How to run a prediction contest
that lasts a whole season
Everyone knows a prediction contest that died in November. It is never the players' fault: it is the scoring, the launch date, or the exhaustion of whoever keeps the score. Here is how to avoid all three.
Ligue 1 2026-27: kick-off on 21 August
Free, no stakes, no ads.
The four decisions to make before you launch
You make them once, in five minutes, and they decide everything else.
-
1
Who plays
Below four players, two bad weeks are enough to settle the table and interest collapses. Between six and twenty, a different person can win each matchday: that is the sweet spot. Beyond that, the overall table feels out of reach and you need per-matchday standings to keep everyone in the race. And go for a group that already exists, a gang of mates, a family, a team at work: a group assembled for the occasion never survives matchday three.
-
2
Which competition
A tournament (World Cup, Euros) lasts a month and gets played flat out, but it ends. A league season lasts all year and sets a weekly rhythm: it is the better choice for a first contest, because missing one matchday knocks nobody out. One competition at a time, though: with two, nobody knows where they stand any more, nor what they are playing for this weekend.
-
3
Which scoring
This is the decision that changes the game the most, and the one people make fastest. Scoring that under-rewards the exact score pushes everyone to play safe, and the table turns into a raffle. Set it before matchday one, write it to the group once, and never touch it again: changing the scoring mid-season, even to improve it, costs more trust than it buys fairness.
-
4
What is at stake
The most common mistake is believing you need money for people to play. A pot adds rules, accounts to keep and arguments, it shuts out those who do not want in, and it changes the nature of what you are running. What brings people back every week is the visible table and the right to take the mickey on Monday. If you want a prize, make it symbolic and pay for it yourself, never collect it from the players.
The scoring: the only rule that really matters
A prediction contest is played on the ratio between the exact score and the right result. At 3 against 1, nobody takes a risk and everyone puts 1-0. At 5 against 1, going for a 3-1 becomes worth it, and the game begins. Here is the scoring PronoArena uses by default, and why every line is there.
| What you got right | Points | Why |
|---|---|---|
| The exact score (2-1 called, 2-1 played) | 5 points | It is the only genuinely hard call: it has to be worth far more than the rest, or nobody attempts it. |
| The right goal difference (2-1 called, 3-2 played) | 3 points | Rewards the person who read the match, not just the winner. Without this line, a sharp prediction earns nothing extra. |
| The right result (win, draw or loss) | 1 point | The safety net: without it a whole matchday can score nothing, and that is when a player drops out. |
| The joker, placed on one match per matchday | Points doubled | Gives a lever to whoever has fallen behind. A contest you can no longer come back in is a contest you leave. |
| The stage multiplier (last 16, semis, final) | × 2 to × 4 | In a tournament a final cannot be worth the same as a group game: that is what keeps the suspense alive to the last day. |
This scoring is not an opinion: it has been there since the first version of the game and has run hundreds of matchdays. If you are running your contest by hand, copy it as is and you will save yourself two seasons of tinkering.
The five mistakes that kill a prediction contest
-
The hand-kept spreadsheet
It works for three matchdays. On the fourth, the organiser adds up ninety rows on a Sunday night, and the week they have not got time, the contest stops dead. A prediction contest almost always dies of its organiser's exhaustion, never of the players losing interest.
-
Predictions arriving by message
In a group where everyone posts their scores, latecomers publish after kick-off, nobody can prove the time, and the organiser spends the week refereeing. Predictions have to lock themselves at kick-off: that is the condition for a table nobody ever disputes.
-
Launching on the day
A contest launches the week before, not on the morning of matchday one: the link needs time to go round, people need time to sign up, and the undecided need time to decide. Launching at kick-off easily costs you half the group, and those people do not come back.
-
Charging, even five euros
A stake shuts out those who do not want in, forces someone to keep the kitty, and tips a game between friends into another category, with all the questions that follow. Experience is stubborn: free contests last longer than contests with a pot, because nobody has a reason to leave angry.
-
No ritual
A table nobody looks at does not exist. You need a fixed moment, always the same: the Monday morning message naming the winner of the matchday. It is that gesture, not the game, that means people are still talking about it in February.
By hand or with a tool?
The question deserves an honest answer, and it depends on the size of the group and how long it runs.
The shared spreadsheet
Free, your own rules, and everybody already knows how to use one. For five players over a month-long tournament, it is plenty.
It does not lock predictions at kick-off, does not fetch the scores, does not chase anyone, and it costs you an hour a week. That hour is the one that eventually runs out.
A tool that keeps the score
Results come in on their own, the table recalculates after every match, predictions lock at kick-off and the forgetful get a reminder. You go back to being one player among others.
The scoring is the tool's, and you take its rules instead of your own.
In short: for a short tournament with a handful of people, the spreadsheet does the job. Over a league season, or beyond a dozen players, the weekly data entry is the breaking point, and that is always where it breaks.
The checklist before matchday one
-
Pick the competition and note the date of the first matchday.
-
Create the league and give it a name that raises a smile: that is what will get passed around.
-
Write the scoring to the group once, in black and white.
-
Send the link a week ahead, where people already talk to each other.
-
Chase once, two days before. Not three times.
-
Post the table on the Monday after matchday one: that is the gesture that sets the ritual.
Running it at the office?
Among colleagues the questions are different: the first one is not the scoring, it is “are we putting money in?”. A separate page answers it, with what you need to propose the contest without anyone having to sign it off.
See the “Predictions at work” pageFrequently asked questions
How many people do you need for it to be fun?
Four at the very least, otherwise the table settles too quickly. The sweet spot is between six and twenty: enough for a different person to win each matchday, not so many that people get lost. Beyond that you need per-matchday standings so newcomers still have something to win.
When should the contest be launched?
A week before the first matchday. The link needs time to travel round the group, and the undecided need time to decide. Launched on the morning of the first match, a contest loses half its potential players.
Can people join mid-season?
Yes, provided every matchday is ranked separately: a newcomer then plays on equal terms from their very first matchday, even if they never catch up on the total. Without per-matchday standings, arriving in November is pointless and nobody does it.
Do you need a cash pot for people to play?
No, and it is actually counter-productive. A stake adds admin, shuts out part of the group, and takes the game out of pure banter. The visible table and a winner named every Monday are enough to bring people back, week after week.
Who enters the match scores?
Either you do, every week, by hand. Or a tool plugged into the official results does it, fetching them and recalculating the table without anyone having to think about it. It is the one job in the contest that comes back every single week, so it is the first one to get rid of.