Methodology

How WinSight works

WinSight turns one statistical model into every number on this site. This page explains how that model thinks, where it is strong, where it is limited, and how to read a probability without fooling yourself. None of it is betting advice.

Frequently asked questions →Guides →

One model, four markets

Every prediction starts from the Dixon–Coles model. It estimates how strong each team is at scoring and at preventing goals, adjusts for home advantage, and produces a full grid of scoreline probabilities — the chance of 0–0, 1–0, 2–1, and so on. The match result, over/under, both-teams-to-score and correct-score numbers are all read off that single grid. That is why they always agree with one another: they are different views of the same underlying picture.

Why recent matches count for more

A team is not the same from one season to the next — players, managers and form all change. So the model weights recent matches more heavily than old ones, letting the distant past fade gradually rather than counting a result from five years ago as if it were yesterday's. How quickly the past fades is not guessed; it is tuned on the walk-forward backtest, where we can see which setting would have predicted best without ever letting the model see the future.

Teams new to a league

A newly promoted side has no recent record in its new division, so the model cannot yet judge it directly. It starts that team from the average of comparable teams — those around it in strength — and then lets the team's own results take over as they arrive. Within about ten matches the real evidence has replaced the starting assumption. Early predictions for these teams carry more uncertainty, and that uncertainty is meant to be there.

Why we don't claim to beat the bookmakers

This is the plainest thing we can tell you: WinSight does not claim to beat the bookmakers, and you should be wary of anyone in this field who does. By the time a match kicks off, the betting market has absorbed almost everything that can be known about it — team news, injuries, and the weight of real money pushing prices toward the truth. Closing odds are the sharpest forecast in the sport, and no public model reliably beats them. On the same 16,322 matches, our honest walk-forward backtest trails Bet365's closing favourite by 2.16 percentage points of accuracy. We publish that gap rather than hide it.

So what is WinSight for? For probabilities that are honest about their own uncertainty, and for a record that is public, permanent, and never edited — including the predictions we got wrong. That record is the one thing a confident-sounding tipster can never show you, and it is the only thing we ask you to judge us on.

See how we're measured →

How to read a probability

A 60% chance is not a promise. It means that if the same situation played out many times, we would expect it to happen about six times in ten — which also means it should fail about four times in ten. A single high-confidence prediction going the wrong way is not a broken model; a model whose 60%s only come true 40% of the time is. That difference is exactly what calibration measures, and it is why we publish our calibration curve rather than just a headline accuracy number. Read the probabilities as a careful estimate of uncertainty, not as a forecast of what must happen.

