There is no such thing as a 100 sure football prediction. Every match has three possible results, and the best forecasters in the world lose a large share of their calls. Our own public record, read off our live homepage this morning, is 109 wins and 78 losses from 187 graded picks, 58 percent. Our most confident band, the picks we rate at 70 percent or higher, has won 25 of 30, 83 percent, which still means one strong pick in six has lost. Any site that promises you a sure win is either hiding its losses or has never published them.
Why no football prediction can be 100 percent sure
A football match is not a two-way coin. It can end in a home win, a draw or an away win, and the draw is the result that quietly destroys most confident predictions. On our own record, 42 of our 78 losses were draws, 53.8 percent of them. Every graded pick we have published so far backed a team to win, so a draw lost it. The favourite did not lose those games. It simply failed to win them, and that is enough.
The betting market says the same thing in a different language. If a result were certain, a bookmaker would have to price it at odds of 1.00, which pays nothing. No bookmaker prices a football match that way, because even the heaviest favourite in a match between two professional sides loses or draws often enough that the price has to leave room for it. When a website sells you a sure pick at odds of 1.50 or 2.00, the odds themselves are telling you the pick is not sure. Our odds explained guide shows how to turn any price into the probability it implies.
Our graded record, losses included
These figures were read from our live homepage with a cache-busting request on 7 October 2026. We do not carry a figure over from yesterday, because our results are graded every day and the record moves. Draws count as losses. Postponed or unverifiable fixtures are voided and never touch the record.
| Measure | Figure | What it means |
|---|---|---|
| Graded picks | 187 | Every published pick that has been settled |
| Wins | 109 | 58.3 percent of graded picks |
| Losses | 78 | 41.7 percent, published in full on the homepage log |
| Losses that were draws | 42 of 78 | 53.8 percent of our losses were draws, not defeats |
| Strong band, 70% and above | 25 of 30 | 83 percent, our most confident calls still lost five times |
| Overall | 109W-78L | 58 percent, not 100 |
Here are ten rows from our live pick log, read the same morning, with the losses left in. Look at the Tunisia row. It was a 75 percent strong-band call, the kind of pick a sure-win site would sell as a banker, and it lost.
| Fixture | Competition | Our pick | Confidence | Result |
|---|---|---|---|---|
| Spain v Italy | Nations League | Italy win | 45% | WIN |
| Sweden v Poland | Nations League | Poland win | 44% | LOSS |
| Belgium v France | Nations League | France win | 42% | WIN |
| Georgia v Ukraine | Nations League | Ukraine win | 46% | LOSS |
| Tunisia v Botswana | AFCON qualifier | Tunisia win | 75% | LOSS |
| Zimbabwe v Congo DR | AFCON qualifier | Congo DR win | 44% | WIN |
| Eritrea v South Africa | AFCON qualifier | South Africa win | 74% | WIN |
| Hungary v Georgia | Nations League | Hungary win | 46% | WIN |
| Ukraine v Northern Ireland | Nations League | Ukraine win | 50% | LOSS |
| Football rows shown | 5 won, 4 lost |
The tenth row in that part of the log was an NFL pick, Pittsburgh to win in Cleveland at 57 percent, which also lost. We have left it out of the football table above but not out of this page.
The accumulator arithmetic: why sure-win slips collapse
This table is built from our own live hit rates, not from anyone else’s claims. If picks win independently of each other, the chance that every leg of an accumulator wins is the hit rate multiplied by itself once per leg. We ran it twice: once at our strong-band rate of 83.3 percent, which is the best figure we have, and once at our overall rate of 58.3 percent.
| Legs on the slip | All legs win, strong-band picks only (83.3%) | All legs win, all picks (58.3%) |
|---|---|---|
| 2 | 69.4% | 34.0% |
| 3 | 57.9% | 19.8% |
| 4 | 48.2% | 11.5% |
| 5 | 40.2% | 6.7% |
| 8 | 23.3% | 1.3% |
| 10 | 16.2% | 0.5% |
| Break-even point | Four strong legs already lose more often than they win | Two ordinary legs already lose two times in three |
Read the strong-band column carefully. Even if you only ever used our most confident calls, a four-leg slip would lose more often than it wins, at 48.2 percent. A ten-leg sure-win slip built from ordinary picks lands about one time in two hundred. This is why the long accumulators sold as 100 sure by tipsters on social media almost never come in, and why the screenshots of the ones that did are not evidence of anything. The independence assumption is a simplification: real picks on the same weekend can be linked, which can make the true figures slightly better or worse, but it cannot make a ten-leg slip anything close to certain.
Losing streaks are normal, even for an honest 58 percent
This is the second original dataset on this page. Take a forecaster who really does hit 58.3 percent, our live rate, and give them 187 picks, our live sample. How likely is it that somewhere in that run they suffer a streak of consecutive losses? We calculated it exactly, assuming each pick loses with probability 41.7 percent independently of the others.
| Longest losing run of at least | Chance of seeing it in 187 picks |
|---|---|
| 3 losses in a row | Above 99.9%, effectively certain |
| 4 losses in a row | 97.1% |
| 5 losses in a row | 75.6% |
| 6 losses in a row | 43.6% |
| 7 losses in a row | 20.9% |
| 8 losses in a row | 9.2% |
A run of five straight losses is more likely than not for a forecaster who is genuinely good. That matters for two reasons. First, if a site shows you a long unbroken winning run and no losing runs, its record is not real, because a real record of this size contains losing runs almost by arithmetic. Second, if you stake the same amount on every pick and cannot survive five losses in a row, your stakes are too big for any forecaster, including us. Our bankroll management guide covers stake sizing that survives a bad run.
How to spot a fake 100 sure prediction site
- No losses anywhere on the site. A real record over any meaningful number of matches contains losses and losing runs. A page of green ticks is a page someone has edited.
- No settled history you can check. Ask for every past pick with its date, its fixture and its result. If you only ever see the winners from last weekend, you are looking at marketing.
- Paid VIP groups and fixed match offers. Someone who really knew the result of a match in advance would bet on it themselves rather than sell it to you for a small fee.
- Confidence written as certainty. Words like banker, sure two odds and zero risk describe a feeling, not a probability. A real forecaster writes a number and lets you check it later.
- Correct score sure tips. Correct score is the hardest market in football. Our own correct score predictions page states its expected hit rate openly, and it is nowhere near 100 percent.
What to look for instead
Look for a site that publishes a confidence figure before kick-off, grades every pick afterwards, counts draws as losses when it backed a winner, and leaves the losses on the page. That is the whole of our method, and it is why we would rather show you 109 wins and 78 losses than a headline number we could not defend. Our explainer on what makes a good football prediction site goes through the checks in more detail, and how accurate football predictions really are sets our figures against the wider research.
Read next
- Today’s predictions, every fixture with its confidence figure
- Our full graded pick log on the homepage, wins and losses
- Bet of the day, including the days we say there is no bet
- Value betting: why the price matters more than the pick
- Double chance meaning, the market that covers the draw
- Draw no bet meaning, and what it costs you
- URC log and fixtures, our new rugby hub
- Responsible gambling and where to get help
Predictions on this site are published for information only and are not betting advice. No prediction is certain and past results do not guarantee future outcomes. 18+ only. Betting involves risk and you can lose money, so never stake more than you can afford to lose. Free and confidential help is available from the South African National Responsible Gambling Programme on 0800 006 008, 24 hours a day. Outside South Africa, contact your national gambling regulator for local support services.