Puzzle design research

Designing a Hint That Does Not Leak the Answer

A reveal supplies an answer value; a public-rule hint supplies a reason derived from the current position and published constraints. In the Bit Balance behavior reviewed here, Nudge explains a public deduction without filling the cell. This guide compares those information models and does not claim that one produces a measured learning outcome.

By BeanNest Studio Reviewed September 20, 2026

What this guide helps you decide

A hint that tells you the right cell is just a solution. What can a hint usefully do instead?

This research explains the public-rule hint model evidenced by the published Bit Balance behavior. It does not claim measured learning outcomes, publish hint output for a live puzzle, or describe an internal solver.

Try it on the anchor product: BeanNest Games /games/.

The answer-oracle model

An answer oracle compares the current board against a stored solution and can report a value or mismatch from that hidden reference. Its defining dependency is the stored answer. That is different from a rule-based explanation, which can justify a constraint from public state. This is a structural distinction, not a claim about how players learn or how often they request help.

The public-rule model

A public-rule hint looks only at the current board and the published constraints, then identifies a rule that applies. It can say that a constraint is currently violated, or that a particular pattern makes one value impossible. It never needs a stored solution, and the player receives a reason rather than a value.

Why the rule model also improves verification

Because the rule model works from the board and the constraints, its output can be checked the same way any move can be checked. The repository's evidence contract separates canonical puzzle state from presentation precisely so rules can be evaluated outside the browser. A hint system built on public rules inherits that verifiability, while an oracle cannot be checked without the secret it consults.

The conflict-feedback distinction

A related design decision is whether ordinary play tells the player that a move was wrong. Public-rule-only conflict feedback means a move is rejected only when it violates a stated rule, never because it differs from a hidden answer. That keeps the mistake the player's own and keeps the hint honest, because both operate on the same public information.

What a player experiences

In the reviewed Bit Balance flow, a rule-based Nudge returns an explanation and leaves the move to the player. An answer oracle, by definition, would return information derived from a stored solution. The repository has no player-study data for recognition, learning, or request frequency, so this guide does not infer those outcomes.

Where the line is drawn

A product can expose a rule explanation and a separate reveal action without treating them as the same feature. The useful distinction is what information each action returns: a reason derived from public constraints versus an answer value. This guide describes that boundary rather than prescribing a measured learning strategy.

What the oracle model requires

An oracle requires a stored solution or another authoritative answer source so it can compare the current state with that answer. A public-rule model instead requires the current state plus the published constraints it evaluates. The evidence here supports that dependency difference; it does not claim one implementation is universally simpler.

How a rule hint is computed

A rule hint examines the current board and the published constraints, then identifies a constraint that applies to the current position. It can report that a stated constraint is violated without needing to know the answer, because the violation is a property of the board. The computation is local to the rules, which is why it works identically for a board nobody has solved before.

Why the two action labels should stay distinct

A rule hint and an answer reveal are different information contracts even if a product chooses to expose both. The rule hint derives a reason from public constraints; the reveal depends on an answer source. Keeping those labels distinct prevents documentation or UI copy from describing an answer lookup as though it were a public-rule deduction.

How this interacts with conflict feedback

If ordinary play rejects moves by comparing them to a hidden answer, then the game is already using an oracle and the hint is only one surface of it. Public-rule-only conflict feedback keeps the same information basis for rejection and for assistance, so a player's mistakes remain their own and the help remains explanatory rather than authoritative.

What repeated oracle use changes structurally

Each oracle answer resolves part of the current board using information from the hidden solution rather than from a public-rule deduction. Repeating that action can therefore reduce how much of the current board remains for the player to deduce. This guide makes no claim about whether repeated use changes learning, skill, or future help-seeking behavior.

How a rule hint stays useful across puzzles

A rule explanation describes a constraint that can apply to other positions governed by the same rule, while an oracle answer describes an answer for the current state. The first therefore contains rule-level information and the second contains state-specific answer information. Whether a player later remembers or applies that information is outside the evidence collected here.

What the Bit Balance evidence actually proves

The cited Bit Balance source proves that its Nudge flow explains public deductions without filling a cell and that ordinary conflict feedback uses public rules. The oracle model in this guide is a comparison model, not evidence that Bit Balance or every BeanNest game exposes a separate reveal control.

How the two models differ in what they require

An oracle requires a stored solution for every board it will ever advise on, which means the game must possess answers for boards it has not yet served. A rule model requires only the constraints and the current position, which is the same information any played move already uses. The second requirement is smaller and is satisfied by the rules the game already publishes.

What information each model returns

A rule hint returns information about a constraint and why it applies to the current position. An oracle returns answer information tied to the stored solution for the current state. That difference can be described directly from the two models without inferring what a player will learn from either one.

What coexistence would not prove

Defining both a rule-hint model and an oracle model does not establish that a particular product offers both, nor does it establish which one improves learning or engagement. Those are separate product and outcome claims. This guide uses the oracle only to clarify what information the evidenced Bit Balance Nudge does not consult.

How a player would notice the difference

In a rule-hint flow, the interface can present an explanation and leave the next move unfilled; in an oracle/reveal flow, the interface can present or apply an answer value from the stored solution. That interaction difference is directly observable. Any claim about what the player understands afterward would require separate outcome evidence.

What the guide declines to claim

It does not claim that rule hints improve learning or that any measured outcome follows from them, because this repository has not collected such data. It describes the two models and the reasoning behind the preference, and it stops there.

First-party evidence and provenance

How we checked this

We reviewed the evidence contract and the published rule metadata for a representative logic game to confirm that conflict feedback and hints operate on public constraints rather than on a hidden solution. We did not inspect or publish any stored solution state.

Method: We reviewed this repository's evidence contract and the published game metadata, then traced the shipping behaviour to confirm the mechanism described here. We deliberately did not read or publish a hidden solution state.
Environment: Repository source review of the current main branch. No live traffic, account data, or player-identifying information was read.
Captured: September 20, 2026
Reviewed by: BeanNest Studio

Repository evidence artifacts:

docs/gameplay-reconstruction/EVIDENCE-CONTRACT.md
games/bit-balance/game.json