Glossary

Shots on goal the model projects for this player tonight

The model's projected shots on goal for the player tonight, the average across every outcome it simulates.

Compare it to the SOG Line: a projection well above the line points toward the over, well below toward the under. The probabilities give the precise read.

Goals the model expects this player to score tonight

The model's expected goals for the player tonight. A value of 0.50 means the model expects half a goal on average, not that he scores half the time. Model ATGS is the chance he scores at least one.

The sportsbook's shots-on-goal line

The shots-on-goal total the sportsbook is offering. Over 3.5 wins on 4 or more shots, under 3.5 wins on 3 or fewer.

Sportsbook odds for the over on the SOG line

American odds for the over. -150 means risk 1.50 units to win 1; +125 means risk 1 unit to win 1.25.

Sportsbook odds for the under on the SOG line

American odds for the under, read the same way as SOG Over.

Sportsbook odds for this player to score a goal

American odds for the anytime goal scorer bet: the player scores at least one goal in the game.

Model's chance of at least N shots on goal

The model's probability the player records at least that many shots on goal. Use these to price alt lines: Model SOG 4+ is the chance of clearing an over 3.5.

Model's chance this player scores at least one goal

The model's probability the player scores at least one goal. Compare it to the chance implied by Book ATGS; ROI ATGS does that math for you.

Model's chance the player goes over the SOG line, with its fair odds

The model's probability the player goes over the SOG line, followed by the fair odds for that probability. Fair odds are the price with no sportsbook margin: if the book pays better than fair, the model sees value.

Model's chance the player stays under the SOG line, with its fair odds

The model's probability the player stays under the SOG line, followed by its fair odds. Over and under add to 100%.

Expected return per unit risked on the over at the listed odds

Expected profit per 1 unit risked on the over, using the model's probability and the SOG Over odds. +10% means the model expects 0.10 units of profit per unit risked over the long run. Negative means the price is worse than the model's fair odds.

Expected return per unit risked on the under at the listed odds

Same as ROI SOG Over, for the under.

Expected return per unit risked on the anytime goal bet

Same as ROI SOG Over, for the anytime goal scorer bet at the Book ATGS odds. Goal props are long shots with high variance, so expect big swings even when ROI is positive.

Model probability minus the probability the odds imply

How much more likely the model thinks a bet is to win than the sportsbook's price implies. -150 implies 60%; if the model says 66%, that's a 6-point edge. ROI turns that edge into expected profit.

Expected ROI scaled down for long shots: P^0.9 x ROI

Expected ROI multiplied by the model probability raised to the 0.9 power. Raw ROI treats a +300 long shot and a -150 favorite the same if their ROI matches; Adj xROI trims the long shot because its result swings much more from bet to bet.

The Discord signals post plays at 10% Adj xROI or higher.

The no-margin price for a probability

The odds that would make a bet break even at the model's probability. A 66.4% chance is fair at -198. Any price better than fair, such as -154, is positive expected value by the model's numbers.

The full range of outcomes behind each projection

The probability of every shot total (0 through 7+) and every goal total (0, 1, 2+) for each player. A projection is one number; the distribution shows how likely each outcome is, which is what lines and alt lines are priced against.

How bet size is measured

Results are tracked in units instead of dollars. On minus odds a play risks enough to win 1 unit; on plus odds it risks 1 unit flat.