Football Predictions Today, Backed by Statistics

What football predictions are on this page

The football predictions on LMEX are produced by a fixed rule that works on numbers. The rule looks at the last five results of each team, how many goals each team scores and concedes per match, how the two sides did when they met before, and the Asian handicap and over/under prices for the match. It then asks one question: for the line the market has set, which side is easier to back than the price suggests? When the answer is clear enough, we publish a prediction. When it is not, we say no prediction and give the reason.

A prediction here is not a tip from an insider and not a result we claim to know. It is a statistics-based opinion, labelled that way every time, and it can be wrong. Football has few goals, so the result of one match depends heavily on luck, and even a good method loses often. The point of this page is to be open about the method and to publish a record that counts every win and every loss.

How a prediction is made

The short version is six steps. First, the rule estimates the goals each team will score from its recent goal averages, with a small adjustment for home advantage. Second, it turns those averages into a probability for every possible score. Third, it adds up the scores that make each side win the market line. Fourth, it removes the bookmaker's margin from the market price to find the probability the market itself implies. Fifth, it pulls its own figure halfway towards the market, because the market knows far more than ten matches of form. Sixth, it takes the side with the biggest gap between its figure and the market's, and publishes a prediction only if that gap is at least 6 percentage points.

The page how LMEX makes football predictions gives every step with the exact constants and a worked example. If a constant in the code changes, that page is updated at the same time, and an automated test fails if they disagree.

What a prediction looks like

Each prediction names a line and a price, for example Man City -0.75 at 0.92, or Over 2.5 at 0.90. The first is an Asian handicap pick: Man City must win by at least one goal for the bet to win, and by one goal exactly wins half. The second is a total goals pick. Underneath, a paragraph shows the statistics behind the pick: the expected goals, the probability our figure gives, the probability the market implies and the gap between the two. The time of publication is shown in UTC.

Once published, a prediction is frozen. We do not change the line, the price or the wording, and we do not remove predictions that turned out wrong. If a match had no prediction at the first assessment about a day before kick-off, the system looks again in the last hours before the match and can publish then. After kick-off nothing is added.

The public record

The record near the bottom of this page counts every prediction that has been settled. The results are win, half win, push, half loss and loss, and they are settled on the score at full time. It shows the share of predictions on the winning side and the profit or loss if every prediction had been staked at 1 unit at the published price. Both figures include all predictions, not a selection. A short list of the most recent settled predictions with their scores is under the totals so you can check them.

The record starts small. With few predictions the percentage means very little, and a run of ten or twenty can swing widely in either direction. Treat the record as evidence that builds up over months, not as proof after a weekend. If it shows that the rule does not work, we will say so and change or retire the rule.

What we do not do

We do not sell predictions, take payment for them, or ask you to join a group. We do not offer fixed matches, sure wins or guaranteed tips, and no such thing exists. We do not link to bookmakers from this page. We do not use team news, injuries or line-ups, because they are not in our data, which is why a prediction can look odd to a fan who knows that a key player is missing. And we do not publish a prediction for every match: many matches have none, and that is the correct output when the statistics agree with the market.

Using predictions sensibly

Treat any prediction, ours included, as one input among many and never as a reason to stake money you cannot afford to lose. The bookmaker's margin on a handicap market is roughly 3% to 5%, which means a method has to be right clearly more than half the time just to break even. Set a budget before you play, never chase losses and take a break if betting stops being fun. Gambling is for adults aged 18 and over and is not legal in every country. If you are worried about your own gambling, contact a local support service. Our policy page lists where to find help.

Frequently asked questions

Which football prediction site is the most accurate?

No site can be shown to be the most accurate from a few weeks of results, and any site that claims certainty is not being honest. Look for a site that publishes every prediction with its price and time, counts losses as well as wins and explains its method. LMEX publishes its full record on this page and its method on a separate page so that you can judge for yourself.

How can I analyse a football match for a prediction?

Start with recent form (the last five results), goals scored and conceded per match, and the head-to-head record. Then compare what those numbers imply with the market price once the bookmaker's margin is removed. The market already contains team news, so a view that differs from it needs a good reason. Our method page shows each step with a worked example.

Are the predictions here free?

Yes. Reading predictions and the record costs nothing and needs no sign-up. LMEX does not sell predictions or tips.

Can football results be fixed or predicted with certainty?

No. A prediction is a probability, not a result. Offers of fixed matches or sure wins are scams, and you should not pay for them. Even a prediction with a clear edge over the market loses a large share of the time.

Why does a match say no prediction?

Because the statistics did not differ from the market price by enough, the data on one of the teams was too thin, or the prices for the match were incomplete. Passing is the correct output of the rule when there is no clear edge, and it happens about as often as a published prediction.

Sources

• https://en.wikipedia.org/wiki/Poisson_distribution
• https://en.wikipedia.org/wiki/Asian_handicap