Using xG to Spot Hidden 1-1 Potential in Lopsided Fixtures

A big favourite at home against a struggling visitor looks, on paper, like the last place to find a score draw. Expected goals data tells a slightly more interesting story once you actually run the numbers rather than eyeballing the table gap. This piece walks through exactly how, with a transparent worked example rather than a vague claim.

What expected goals actually measures

Expected goals, usually written xG, estimates the number of goals a team would be expected to score on average given the quality and quantity of chances created in a match, based on historical conversion rates for similar chances. It is a measure of chance quality, not a guarantee of a specific scoreline, and that distinction is exactly why it is useful for probability work rather than simple prediction.

Turning xG into scoreline probabilities

A standard, simplified way to convert a team’s xG figure into scoreline probabilities is the Poisson distribution, which estimates the likelihood of a team scoring exactly k goals given an average rate (their xG) using the formula P(k) = e^(−λ) × λ^k / k!, where λ is the expected goals figure. Multiplying the two teams’ individual probabilities for a given goal count together gives an estimate for that specific scoreline, assuming independence between the two sides, which is a simplification but a widely used and reasonable one for illustrative purposes.

Two illustrative matchups

Consider two scenarios, both purely illustrative:

• Evenly matched sides: both teams with an xG of 1.3 for the match.
• Lopsided matchup: a stronger side with an xG of 1.9 against a weaker side with an xG of 0.6.

Running the Poisson calculation for both scenarios gives the following illustrative results:

The counterintuitive part

Look at the P(1-1) column. The evenly matched fixture is, unsurprisingly, more likely to finish 1-1 at 12.6% in this illustrative model. But the lopsided matchup still carries a 9.4% chance of the same result, not a dramatically smaller number, and notably higher than its own chance of finishing 0-0. A big favourite can create enough chances to be individually likely to score once, while a weaker side can still have enough of a single opportunity in an otherwise one-sided match to nick a goal of its own. The favourite scoring once rather than twice or more, combined with the underdog scoring exactly once, is not nearly as rare as the table gap between the two sides might suggest.

Why this happens, mechanically

A Poisson distribution with a moderate λ, say 1.9, still places a meaningful chunk of its probability on exactly 1 goal rather than 2 or more; favourites do not score as freely as their overall quality gap implies, because converting chances into goals always carries variance. Meanwhile, even a weak side with a low λ of 0.6 has a non-trivial chance of scoring exactly once rather than being shut out completely. Multiply those two individually plausible outcomes together and a 1-1 becomes a live possibility rather than a surprise.

A practical health check before trusting any xG-based reasoning

• Use a recent, match-relevant xG figure rather than a season-long average that may be skewed by one outlier performance.
• Check for injuries to key attacking or defensive players since the xG figures were compiled; the numbers reflect the past, not necessarily the team sheet on the day.
• Remember that the Poisson approach assumes independence between the two teams’ scoring, which is a simplification; game state, where one team starts controlling tempo after taking a lead, can shift actual behaviour away from the model.
• Treat every percentage here as a rough guide to likelihood, not a forecast; a 9.4% illustrative chance is still, by definition, an unlikely outcome in isolation.

Extending the model one step further

It is worth asking what happens at the extremes, where the quality gap is even larger than the lopsided example above. Pushing the favourite’s xG up to 2.4 and the underdog’s down to 0.4, still purely illustrative, the Poisson model gives a P(1-1) of roughly 5.8%, smaller than the 9.4% seen at 1.9 versus 0.6, but still not negligible, and still higher than many players would guess from the scoreline gap alone. The pattern holds in direction even as the specific numbers shrink: a 1-1 rarely disappears from the realistic range of outcomes just because one side is heavily favoured, it simply becomes gradually less likely alongside every other scoreline as the favourite’s dominance grows.

How to use this on coupon night without overreaching

This is not an argument that favourites should be avoided, or that lopsided fixtures are secretly draws in disguise. It is a reminder that “big favourite at home” and “score draw is basically impossible” are not the same statement, and that a genuine look at both teams’ expected goals, rather than just the table gap, can surface the occasional live 1-1 possibility that a glance at form alone would miss. Use it as one extra piece of context when weighing up a fixture, not as a standalone system.

It is also worth being candid about what this approach cannot do. It cannot tell you which specific lopsided fixture, out of a whole weekend’s worth of them, will actually produce the live 1-1 rather than the routine home win the table gap suggests. It can only tell you that, across a reasonable spread of similarly shaped matchups, a 1-1 is not nearly as rare an outcome as the headline quality gap might imply, which is a statement about tendencies, not about any single result you can rely on in advance.

All figures here are illustrative outputs of a simplified statistical model, not predictions for any real match, and no model removes the underlying uncertainty of football. Keep your coupon spend modest and decided in advance, treat any dividend as a bonus rather than an expectation, and play only if you are 18 or over, with pools enjoyed strictly as entertainment rather than a source of income.