A team wins five matches in a row. The newspapers call them “the form team”. Your friends say they “can’t stop winning”. The odds get shorter every week. So on Saturday you back them at home against a team from the bottom of the table.
It ends 1–1.
Was that bad luck? Maybe a little. But mostly it was something much more ordinary — something that happens to almost every team on a hot streak. It even has a name: regression to the mean.
This article explains it in plain words, with numbers from a simple computer test. No maths degree needed.
1What “Form” Really Measures
“Form” usually means the last five or six results: W W W D W. It looks clear and simple. That is exactly why it fools people.
It is a very small sample. Five games is five games. A lot can happen by luck in five games.
A win is a win. A 4–0 win and a lucky 1–0 win from one shot on target both show as “W”.
It hides who they played. Five wins against the bottom five teams is not the same as five wins against the top five.
It hides where they played. Four home games and one away game is an easier run than the other way round.
So the form line is not a lie. It is just a short, rough summary — and a short, rough summary is full of luck.
2Regression to the Mean, in Plain Words
Every team has a normal level — how good it really is, over a long time. Its results go up and down around that level, because football has a lot of luck in it: a deflection, a post, a penalty, a red card.
When a team has an extreme run — very good or very bad — it is usually a mix of two things: its real level, plus some extra luck. The real level stays. The extra luck does not. So the next results usually move back towards normal.
What it is not: “They won five, so now they must lose.” Football has no memory. Nothing pushes a team to lose.
What it is: “They won five, and part of that was luck. Their next games will probably look more like their normal level.”
Think of a student who usually gets 7 out of 10 in tests, and one week gets 10 out of 10. Next week, the best guess is not another 10. It is something closer to 7. The student did not get worse. The 10 was partly a lucky week.
3What the Computer Test Shows
We built an invented league of 20 teams, some strong and some weak, and played 5,000 full seasons on a computer — 38 rounds each, with goals decided by chance from each team’s real strength. The teams never changed. No injuries, no new coach, no “confidence”. Only luck went up and down. Here is what happened:
Teams that won 5 in a row took 3 points per game during the streak, of course. In their next 5 games they took only 1.62 points per game — almost exactly their normal level (1.56 in the rest of their season).
Teams that lost 5 in a row took 0 points during the run. In their next 5 games they took 1.15 points per game — again, close to their normal level (1.20).
The league leader after 6 rounds averaged 2.52 points per game in those first 6 rounds. Over the remaining 32 rounds, the same team averaged 1.65.
Who had a 5-win streak? About 1 team in 5 in every season. And about 1 in 6 of those teams was actually a below-average team.
Read the first line again. After the streak, the “form team” played like… itself. Not like a team that wins every week. The streak told you almost nothing about the next five games that you didn’t already know from the rest of the season.
Real football is not a computer, of course. Real teams do change — injuries, transfers, a new manager. But the luck part is always there too, and it is bigger than most people think.
4Why This Matters for Your Bets
Here is the problem. You are not the only one who sees the form line. Bookmakers see it. Other bettors see it. Newspapers write about it. When a team is on a hot streak, a lot of money goes on it, and the odds get shorter.
So by the time you back the “form team”, you are usually paying for the streak in the price. If the streak was mostly luck, you are paying for luck that is already gone.
What it feels like: “They’re flying. This is an easy win.”
What it often is: A normal team, a normal match, and odds that are a little shorter than they should be.
The opposite happens too. A team that lost five in a row looks like a disaster. Nobody wants to back them. If the bad run was partly bad luck, the odds can end up a little too long.
This does not mean “always bet against the streak”. That is just the same mistake turned upside down. Sometimes a streak is real — a new manager, a key player back from injury. The honest lesson is smaller: a streak alone is not a reason to bet. Not for, and not against.
Beginners often do the opposite of what helps: they jump on teams when they are hot and drop them when they are cold. In other words, they buy high and sell low.
5Form vs Real Strength
If form is mostly noise, what should you look at instead? Something that changes more slowly, because it is built on more games.
Longer records
In our test, we checked what predicted the second half of the season better: the last 5 games, or the first 20 games. The 20-game record was clearly better — its link with what came next was about 1.5 times as strong. More games, less luck.
The numbers under the result
Goals scored and conceded over many games, shots, shots on target. A team that keeps winning 1–0 while being outshot every week is living on luck. A team that keeps losing while creating more chances than its opponents is often better than its results.
Ratings like ELO
An ELO rating moves a little after every match — more for a surprise result, less for an expected one. It is designed to be slow, so one lucky week cannot push it very far. That is exactly what you want from a measure of strength.
Use form as: a warning light — “something may be going on here, let me look.”
Use long-term strength as: your starting point — “this is how good the team really is.”
6The Same Trap at Home and Away
Home advantage is real. In our invented league, home teams won about 44% of matches, draws were about 23%, and away wins about 32% — similar to many real European leagues.
But it is smaller than most people think, it is different from league to league, and in many leagues it has become smaller over the years. When matches were played in empty stadiums in 2020 and 2021, home advantage dropped sharply in many leagues — a good reminder of how much of it comes from the crowd.
And here the luck trap appears again. In our test, every team had exactly the same home advantage. Even so, about 1 team in 6 finished a season with a better away record than home record. Nothing was special about those teams. One season is simply not enough games to measure home and away form separately.
7Three Simple Rules
Rule 1: Don’t judge a team on fewer than 10 games
Five games can say almost anything. Look at 10, 20, the full season, or last season too. If a number changes a lot when you add more games, it was mostly luck.
Rule 2: When results and the numbers under them disagree, trust the numbers
Winning without creating chances usually doesn’t last. Losing while creating chances usually doesn’t last either. Results catch up with performance more often than the other way round.
Rule 3: A hot or cold team is a question, not an answer
When you see a streak, don’t ask “should I bet on it?”. Ask “why is it happening, and is the reason still there?”. A new manager or a key player back is a reason. “They’re in form” is not.
DO: look at long records, the numbers under the results, and slow ratings like ELO.
DO: treat every streak as a reason to look closer.
AVOID: betting on a team just because it is hot — or against it just because it is cold.
AVOID: trusting one season of home/away splits as if it were a fact.
8Trying It Yourself in CGMBet26
You don’t need a simulation to see this. In CGMBet26, open the League Tables of any league and pick the team with the best form. Then compare three things side by side: its last 5 games, its full-season record with the home/away breakdown, and its ELO graph over time.
Very often you will see the same picture: the form line jumps up and down, the season record moves slowly, and the ELO line barely moves at all. That calm line is the closest thing to the team’s real level.
Then go back a few seasons in the database and look for teams that started the season top of the table, or went on long winning runs. Check where they finished. You will find some real stories — and a lot of teams that simply drifted back to where they belonged.
9The Short Version
Form is a short, noisy summary. Extreme runs are partly luck, and luck doesn’t last — so hot teams usually cool down and cold teams usually warm up. The market sees the same form line you see, so a streak is often already in the price.
Look at more games. Trust the numbers under the results. Treat every streak as a question. That’s all.
If you haven’t yet, read Ten Losses in a Row Is Not a Signal — it is the same idea seen from the other side: what luck does to your results, not to a team’s. And Reading a Football Stats Table Without Lying to Yourself shows how to check the sample behind any number you see.
Comments