Betting psychology: the gambler's fallacy, chasing losses and other traps
Betting psychology comes down to one thing: the brain sees patterns where chance is at work, and it suggests decisions that cost money. The most common traps are the gambler's fallacy ("after five defeats they're bound to win"), chasing losses, overconfidence and selective memory. In our database of 104,000 matches from 37 leagues (2015–2026), teams on losing or winning runs won practically exactly as often as the odds said they would. Below we describe each trap with numbers and a simple way to defend yourself against it.
Key takeaways
- 1
After 5 defeats in a row a team wins 23.7% of its matches, against 25.3% the market expected, so the streak is already in the price.
- 2
A favourite with a 70–80% chance loses roughly one match in four, and so does a "banker" at odds of about 1.30.
- 3
Raising the stake after losses repairs nothing: after three £20 losses a winning £60 bet at 2.00 only gets you back to zero, and at a 5% margin the loss that doubles your deficit is the likelier outcome (52.4%).
- 4
A 5-leg accumulator at about 75% a leg loses by exactly one leg in about 40% of cases, more often than it lands in full.
- 5
A betting record is the simplest defence against selective memory, because at a typical 5% margin you lose about 4.8% of every stake.
The gambler's fallacy: "the run has to end"
The gambler's fallacy is the belief that after a run of one kind of result, the opposite result is "due". A team has lost five times in a row, so "they've got to win eventually". Its mirror image is the "hot hand": a team has won five times, so it is "on a roll" and will win again.
We checked both beliefs on our data. For each run we compared how often the team won its next match with the chance the market gave it in the closing odds, with the margin removed:
| Before the match | Matches | Team won | Market expected |
|---|---|---|---|
| after 3+ defeats in a row | 12,861 | 29.4% | 29.3% |
| after 5+ defeats in a row | 2,688 | 23.7% | 25.3% |
| after 3+ wins in a row | 13,675 | 48.1% | 48.9% |
| after 5+ wins in a row | 3,387 | 56.4% | 56.7% |
The conclusion is simple: a run does not have to end, and it does not have to go on. The market already prices it. A team after 5 defeats wins 23.7% of its matches, slightly less often even than the odds suggested. Anyone who backs it because "they're due one" is paying for an illusion. In virtual football the same holds by design: every result is a fresh draw by a random number generator.
Tip: before you bet on a run ending or carrying on, ask yourself what you know beyond what is already in the price. If the only answer is the run itself, you have no argument.
Overconfidence
After a few winning bets it is easy to believe you "read" football better than the bookmaker. Yet the market is very well calibrated. In our database, favourites given a 70–80% chance by the odds won 75.5% of their matches, against an average price-implied chance of 74.2%. That means even a "banker" at odds of about 1.30 loses roughly one match in four.
Overconfidence is dangerous because it leads to bigger stakes. Our model test, described in why a betting model is not enough, showed that even a good statistical model does not beat the closing odds. A hunch after watching the highlights stands even less of a chance.
Tip: log every bet in a betting record and judge yourself after several hundred bets, not after a good weekend.
Confirmation bias: fan and tipster in one person
Confirmation bias means looking for information that fits what you already think and skipping the rest. A Newcastle fan before the Tyne–Wear derby reads about his strikers' form and misses the news that the first-choice centre-back is injured. In the end, the "analysis" confirms what he wanted from the start.
Tip: before you bet on your own team, check what chance the market gives it, for example on our list of matches with probabilities. If your view differs from the market's by more than ten percentage points or so, the likelier explanation is that you are wrong. If your feelings run too high, do not bet on your team's matches at all.
Chasing losses, or tilt
Chasing losses means raising your stakes after defeats to win the money back quickly. It is the most expensive trap, because it strikes exactly when emotions are strongest.
An illustrative example, step by step. You bet £20 a time and lose three bets in a row, so you are £60 down. You stake £60 at odds of 2.00 "to get back to level":
- If the bet wins, the return is 60 × 2.00 = £120, a profit of £60.
- The day's balance: −60 + 60 = £0. The best case is getting back to where you started.
- If the bet loses, the day's balance is −£120, twice as much as before you tried to win it back.
- Odds of 2.00 are not a coin toss. With a 5% margin built in, the chance behind them is about 47.6%, so the bet loses 52.4% of the time and is worth −£2.86 on average. Doubling the deficit is the likelier outcome.
Runs of losing bets are also completely normal. With bets at a 50% chance, the probability of at least 5 losses in a row somewhere in 100 bets is about 81%. A run is not a sign that the next one "has to come in", just an ordinary part of betting.
Chasing losses turned into a rule gives you progression systems, in which each stake has to win back all the previous losses. The most popular is the draw progression: in over 300,000 league matches a run of 15 matches without a draw happened 1,797 times, and at draw odds of 3.60 with a £20 target per cycle, 15 matches without a draw already add up to £2,616 of stakes, and 30 matches to £347,440.
Tip: a fixed stake and a daily loss limit, set before you start betting, not while you are at it. The details are in our 10 rules to protect your bankroll.
The illusion of control
The more time you spend on analysis, the more it feels as if you have a say in the result. An hour of reading stats gives you a sense of control, but it does not change the fact that 25.8% of the matches in our database end in a draw, against the 25.6% the market expected. In-play betting works the same way: you are watching the match and you "can see" a goal coming, but the live odds take it in faster than you do. AI betting tips work the same way too: a chatbot will add a convincing reason to every tip, but in tests from 2026 the leading language models predicted matches slightly worse than the odds.
