Most prediction games stop at "who wins." Pick the winner, get a point, move on. Easy.
But anyone who's watched enough sport knows the binary outcome is the least interesting question. Two teams can have wildly different scoring patterns, defensive shapes, and game states — and yet end up with the same winner. Predicting the exact score forces you to engage with all of it.
That's why our scoring system rewards precision exponentially:
- +5 pts — Exact score (you said 3–1, it's 3–1)
- +3 pts — Correct winner and exact goal difference
- +2 pts — Correct winner, wrong margin
- +0 pts — Wrong winner
The gap between +2 and +5 isn't arbitrary. It reflects how much more signal is in the exact prediction versus a coin-flip-tier "team A wins."
What predicting the score actually measures
When you commit to "Finland 3, Sweden 1," you're implicitly making a lot of statements:
- Both teams will score — not a clean sheet, not a high-scoring shootout
- The game won't go to overtime/penalties (in formats where that matters)
- The favorite wins by exactly two goals — not one, not three
- The tempo is moderate — four total goals, average for top-flight hockey
Each of those is a small bet against a distribution. Get most of them right and you'll land on the exact score. That's not luck — that's a model of the game in your head, calibrated against years of pattern recognition.
The honest math: why exact scores are hard
In hockey, the most common scoreline in NHL regulation is 3–2 (~9% of all games). Even guessing 3–2 every single time, you'd be wrong 91% of the time on the exact score. The frequencies for football are even flatter — 1–1, 2–1, 1–0, and 2–0 all hover around 9–11% individually.
So a perfect-score prediction is genuinely rare. We could have made +5 worth +10 to crank the variance, but we landed on +5 because:
- It's enough of a multiplier to matter across a season
- It's low enough that one lucky guess doesn't lock down the leaderboard
- It rewards consistency over flash
Why we don't reward "draws are hard" the same way
We tried. The early drafts of the scoring system had a "+1 bonus for correctly predicting a draw" because draws are statistically less common than wins. But it warped the meta — players started over-predicting 0–0s to farm the bonus, and the leaderboard rewarded contrarianism over actual sport knowledge.
The cleanest rule is the simplest one: predict the precise outcome you believe will happen, and we'll reward you proportionally to how close you got. Everything else is incentive design noise.
What to do with this
If you're new to YourFriendsLeague, the most common mistake we see is predicting too cautiously. People hedge toward 1–0 / 2–1 / 0–0 because those scorelines feel "safe."
They're not safe. They're just common — which means everyone else is also guessing them. The leaderboard rewards being right and uncommon, not just right.
If you genuinely think tonight's game is a 4–3 thriller, predict 4–3. If you're wrong you'll get 0 or 2, which is what you'd get from a safer guess anyway. But if you're right — even roughly right — you'll pull ahead of every "1–0 just in case" prediction in the league.
Trust your read of the game. The points follow.