Expansion Dilution — Is B&C Really Swingier? Does NI Dilute?
“B&C alone felt swingy; JE balanced that; NI diluted things.”
This chapter tests the claim with per-expansion marginal statistics. For each expansion (Base, B&C, JE, NI) we assign every card to its home expansion (the earliest/most-specific set tag) and compute mean, standard deviation, and coefficient of variation (CV) — CV = σ/μ, the standard measure of swinginess independent of scale. A higher CV = wider spread of outcomes relative to the average = swingier.
Data pipeline: scripts/expansion-deltas.py → data/references/deck-stats/per_expansion_deltas.json.
Headline result
B&C is the swingiest expansion for Fear + Event decks. JE is the most balanced. NI does NOT dilute Fear — it actually brings the heaviest-hitting Fear cards per card added — but NI DOES dilute Major Powers. Full breakdown below.
Fear deck: swinginess per expansion (net-favor)
Net-favor = (#spirit-favoring keyword hits) − (#invader-favoring keyword hits) per card, summed across all three terror-level stages. Higher CV means the card-to-card swing is wider.
| Expansion | n | mean net-favor | σ | CV | median |
|---|---|---|---|---|---|
| Base game | 15 | 0.40 | 0.51 | 1.27 | 0 |
| Branch & Claw | 15 | 0.13 | 0.35 | 2.64 ← swingiest | 0 |
| Jagged Earth | 6 | 0.50 | 0.55 | 1.09 ← most balanced | 0 |
| Nature Incarnate | 9 | 0.33 | 0.71 | 2.12 | 0 |
Reading: B&C’s Fear-card net-favor has the highest CV (2.64) — individual draws are nearly 3× as variable relative to the mean as JE’s (1.09). Your intuition is measurably correct: within the same Terror level, B&C Fear cards are the most unpredictable swing-cards in the game.
Fear deck: dilution analysis
The “dilution” claim is about cumulative mean — when you add a later expansion’s Fear cards to the running pool, how does the combined average shift? Negative delta = new expansion is weaker (dilutes the pool); positive delta = stronger (concentrates variance).
| Added | +cards | Mean per added card (fear_amount) | Cumulative-pool mean | Δ vs. previous |
|---|---|---|---|---|
| Base game | +15 | 0.00 | 0.00 | (baseline) |
| Branch & Claw | +15 | 0.00 | 0.00 | +0.00 |
| Jagged Earth | +6 | 0.17 | 0.03 | +0.03 |
| Nature Incarnate | +9 | 0.33 | 0.09 | +0.06 |
NI does NOT dilute the Fear deck — NI’s 9 Fear cards carry the highest per-card embedded “N Fear” directives (mean 0.33), so adding NI concentrates Fear-generation in the pool. What NI does change is the mix: more Fear cards that explicitly generate Fear on draw (via event-style triggers), vs. base/B&C Fear cards that mostly change game rules.
Where’s the “dilution” feeling coming from then? Probably from the Event deck, which is larger and shuffled every game — see below.
Event deck: the real swingy one
| Expansion | n | mean net-favor | σ | CV |
|---|---|---|---|---|
| Branch & Claw | 23 | 0.09 | 0.95 | 10.92 ← extreme swing |
| Jagged Earth | 30 | 0.53 | 1.17 | 2.19 |
| Nature Incarnate | 9 | 0.78 | 0.83 | 1.07 ← tight |
B&C’s Event deck is enormously swingy — CV of 10.92 because the mean is near-zero (0.09) but the stdev is almost 1.0. That means B&C Events individually swing wildly positive or negative but cancel out on average. JE adds 30 more Events with a consistently pro-spirit lean (mean +0.53, CV 2.19). NI’s 9 Events are the most tightly clustered (mean +0.78, CV 1.07) — NI Events lean pro-spirit with low variance.
Interpretation
- B&C Events feel random because they are: each card is either a big boon or a big bust with few middle-ground cards.
- JE Events balance by adding many consistently-themed cards that pull the mean pro-spirit.
- NI Events concentrate — few cards, all pro-spirit.
Minor deck: mean Fear output per card
| Expansion | n | mean fear_amount | CV |
|---|---|---|---|
| Base | 50 | 0.58 | 1.81 |
| B&C | 30 | 0.67 | 1.64 |
| JE | 33 | 0.58 | 1.89 |
| NI | 1 | 1.00 | 0.00 |
B&C Minors generate the most Fear per card (mean 0.67 and lower CV — more reliable). JE Minors match base in mean Fear but have slightly higher variance. NI only adds 1 new Minor (most NI content is Uniques + Majors).
Major deck: NI does dilute the Major pool
| Expansion | n | mean fear_amount | CV |
|---|---|---|---|
| Base | 41 | 1.05 | 1.36 |
| B&C | 22 | 2.00 | 1.41 |
| JE | 23 | 1.04 | 1.43 |
| NI | 12 | 0.83 | 1.60 |
B&C Majors are the peak — 2.00 mean Fear per Major card (double any other expansion). NI’s 12 new Majors average 0.83 (lowest), so drafting from an all-expansions Major deck does pull your expected Fear-output down compared to drafting from just base+B&C. The NI-dilutes-Majors claim is measurably true.
Rough per-card spread for Majors
- B&C Majors: highest fear + widest range of big effects. The 2.00 mean comes from power-level-5 Majors like
Vengeance of the Deadfamily andCast Down into the Briny Deep. - NI Majors: trend toward non-Fear effects (Beast-summoning, land-transformation, presence-redistribution). Different strategic axis, lower direct Fear output.
So what’s the full verdict?
| Claim | Verdict | Evidence |
|---|---|---|
| “B&C alone feels swingy” | ✅ Confirmed | B&C Fear CV 2.64 + Event CV 10.92 — both highest |
| “JE balances” | ✅ Confirmed (for Fear + Events) | JE Fear CV 1.09 (lowest); JE Events CV 2.19 (lower than B&C) |
| “NI dilutes” | ⚠️ Partial | NI concentrates Fear cards (mean 0.33 > JE 0.17); NI dilutes Majors (mean 0.83 < B&C 2.00) |
Strategic implications
- Playing B&C-only: expect high variance. Plan for worst-case Event draws; don’t build strategies that depend on a specific Event card appearing.
- Playing B&C + JE: JE’s 30 more Events pull the combined Event-pool mean pro-spirit and reduce CV from 10.92 → somewhere around 4-5. Less swingy; tactical planning becomes more reliable.
- Playing all expansions: Fear drawn stays roughly constant per-game (threshold-driven), but the Major deck’s mean Fear output drops. Pivot Major-shopping toward JE-era Majors (still 1.04 mean, lower variance than NI) when the draft offers cross-expansion Majors.
Source + methodology
- Every card classified by its
setfield on the Wiki, assigned to the earliest matching expansion for marginal analysis. - Fear-amount: sum of
N Fearregex matches across all three stages (Fear cards) or the full text (Events/Minors/Majors). - Net-favor: regex-based keyword count of pro-spirit (
Destroy N Explorer,Push N,N Fear,Remove Blight, etc.) minus pro-invader (add N Explorer,extra build,destroy Dahan, etc.) matches per card. - Scripts:
scripts/expansion-deltas.pycomputes per-expansion moments; output atdata/references/deck-stats/per_expansion_deltas.json. - Cross-reference: Deck Expansion Impact for the combo-pool view and Monte Carlo per-game outcomes.