Free tool

    xG Calculator.

    An expected goals calculator: enter the xG for each side and a Poisson model turns those two numbers into match probabilities. You get the three results, the goals lines, both teams to score, and the most likely scorelines. This is the same model family our engine fits for every match it rates.

    46.4%

    Home win

    25.8%

    Draw

    27.8%

    Away win

    73.3%

    Over 1.5

    48.2%

    Over 2.5

    26.4%

    Over 3.5

    51.8%

    BTTS Yes

    Most likely scorelines

    1-1 12.3%1-0 11.1%2-1 9.2%2-0 8.4%0-1 8.2%

    A worked example

    Say the home side is worth 1.8 expected goals and the away side 0.9. Poisson turns 1.8 into the chance of that team scoring exactly 0, 1, 2, 3 goals and so on, and does the same for 0.9. Multiplying the two together gives every scoreline a probability, and that grid answers every market at once.

    On those numbers the home side wins 58.6% of the time, the draw lands 22.9%, and the away side wins 18.5%. The single most likely scoreline is 1-0, but notice it carries only 12.1% on its own, which is exactly why a correct-score bet is a long shot even when the model is confident about the match. Over 1.5 goals clears 75.1% of the time, Over 2.5 is almost a coin flip at 50.6%, and both teams score in 49.5%.

    The lesson worth taking from that: the most likely single scoreline and the most likely result are different questions, and a low-scoring favourite can be a strong home-win call and a weak Over 2.5 call at the same time.

    How it works

    Expected goals (xG) measures the quality of the chances a team creates. A Poisson distribution turns an xG number into the probability of scoring exactly 0, 1, 2 or more goals. Multiply the home and away distributions together and you get a probability for every possible scoreline, and from that grid every market follows: add up the cells where the home side scores more and you have the home win chance, add the cells with three or more total goals and you have Over 2.5.

    Our prediction engine runs the same idea with a deeper input set: attack and defence ratings fitted per team and adjusted for opponent strength, refreshed every day of the season. How the engine works, or see it applied to today's matches.

    For interest and analysis only. Past probabilities do not guarantee outcomes. 18+, bet responsibly.

    Frequently asked

    What is an xG calculator?
    It converts expected goals into probabilities. Expected goals (xG) measures the quality of the chances a team creates, so a side worth 1.8 xG created chances that would typically yield 1.8 goals. A Poisson model turns that single number into the chance of scoring exactly 0, 1, 2 or more goals, and combining both sides gives a probability for every scoreline.
    How do you calculate match probability from expected goals?
    Take each side's xG and apply the Poisson formula to get its distribution of goals scored. Multiply the two distributions to get a grid covering every scoreline. Every market is then a sum of cells: add the cells where the home team scores more for the home win, the cells totalling three or more goals for Over 2.5, and the cells where both sides score at least one for BTTS.
    Is Poisson accurate for football predictions?
    It is a good baseline and a poor final answer. Poisson assumes the two sides score independently, which real football violates: goals change how teams play, and a side chasing a game attacks differently. It also slightly understates draws. Serious models start from Poisson and correct it, which is what our engine does by fitting attack and defence ratings per team and adjusting for the opponent faced.
    Where do the xG numbers come from?
    This calculator does not fetch them; you supply both figures, which is what makes it useful for testing your own numbers. Most published xG comes from shot-level models that score every chance by its position, type and build-up. Our own engine does not take xG as an input at face value either: it fits attack and defence ratings from the last ten matches, weighted by the quality of the opponent.
    Why is the most likely scoreline still unlikely?
    Because the probability is spread across dozens of possible scorelines. Even a clear favourite rarely gives any single scoreline more than about 15%, which is why correct-score prices are long and why we publish the scoreline as the model's working rather than as a tip.