How scores work
Each card is assigned an affinity score for each investigator — a number between 0 and 100 that expresses how relevant that card is in current deck-building for that investigator.
The problem with raw frequency
A naive approach would count what percentage of decks include a card. But this disadvantages newer cards: a card from the Core Set (2016) has had years to accumulate appearances, while a card released last month has barely any. A simple count would always rank old cards higher, regardless of current meta relevance.
Exponential time decay
We weight each deck by how recently it was created. A deck from today counts fully; a deck from 90 days ago counts half as much; a deck from 180 days ago counts a quarter, and so on.
Maturity factor
Recency alone is not enough. When a new expansion launches, players immediately include its cards in decks — not because those cards are necessarily staples, but because they are new. To counteract this day-1 hype, we apply a maturity factor to each (deck, card) pair: a card that was brand-new when the deck was built contributes less than one that had been in the pool for months.
The formula
Recency weight of a deck:
recency_weight(deck) = e−λ × days_ago
Maturity factor of a card within a deck:
maturity_factor(deck, card) = 1 − e−card_age_at_deck / MATURITY_DAYS
Affinity score of a card:
score(card) = Σ(in_deck × recency_weight × maturity_factor) / Σ(recency_weight) × 100
where λ = ln(2) / 90 (recency decay) and MATURITY_DAYS = 60. card_age_at_deck = days between pack release and deck creation. Cards with unknown release dates are treated as fully mature.
Current parameters
| Recency half-life | 90 days — A deck created 90 days ago contributes half as much as one created today. |
|---|---|
| Maturity half-life | 60 days — A card included in a deck 60 days after its pack release contributes 63% of full weight; after 180 days, 95%. |
The "X / Y decks" shown below each score is the raw count — how many decks out of the total actually include the card, regardless of weighting. It gives context to the score.
What is excluded
Investigator identity cards, signature cards (unique to each investigator), and basic weaknesses are excluded from the ranking. Their presence is mandatory and carries no information about deck-building choices.
Exclusive Cards
Each investigator page shows the top 5 cards with the highest exclusivity ratio — cards this investigator uses disproportionately more than others. A card is "exclusive" not because only this investigator can play it, but because this investigator plays it far more often than the average.
Exclusivity ratio of a card for an investigator:
exclusivity(card, investigator) = frequency(card, investigator) / avg_frequency(card, active_other_investigators)
Active other investigators are those who include the card in at least 5% of their decks. Cards used by the target investigator in fewer than 5% of decks are excluded. The ranking uses the pure exclusivity ratio.
Often Paired
For each card we compute a <strong>co-play score</strong> against every other card: how often do the two appear together in the same deck? If 80% of decks that include Flashlight also include Perception, their co-play score is 0.80. This is measured globally across all investigators and all decks.
Co-play score of card B relative to card A:
co_score(A→B) = decks_with_both(A, B) / decks_with(A)
This is a directional score: co_score(A→B) can differ from co_score(B→A). The top 10 pairs are stored per card; 5 are shown by default on each card's detail page under Often Paired.
Usage Over Time
Each card page shows a chart of its global usage frequency by quarter since release — how often it appeared across all decks, regardless of investigator. This differs from the affinity score, which is per-investigator and time-weighted.
| Label | Meaning |
|---|---|
| Staple | Usage is stable or growing over time — a consistently useful card across the meta. |
| Declining | Usage has dropped moderately since release — still played but less central than at launch. |
| Hype | Usage peaked sharply at release then fell significantly — typical of cards that excited players early but proved situational or were outclassed by later options. |
| Late Bloomer | Near-zero early usage, growing later — often cards whose synergies or supporting investigators arrived after the initial release. |