Tip: analysis is there to help you understand a match, not to make you feel more certain. If after an hour of work your view does not differ from the price, the best decision is not to bet.
The "near miss"
An accumulator that "missed by one match" stays in your memory as an almost-win, not as a loss. This is the near-miss effect: the brain treats it as a sign that the strategy works and that you only need to try once more.
The numbers say otherwise. In our database, an accumulator of 5 favourites with a 70–80% chance each landed in 24.1% of cases. If each of the five matches has about a 75% chance, an acca with exactly one miss happens in about 40% of cases, more often than one that lands in full. "Just one match let me down" is simply the most common result of such a bet. How the chances and the margin multiply is covered in our article on accumulator bets.
Tip: in your record, log the acca that "almost came in" the same way as any other loss. Losing by one leg and losing by five legs cost exactly the same.
Memory of wins
A win at odds of 8.00 stays in your head for years. Twenty losing bets at 1.80 are gone within a week. That is why many casual bettors sincerely believe they "roughly break even", although at a typical 5% margin each £100 staked returns about £95.24 on average, so you lose about 4.8% of every stake.
Tip: count, do not remember. A spreadsheet has no selective memory.
The recency effect
The last few matches weigh more in your head than the whole season. A team that conceded in stoppage time in the last two rounds immediately "concedes late". We checked this: the minutes in which a team conceded its goals do not predict its next matches. We describe it in our article on whether teams really concede late goals.
Tip: before you decide something is a trait of a team, check how many matches you are seeing it in. Two or five matches are an anecdote, not a statistic.
The traps at a glance
| Trap | How it sounds in your head | What the data says | What to do |
|---|---|---|---|
| Gambler's fallacy | "After 5 defeats they're bound to win" | they win 23.7%, the market expected 25.3% | ignore the run, look at the price |
| Hot hand | "They're on a roll" | 56.4% wins against 56.7% in the odds | the run is already in the price |
| Overconfidence | "It's a banker" | a 70–80% favourite loses one match in four | fixed, low stakes |
| Chasing losses | "I'll win it back with one bet" | at 2.00 a win only gets you back to zero, a loss (52.4% at a 5% margin) doubles the deficit | a daily loss limit |
| Near miss | "Just one match let me down" | the most common result of a 5-leg acca | log it as an ordinary loss |
| Memory of wins | "I roughly break even" | the margin costs about 4.8% of every stake | keep a record |
When betting stops being entertainment
Betting should be paid entertainment with a set budget. Warning signs that you are losing control:
- you bet mainly to win back earlier losses,
- your stakes keep rising, although you did not plan to raise them,
- you deposit more than your budget allowed, or you borrow to bet,
- you hide how much you bet from the people close to you,
- you think about betting at work, at home and before you fall asleep,
- you feel anxious or irritable when you cannot bet.
If you recognise several of these in yourself, stop. Licensed bookmakers let you set limits on your deposits and spending, take a break and self-exclude, which blocks your account for a period you choose. GAMSTOP lets you exclude yourself, free of charge and in one go, from online gambling sites and apps licensed in Great Britain. Free, confidential support is available 24 hours a day from the National Gambling Helpline run by GamCare, on 0808 8020 133 or by live chat, and Gamblers Anonymous holds meetings every day across England, Wales and Ulster. Talking to someone who has been through the same thing often helps more than another "rule".
Summary
- A losing or winning run does not change the chances beyond what is already in the price.
- A "banker" at 1.30 loses roughly one match in four.
- Raising the stake after losses usually makes the loss bigger: a win only gets you back to level, while the likelier loss doubles the deficit.
- "Just one match let me down" is the most common result of an accumulator, not bad luck.
- A betting record is the simplest defence against selective memory.
FAQ
It is the belief that after a run of one kind of result the opposite is "due", for example that a team has to win after five defeats. In our database, teams after 5 or more defeats in a row won 23.7% of their matches, against the 25.3% the market expected. The run is already in the price.
Set a fixed stake and a daily loss limit before you start betting, and end the day once you reach the limit. A deposit limit in your bookmaker account helps too. After three losses, even a winning bet at 2.00 for three times your stake only gets you back to zero, and with the margin, losing it is the likelier outcome.
Form on its own is not an argument, because the market already prices it. In our data, teams on 5 or more straight wins won 56.4% of their matches, against 56.7% implied by the odds. Form raises a team's chances, but it also shortens its odds, which lowers the payout. How often football streaks carry on, BTTS and over 2.5 runs included, we checked separately on 299,000 matches.
Yes. With bets at a chance of about 50%, the probability of at least 5 losses in a row somewhere in 100 bets is about 81%. A run of losses does not mean the next bet "has to come in", nor that your rules do not work.
Because that is the most likely result. With five legs at about a 75% chance each, an acca with exactly one miss happens in about 40% of cases, while the whole bet lands less often. In our database, an accumulator of 5 favourites at 70–80% landed in 24.1% of cases.
Start with the tools in your bookmaker account: deposit limits, a break or self-exclusion. GAMSTOP blocks you from online gambling sites and apps licensed in Great Britain in one go, free of charge. The National Gambling Helpline run by GamCare (0808 8020 133) gives free, confidential support 24 hours a day, and Gamblers Anonymous holds meetings every day. It is also worth talking to someone close to you or to an addiction therapist.