Probability
How likely we think an outcome is, after calibration. For match result and total goals 2.5: the market-weighted blend.
Example: —% means similar ratings landed about —% of the time.
Every term behind a (?) on the site, in plain words. Examples show the shape of the number; the cards on the site fill in their own.
How likely we think an outcome is, after calibration. For match result and total goals 2.5: the market-weighted blend.
Example: —% means similar ratings landed about —% of the time.
A 0–100 ranking score that orders today’s published tips.
Example: A TipScore of — ranks this tip against today’s other tips.
The probability band: Very high 90% or more, High 80–89%, Strong 70–79%, Moderate 60–69%, Slight 50–59%.
Example: — because the probability is —%.
Our probability minus the market's, margin removed: a measurement, not a promise. Close to zero where we use the blend.
Example: —% − —% = — percentage points.
The gap vs market recorded at publish, kept on the public record. Not a claim that our models beat the market.
Example: The gap was — points when this tip was recorded.
The decimal odds that exactly match our probability.
Example: 1 ÷ — = —.
The highest odds found among tracked bookmakers at the last check.
Example: The best checked price here is —.
The chance suggested by bookmaker prices after removing their margin.
Example: The margin-adjusted market probability is —%.
How the best odds changed since bookmakers first priced the match.
Example: The price moved from — to —.
How many of our five models back this pick.
Example: — of 5 models lean the same way.
A rating that rewards wins over stronger teams and penalises losses to weaker teams.
Example: An Elo gap of — rates the home side higher.
A model that estimates each team’s likely goals from attack and defence strength.
Example: Its score distribution contributes to —%.
A tuned goal model that adjusts low scores and gives recent matches more weight.
Example: It adjusts the low-score part of this fixture.
A machine-learning model trained on past matches to estimate each market.
Example: It is one of — agreeing models here.
Match result and total goals 2.5: our models blended with the margin-free sharp price, weighted to the market.
Example: For this selection the market-weighted blend gives —%.
The mean score across every outcome considered, not an exact-score pick.
Example: The model mean is —.
The three strongest model-input reasons, written only from recorded numbers.
Example: These reasons were stored before kick-off.
The tip and its numbers entered the public record at this time.
Example: Recorded at —, before —.
Team sheets are out and the engine has re-run with them.
Example: This fixture was updated at —.
The next scheduled time the engine recalculates every fixture.
Example: The next run is at —.
Today’s published tips ordered by TipScore.
Example: This list currently contains — tips.
The published tip with today’s highest TipScore.
Example: This tip ranks first with —/100.
Free accounts see the top three tips in full each day.
Example: You have viewed — full tips today.
Starred tips followed for lineup and settlement alerts.
Example: The star follows one tip; it does not add a slip leg.
Every modelled market with its probability and gap vs market, beyond published tips.
Example: Scanner rows include published tips and unpublished markets.
Prices we track where a bookmaker offered more than our fair odds. They are not tips and have their own public record.
Example: This signal was — points above fair when recorded.
The share of settled, non-void tips that won.
Example: — ÷ (— + —) = —%.
The mean odds at which tips were published.
Example: Average published odds in this period are —.
The return from the same one-unit stake on every tip.
Example: This period returned —% across all published tips.
The gap between published odds and final market odds at kick-off.
Example: Closing-line value for this period is —%.
The largest peak-to-trough fall in running return for this period.
Example: The largest fall here was —%.
Whether predicted percentages match how often outcomes landed.
Example: We said —% and —% landed.
Outcome labels settled from the full-time score.
Example: This prediction settled as —.
A prediction that did not count; its odds are treated as 1.00.
Example: This leg contributes odds of 1.00.
Not settled yet: the match is still on, or its result has not reached us. A tip settles when its result arrives.
Example: Results usually arrive within a few hours of full time; otherwise at the next results update (Wednesday or Saturday).
Settled means a result exists; landed means it won.
Example: This item is —.
The engine generation that produced a recorded tip.
Example: This tip used model —.
A short code made from every input used for a tip.
Example: Fingerprint: —.
A public page showing exactly what was recorded and when.
Example: This link opens the recorded tip proof.
A tip taken down before the match started, with the reason kept on the public record.
Example: Withdrawn: —.
Your personal coupon of selections combined together.
Example: This slip has — legs.
One selection inside a slip.
Example: — is one leg at —.
The leg odds multiplied together; void legs count as 1.00.
Example: The independent leg odds multiply to —.
The chance every independent leg wins; each added leg lowers it.
Example: The independent probabilities multiply to —%.
Several likely picks become less likely when every one must land.
Example: Three 75% legs land together about 42% of the time.
A total odds level used to suggest the highest-probability legs.
Example: The current target is —×.
Picks from one match affect each other, so their chances are shown separately.
Example: These legs are excluded from the multiplied chance.
Engine tips were published; Your picks are modelled markets that were not published as tips.
Example: This leg is labelled —.
A short code that opens a shared TipSignal slip.
Example: Share code —.
Stake multiplied by combined odds, including the stake.
Example: — × — = —.
Stake is an amount; flat stake uses the same one-unit amount each time.
Example: No stake is entered by default.
Home win, draw or away win after 90 minutes plus stoppage time.
Example: This market has Home, Draw and Away outcomes.
Covers two of the three match results.
Example: Home or draw covers two outcomes.
A draw makes the selection void and returns the stake.
Example: A draw would settle this market as void.
Whether total goals finish above or below a line.
Example: Over 2.5 needs at least three goals.
Each team must score at least once.
Example: A 1–1 score makes Yes land.
Whether one named team scores the required number of goals.
Example: Home over 1.5 needs at least two home goals.
The exact final score; shown as probabilities but never published as a tip.
Example: The most likely exact score still has a small chance.
The last five results, with the most recent on the right.
Example: Form here is —.
Results from the latest meetings between these teams.
Example: The last five meetings are summarised here.
Dates and kick-offs use one zone: your globe-menu choice, else this device's. UK time follows the clock change.
Example: This kick-off is —.
Display odds as decimal, fractional or probability.
Example: The current format is —.
Flat colours, no blur or animation, deferred images and smaller pages.
Example: Low-data mode is —.
A morning email with today’s tips and yesterday’s results.
Example: The digest is scheduled for —.
Hides tips and stops alerts for a chosen period.
Example: Tips and alerts are currently —.
Hides tips, alerts and slips until the selected period ends.
Example: Slip and tip controls are disabled while active.
Download your stored data or delete your account and data.
Example: These controls are available in Account.
Seven days of Pro at no charge, with a reminder before it ends.
Example: This trial ends on —.
The country on your account sets the price you pay and the provider; the display region only changes the currency shown.
Example: The current billing country is —.