The Spire Intelligence Observatory
Make every draft decision on evidence.
SLAY2 is an independent strategy companion for Slay the Spire 2. It calculates reward, potion, “?” room and draw odds exactly, tracks the hidden counters behind them, and explains its card advice — without passing off a guess as a fact.
- Exact arithmetic
- Main v0.107.1 verified
- No account · runs in your browser
Try it — runs in your browser
How likely is a rare in your next reward?
Each card in a reward pack rolls its rarity against a hidden counter that starts at −5 points and climbs one point per non-rare card. Drag the slider to see how the counter changes the next pack. Example inputs, Ascension 0 — not your run.
- After a normal fight
- 29.5%chance of at least one rare in the 3-card pack
- After an elite
- 44.9%chance of at least one rare in the 3-card pack
What SLAY2 does
Exact odds
Numbers you can check
Card-reward rarity with the hidden pity counter, potion drops, “?” room outcomes, draw probabilities and shop prices — computed in exact rational arithmetic, verified against an independent implementation and the published tables.
Advisor
Advice that shows its work
Enter the cards you are offered. SLAY2 ranks them against your deck, breaks the score into reasons, and tells you how stable the call is. It is a transparent heuristic today — labelled as one, never dressed up as a win probability.
Evidence
A ledger, not a leaderboard
Every rule has a source, every model a measured result with its sample size — and “not yet evaluated” where nothing has been measured. Third-party simulator results are research output with unverified live-game parity.
Coverage notice
What is — and is not — supported today
| Area | Status | Detail |
|---|---|---|
| Game build | Supported | Main v0.107.1 (Steam build 23811903). Beta v0.111.0 is refused until its rules are verified. |
| Characters & ascension | Supported | Ironclad, Silent, Regent, Necrobinder, Defect · Ascension 0–10 (STS2 has no A11–A20, no Act 4, no keys). |
| Exact odds | Supported | Rarity + pity counter (incl. A7), potion counter, “?” rooms, draws, shop prices, gold. |
| Card advisor | Heuristic v0 | Transparent and uncalibrated. Class coverage is limited by how much each card’s effect is tagged. Where simulated runs tested a class’s card values, the advisor adds a labelled second opinion beside it. |
| Simulated card values | Research | Ironclad: playing by the measured values got past the Act-1 boss in 65.1% of 1,991 held-out simulated games against 43.3% for the hand-written ratings (p = 8.8e-57) · Silent: playing by the measured values got past the Act-1 boss in 18.5% of 1,999 held-out simulated games against 11.5% for the hand-written ratings (p = 8.5e-13) · Regent: playing by the measured values got past the Act-1 boss in 49.7% of 1,978 held-out simulated games against 14.0% for the hand-written ratings (p = 3.9e-154) · Necrobinder: playing by the measured values got past the Act-1 boss in 62.1% of 1,997 held-out simulated games against 47.8% for the hand-written ratings (p = 5.9e-30) · Defect: playing by the measured values got past the Act-1 boss in 76.2% of 1,996 held-out simulated games against 34.5% for the hand-written ratings (p = 7.2e-167). Simulator unverified against the live game; see Evidence & methods. |
| Combat lab, map, shops, events | Planned | Not available yet; nothing on this site pretends otherwise. |
| Co-op | Not supported | Per-seat odds exist in the game; scaling and shared decisions are unverified. |