Straight answers

Questions, answered plainly

Can this guarantee a profit?
No. No probability model can guarantee a profit. WinSight is for 18+ (21+ where required). It publishes statistical probabilities for information only. They are not betting advice, and they are not guarantees. No model removes uncertainty. A historical record or a backtest cannot tell you what will happen next. If you gamble, that decision and its consequences are yours, and gambling can lead to financial loss. WinSight does not accept or place bets. Safeguards and independent help are on the Responsible Gambling page.
Does WinSight beat the bookmakers?
No. Closing odds are extraordinarily hard to beat, and we make no such claim.
Is this betting advice?
No. WinSight publishes statistical information only. It is not betting advice.
What is WinSight?
WinSight publishes football probabilities you can audit. Every prediction we publish is settled after the match and stays online permanently, including the ones we got wrong. WinSight is not a bookmaker and does not accept or place bets.
How are the predictions made?
Predictions come from a Dixon–Coles model. It estimates each team's attacking and defensive strength, adjusts for home advantage, and builds one grid of scoreline probabilities. Match result (1X2), total goals over or under 2.5, both teams to score, and correct score are all read from that grid. The production model is retrained once a week. Upcoming fixtures are predicted once a day, after the evening fixture sync. Recent matches count for more than older ones. A team new to a league starts from comparable teams until its own results take over. A published probability is an estimate of uncertainty, not a promise that an outcome must happen.
Which competitions does WinSight cover?
WinSight currently covers 7 active competitions: Premier League, Championship, La Liga, Bundesliga, Serie A, Ligue 1, and UEFA Champions League. Eredivisie and Primeira Liga are registered as pending and are not covered. Only competitions marked active are published.
How does the public track record work, and how are results settled?
Every published prediction is settled after the match and stays online, including the ones we got wrong. For a finished match, the published pick is compared with the final score. The match-result pick is the home win, draw, or away win with the highest probability. Total goals are over 2.5 when that probability is at least one half, and the match is over when the teams score three or more. Both teams to score is yes when that probability is at least one half. Correct score uses the single most likely scoreline. The call is a win when the pick matches the final score, and a loss when it does not. A cancelled match is void. A postponed match is voided only after it has stayed unplayed for 14 days past kickoff; if it is rescheduled, it settles on the real result. Settlement runs after each of the two daily result syncs. On the public record, a match is counted once. Forecasts count each settled match-result (1X2) call from a non-development model version. Void results are excluded, and the other markets are not added to those totals. Hit rate is wins divided by wins plus losses, for that model version only. Versions stay separate and are never pooled. A live sample smaller than 100 settled forecasts is marked as not yet strong evidence. Each Monday, the previous week's settled predictions are published as a weekly report, and a database trigger rejects later updates and deletions of that report.
What is free, and what does Pro include?
Free is $0. Completed-match reviews, the accuracy record, and the methodology stay open without a subscription. A signed-in reader can also see basic match-result (1X2) probabilities for the earliest two non-marquee matches in the next seven days. On those free matches, confidence and the other markets stay locked. Guests do not see pre-match probabilities. Pro includes every upcoming match-result (1X2), goals over or under 2.5, both-teams-to-score, and correct-score probability, plus the model evidence for every upcoming match. Pre-match probabilities outside the free matches stay behind the subscription.
What does WinSight cost?
Prices are in USD. Applicable taxes may be added at checkout. Free is $0. The public monthly plan is Founding at $5.99 per month, with about 90 places left. Pro annual is $89 per year, an effective $7.42 per month. Each paid plan renews at the price and billing cycle confirmed at checkout until it is cancelled. An eligible subscriber keeps that founding monthly price while the subscription stays active. Cancelling ends the locked price. A new-user welcome offer ends at 19 November 2026, 23:59 UTC+8 (15:59 UTC). Annual code welcome75 takes $14 off the first payment, so the first year is $75 instead of $89. Monthly code welcome3 takes $3 off the first payment. The codes are for new subscribers only, once per person, and only on the first payment. After that first billing period, the plan renews at its regular price.
How do I cancel or manage a subscription?
If you are signed in and have a subscription that can be managed, Manage billing on the account page opens the Whop billing portal for payment details and invoices. You can cancel at any time on WinSight's cancellation page. An ordinary cancellation is accepted automatically, asks for no reason, and stops the next renewal. Unless a refund or a mandatory law applies, paid access continues until the end of the current billing period. The refund policy also says you may cancel through the Whop billing portal linked from your WinSight account or receipt. Cancelling does not itself return money already paid.
What is the refund policy?
WinSight supplies the subscription. Whop hosts checkout and processes payments. WinSight decides refund requests under the published refund policy, and an approved refund is processed by Whop back to the original payment method where possible. Cancellation alone does not return an amount already paid. Eligibility depends on the transaction, the timing, how the service was used, whether the service was available, and the law that applies where you live. The policy does not limit any mandatory refund, withdrawal, or other consumer right. Consumers in the EU, the EEA, and the UK generally have a 14-day right to withdraw. For a digital service delivered immediately, that right is lost after access begins only if you expressly asked for immediate delivery and acknowledged that loss at checkout. If that acknowledgement was not recorded, the statutory right remains available as the law requires. The full text is the Refund Policy.