60% model vs 52% market = +8.0-point edge.

Expected Value (EV)

The estimated average return from a 1-unit wager using the model probability and listed odds. Positive EV means the model sees theoretical value at that price.

ROC-AUC

Measures how well the model separates winners from losers across every possible probability cutoff. A score of 0.50 is roughly random; higher values indicate better ranking ability, with 1.00 being perfect separation.

Brier Score

Measures the squared error of the model's predicted probabilities. It rewards probabilities that are close to the actual outcomes and penalizes probabilities that are too high or too low. Lower is better, with 0 being perfect.

Log Loss

Measures the accuracy of the model's predicted probabilities while penalizing confident incorrect predictions especially heavily. Lower is better, with 0 representing perfect predictions.

Player Impact

The model-based contribution associated with an expected player relative to the matchup context used in the current prediction.

Goalie Impact

The estimated effect of the projected starter on the current matchup prediction.

Top Pick

A game-side opportunity that meets the site's current Top Picks EV threshold.

Locked Prediction

The final pregame model prediction preserved at puck drop for official model-history grading.