Scoring
How Pack One scoring works
What the 100-point score measures, how Top 3 differs from Full Pack, and why support is not win probability.
Draft Run: ten choices, one score.
Matching the trophy drafter earns 100 points for that decision. Other cards earn up to 95 based on their contextual support relative to the model’s strongest option: round(95 × selected support / highest support). An alternative with 90% of the leader’s support earns 86; half the support earns 48. Each of the ten decisions contributes equally to your final average.
The model already accounts for pick depth and the original drafter’s earlier cards, so the same relative support earns the same partial credit at an early or late pick. Rerolls replace a decision before you answer; the discarded pack has no effect on your score.
Match the trophy drafter.
The trophy drafter’s actual choice is deliberately the 100-point target in Draft Run and Powered Cube Run. Strong-player model support determines partial credit for every other card. This is the game’s scoring design.
Daily standings and ties.
Your first Daily attempt is ranked. Today’s position can move as more people play; final percentiles use the closed board after midnight Eastern. Percentiles require at least ten players and include everyone tied at your score. Weekly, monthly, and all-time Draft Run boards compare Daily averages and show days played alongside them.
The other modes use their own scoring.
Pack One scores measure how closely your choices track the strong-player consensus model for the cards that were actually available. A score of 92 does not mean your pick had a 92% chance of being correct, winning the draft, or producing a trophy. The model probabilities are comparative support values inside the decision, not calibrated outcome probabilities.
Full Pack rewards meaningful decisions.
For an individual pick, the selected card is compared with the most-supported card in that pack. Support is transformed so that a near miss remains close rather than collapsing into a binary right/wrong result. The final Full Pack score also accounts for how many real alternatives remained. An early pick from fourteen cards carries more information than a forced final pick, so late decisions receive less weight.
This is why Full Pack can distinguish between a drafter who repeatedly chooses from the consensus cluster and one who consistently takes the least-supported option, without pretending every pick has equal strategic importance.
Top 3 is mostly about the set, then the order.
Top 3 scoring puts most of its weight on whether your three selected cards belong in the model's strongest group, with a smaller component for putting those cards in the same order. That means reversing three very strong cards should still score well, while replacing one of the true contenders with a much weaker card should cost more.
Consensus support is descriptive.
The model summarizes patterns in a selected cohort of experienced 17Lands drafters and conditions those tendencies on the cards already in the replay pool. It is designed to describe strong-player behavior more intelligently than a raw pick-rate table. It does not claim to establish objective card grades.
Scores are for comparison over time.
The most meaningful use of the 100-point scale is relative: compare your own sessions, compare the same seeded pack with a friend, or compare Daily Challenge standings. Do not read small point differences as scientific certainty. A score is a compact feedback signal attached to a richer decision record.