Vault Breakers: Super Score — Exact Results, `draft-v0.2`

DEV ONLY · NOT FOR PRODUCTION · NOT FOR CERTIFICATION. Everything here is a closed form. Nothing is sampled, and nothing sampled appears in it. Monte Carlo results live in ../simulation/draft-v0.2.md.

Supersedes draft-v0.1.md, which is retained unchanged.

Field	Value
Date	2026-08-19
Game ID	`vault-breakers-super-score`
Math version	`draft-v0.2`
Canonical config	`libs/mathengine/games/vault_breakers_super_score/vault-breakers-super-score.yaml`
Config SHA-256	`fa6d0410d7de619e68b42083e7012b975d70be572a252d0dd60f3debe2ba4fc8`
RNG algorithm/version	`sha256-trunc64-be/go1-alfg#1`
Produced by	`.../exact_test.go`, `TestExactSnapshotReport`

What changed from v0.1

Free spins now run their own reel table. That is the whole point of the tuning pass: in v0.1 free spins ran the base reels, so a fifteen-spin feature was worth 15.52× and the reference buy prices of 100× and 500× returned 15.52% and 10.94%. Making the feature worth buying required making its reels different, which required FreeSpinsConfig.AtomicSimUsesVariants so that the simulator plays the same reels the runtime does.

Both tables obey the same draw model, so the same closed forms apply to each.

Per-cell probabilities

Weights out of 50,000, read from the strips the engine executes.

Symbol	Base weight	Base p	FS weight	FS p
`H1`	6,153	0.123060	5,250	0.105000
`H2`	6,153	0.123060	5,251	0.105020
`H3`	6,153	0.123060	5,251	0.105020
`H4`	6,153	0.123060	5,251	0.105020
`L1`	6,153	0.123060	5,251	0.105020
`L2`	6,153	0.123060	5,251	0.105020
`L3`	6,154	0.123080	5,251	0.105020
`L4`	6,154	0.123080	5,250	0.105000
`MULTIPLIER`	30	0.000600	7,250	0.145000
`SCATTER`	722	0.014440	722	0.014440
`SUPER_SCATTER`	22	0.000440	22	0.000440
total	50,000	1.000000	50,000	1.000000

Scatter class = 744 / 50,000 = 0.014880 on both tables, deliberately: the retrigger rate is then the same question as the trigger rate rather than a second thing to reason about. Super share of the class q = 22 / 744 = 0.02956989.

Exact results — one i.i.d. BASE snapshot

Metric	Method	Value
Regular-symbol RTP	EXACT	0.56250204 — 56.2502%
— `H1`	EXACT	0.24552142
— `H2`	EXACT	0.12363634
— `H3`	EXACT	0.07604315
— `H4`	EXACT	0.04936591
— `L1`	EXACT	0.02367917
— `L2`	EXACT	0.01984274
— `L3`	EXACT	0.01437683
— `L4`	EXACT	0.01003648
Hit rate	EXACT	0.19759904 — 1 in 5.0608
Charge cells per snapshot	EXACT	0.0180
P(4+ scatter-class)	EXACT	0.0009860307 — 1 in 1014.2
P(3+ scatter-class)	EXACT	0.0099096597 — 1 in 100.9
E[instant prize \| qualifying snapshot]	EXACT	14.210410×
E[instant prize] per snapshot	EXACT	0.01401190 — 1.4012%

Exact results — one i.i.d. FREE-SPIN snapshot

Metric	Method	Value
Regular-symbol RTP	EXACT	0.22144564 — 22.1446%
— `H1`	EXACT	0.09666009
— `H2`	EXACT	0.04864223
— `H3`	EXACT	0.02973830
— `H4`	EXACT	0.01940891
— `L1`	EXACT	0.00940294
— `L2`	EXACT	0.00789717
— `L3`	EXACT	0.00570532
— `L4`	EXACT	0.00399069
Hit rate	EXACT	0.08233109 — 1 in 12.1461
Charge cells per snapshot	EXACT	4.3500
P(3+ scatter-class), the retrigger	EXACT	0.0099096597 — 1 in 100.9

Reading these two tables together

A free-spin snapshot pays a regular win less often than a base one — 8.23% against 19.76% — because charges occupy 14.5% of the grid and every cell a charge takes is a cell that cannot complete an eight-of-a-kind. What it gains is 4.35 charges per snapshot against 0.018, and the charges are summed and multiply whatever the sequence did win.

That is the shape of the feature: rarer wins, far larger ones. It is also the reason the free-spin regular RTP per snapshot (22.14%) is well below the base (56.25%) while the sequence as a whole is worth roughly a hundred times a stake.

Charge pools

Value	Base weight	Base probability	Super weight	Super probability
2×	692	0.692	—	—
3×	182	0.182	—	—
5×	72	0.072	—	—
10×	30	0.030	526	0.526
15×	13	0.013	200	0.200
20×	—	—	120	0.120
25×	6	0.006	80	0.080
50×	3	0.003	45	0.045
100×	1	0.001	25	0.025
250×	—	—	3	0.003
500×	1	0.001	1	0.001
mean		3.6850×		18.6600×

The Super pool mean is 5.06× the base pool's, which is what puts the Super buy at five times the Standard buy's price and holds both at their reference values.

What this does and does not establish

Internally exact, produced by the same engine that executes the game. Not independent verification, and this document is not certification.
Exact figures cover one snapshot. A spin is one or more, so whole-game figures are larger and are measured.
The instant-prize expectation per snapshot is exact; its contribution per spin depends on how many snapshots a spin produces, which is measured.

#Vault Breakers: Super Score — Exact Results, draft-v0.2

#What changed from v0.1

#Per-cell probabilities

#Exact results — one i.i.d. BASE snapshot

#Exact results — one i.i.d. FREE-SPIN snapshot

#Reading these two tables together

#Charge pools

#What this does and does not establish