How to Read a Match Probability Bar

A match probability bar is a single horizontal bar split into three bands. It shows how the chance of a match result is divided between a home win, a draw and an away win, and nothing else. This page explains the device and how to read it without asking it to say more than it says.

What the bar is

Each bar above answers one question: of the three possible results — home win, draw, away win — how is the chance split? The answer is drawn as three bands laid end to end. Blue is the home win, grey is the draw and ochre is the away win. A band twice as wide as its neighbour carries twice the chance. The three bars at the top of this page are worked illustrations; each is labelled as an example and describes no real fixture.

Why the bands always sum to 100

Full time brings exactly one of three results: the home side wins, the teams draw, or the away side wins. The outcomes are mutually exclusive — no two of them can occur together — and exhaustive — there is no fourth result to account for. All of the chance must therefore sit inside the bar, and none of it can leak out. If the home band holds 61%, the draw and away bands must hold the remaining 39% between them. There is no remainder and nothing outside.

One boundary is worth drawing now. The bar speaks about the result at full time in regulation. A cup tie decided in extra time or on penalties is a different question, and this bar does not answer it. The bar also says nothing about the score or about events inside the match. It carries one share per outcome, and the share is the chance of the result.

A favourite is not a confidence

The widest band is the most likely single outcome. That is all it is. Look again at the first bar above. The home band is 61%, the widest on this page, and the other two bands together still hold 39% of the chance — close to two chances in five. Imagine a long run of matches, each with the home band drawn at exactly 61%. In about four of every ten of those matches the home side would not win. Those four are not the bar failing. They are the 39% doing what it was drawn to do, and a distribution that never produced them would be a distribution drawn wrongly.

That is the discipline that separates a probability from a promise. One match settles nothing about whether a bar was sensible. A distribution is judged over many matches, by how its shares compare with what happened across the whole run — not by whether the widest band was followed by the result on a single occasion. The mistake to avoid is reading a 61% band as a statement that the home side will win. It says that this outcome holds more chance than the other two, and that nearly two fifths of the chance says otherwise.

The second bar above is the counterweight. In an evenly matched fixture the home band is 38%, and 62% of the chance sits outside it. The home band is still the widest, so the same word — favourite — is available for both bars. The two are not making the same statement. The first says one outcome clearly outweighs the other two. The second says the three outcomes sit close enough together that none of them dominates.

What moves the bands

Compare the first and third bars. They show the same favourite one week apart. In the first, the home band is 61%. In the third it is 52%, and the nine points of chance it shed have moved into the draw and away bands. Nothing about the fixture itself changed between the two drawings. What changed is information: a squad change, a suspension, a congested schedule, a different venue. Any of these can move the bands, because a band is a reading of what was known at the moment it was drawn.

A moving bar is often mistaken for an admission of error. It is better understood as the model taking in new information. The first bar reflected what was known in its week; a week later there was more to know, and the same fixture read differently. The earlier bar was not wrong to sit where it sat. A distribution is a snapshot, and snapshots are meant to move when the scene moves. The worrying behaviour would be the opposite: bands that never shifted while the matches around them changed week after week.

How to read a bar honestly

Three habits cover most of it.

Read the whole bar rather than the widest band. The widest band is one number; the distribution is all three, and the gaps between the bands carry as much information as the leader.

Notice how much chance sits outside the leading band. A leader with 39% against it, as in the first bar, is a different statement from a leader with 62% against it, as in the second. The chance outside the leader is the honest measure of how far the bar leans.

Treat two bars with the same leader but different spreads as different statements. The first and second bars share a leader and little else. Reading them as one signal throws away most of what they say.

What the bar is not

The bar is not advice. It is not a selection, and it is not a prediction that one outcome will occur. It is a description of a distribution, published so that the whole shape of the chance can be seen rather than a single band of it.

The record so far

LeagueQuant has published no forecasts and has no accuracy record. Nothing on this page rests on past performance, because there is none to cite. When live forecasts begin, each one will be published with a hash recorded before kick-off and will be evaluated automatically once the match has finished, with misses included in the record alongside everything else. Until then, every number in the bars above is illustrative and carries that label.