Here are some of the common betting terms used in hockey betting:
This one is simple - who will win the game? Moneyline odds may vary greatly, depending the disparity between the teams playing. Identifying and betting the underdog (the team with + odds) with the best chance of winning could payoff nicely.
Commonly known in other sports as the spread. The puckline in an NHL game is usually set at -1.5 goals for the favorite and +1.5 goals for the underdog. Betting the puckline can be difficult to consistently return on investment given the random nature of the sport, but effective if bet correctly. You'll find pulled goalie situations will become your best friend or enemy in these situations.
Also known as game total or over/under. This bet is the total number of goals scored in the game (by both teams combined). A standard total goals bet is set at either 5.5, 6, or 6.5 goals in the NHL depending on the sportsbook's projections for both teams. You can then bet either over or under, with these lines generally hovering close to 50-50 payouts.
A unit is simply a term for the amount of your wager. The term "unit" is used since everyone's standard wager may vary, depending on comfort-level, aggressiveness, and disposable income. Commonly a unit amount is suggested to be between 1%-3% of your sportsbetting bankroll, so someone with $1,000 to play with would have a standard unit amount anywhere between $10 and $30.
Return on Investment (ROI)
ROI stands for return on investment. You may be familiar with this term from investment banking, and its significance in sportsbetting is similar. ROI is the amount of dollars you're earning per dollar risked. While volume betting favorites can earn you money, it may not be worth the amount you have to risk. The key is to identify the best investment opportunities based on likeliness of a wager cashing and potential profit. That is where we come in.
Our advanced hockey and statistic gab can be unfamiliar, so here's some clarity:
This is not to be confused with "expected goals" (see below). We project goals for each team in each game through complex linear regression models based on trends in the last 10 years of NHL player data. Each individual is attributed their own isolated impact based on our models in each of the key phases of the game (even-strength, powerplay, and shorthanded). A cumulative projected goals number is then determined based on projected ice-time for each player in each phase of the game.
Expected Goals (xG)
This is a new-age statistic in hockey which in the public realm is basically a measure of shot quality. Given the shot location and shot type, we can determine the history of a similar shot resulting in a goal. So essentially every shot taken in the NHL we can assign a probability of that shot going in the net, also known as its expected goal number.
Goals Saved Above Expected (GSAx)
Piggybacking off of expected goals, goals saved above expected is the measure of how well a goaltender is performing based on the expected goals (or shot quality) faced. If a goaltender has a positive GSAx, then that goalie is saving above expectation and performing admirably. On the opposite on of the spectrum, if a goaltender has a negative GSAx then that goalie is not performing up to par. This is regarded as a more accurate representation of goalie performance than save percentage since not all shots are equally as difficult to save.
GSAx is usually measured as a cumulative statistic, which tilt both the positive and negative ends of the spectrum heavily towards goalies who play the most minutes. In order to even the playing field for comparison, we display GSAx/60, which is a goaltender's goals saved per expected per game (60 minutes).
Suggested Bet %, Kelly Bet %, and Lineup Data Confidence
You'll find these terms on our PuckLuck Picks page.
The Kelly bet, or Kelly criterion, is a long-trusted way of optimizing your sportsbetting bankroll by allocating dollars appropriately based on win probability and potential return on investment.
Lineup data confidence is a measure of our confidence level in the data used in our model to comprise the game's projections. The more each player projected to be involved in the game has played in the last three seasons, the higher the confidence level.
Suggested bet % is our version of "Fractional Kelly", which is a risk-adverse way of accounting for the margin of error in our data. We are able to apply a confidence level percentage to the initial Kelly bet percentage to lessen the risk on games consisting of more unproven players. This helps avoid putting trust in unreliable data and provides what we consider to be the actual most optimal allocation of your units.