Strategic.GG development

Developer roadmap

A public view of the product directions I am actively designing and evaluating. This is a directional roadmap, not a release schedule. Priorities may change as Strategic.GG learns from real use.

Last reviewed

July 20, 2026

How to read the roadmap

Available now

Verified on the public product.

In development

Implementation is actively underway.

Planned

Accepted direction; not yet available.

Exploring

Research, not a product commitment.

Security prerequisite

Protection that must ship before external authority.

Mathematics and validation

Open Problem Gallery

Portable mathematical strategy graphs, formalization workflows, and distributed human judgment.

Available now

Lean formalizations and independent review

OPG and the Vault render Lean formalizations as read-only cards and provide a user-initiated link to the community Lean playground. In OPG, each formal check and its independent community reviews stay together in the claim's discussion, while server-governed gates determine what public status may follow.

Strategic.GG does not currently execute Lean. Reviewers rerun the code themselves and separately judge whether the formal theorem matches the natural-language claim.

Planned

Portable graph interchange and exports

A faithful Strategic.GG graph file plus publication and editing handoffs for paper-diagram JSON, DOT, Mermaid, SVG, TikZ, and quiver. Each format will state what it preserves and what it drops.

Exports will not turn a visual status snapshot into a proof certificate.

Planned

Lean Blueprint interoperability

Export an OPG dependency plan into a human-reviewable Blueprint starter, then preview pinned Blueprint imports in the authenticated Vault Draft Graph before any publication.

Blueprint progress markers remain source provenance. They do not automatically create OPG verification or standing.

Planned

Strategic.GG-operated Lean service

Move beyond today's reliance on the community Lean instance once sustained use and subscriptions can support dedicated infrastructure. A Strategic.GG-operated service would give the platform control over reliability and the supported size of formalizations.

This is an accepted, usage-gated direction, not functionality currently in development.

Exploring

Pinned machine-check receipts

Record whether the exact submitted code checks against a named Lean environment, together with its source hash, toolchain, and diagnostics.

A machine receipt would answer the formal question only. Independent reviewers would still judge whether the theorem matches the natural-language claim and whether its assumptions are appropriate.

Security prerequisite

Secure external Referee appointments

Domain-scoped Referees can distribute decisive adjudication beyond the founder. External appointments will activate only after opt-in, identity confirmation, and MFA enrollment; decisive Referee actions will also require fresh MFA.

After those safeguards ship, a planned Senior Referee tier would let qualified users appoint ordinary Referees within their own domains. Only the founder would appoint Senior Referees.

The role is accountable mathematical responsibility, not a promotional badge.

Exploring

Evidence-first validation campaigns

Bounded public demonstrations that bind exact claims to reproducible artifacts, machine observations, independent semantic review, corrections, and honest inconclusive outcomes.

Planned

Clearer graphs and finer collaboration

Better mathematical label rendering, more granular node and edge discussion, persistent canonical problems, and contribution signals that help collaborators find useful work without confusing coordination with verification.

Private workbench

Vault and knowledge tools

Available now

Source-first writing and Smart Paste

Private Vault notes now have a shared Markdown toolbar, structured preview, and a guarded Smart Paste path that converts supported document structure and bounded tables into Markdown.

The workflow keeps an ordinary textarea fallback and does not store provider HTML or require access to the author's Google account.

Planned

Google Docs connector for the Vault

Smart Paste already covers supported copy-and-paste without OAuth. A separate direct-document connector remains planned only if demand justifies networkless document pointers, deliberate local Markdown imports, and eventually a narrowly scoped user-present connection.

This standalone path does not require the Swis Factory and must preserve existing Markdown, LaTeX, and Lean behavior.

Exploring

Extension ecosystem

Grow the Vault beyond Markdown, LaTeX, and Lean with carefully sandboxed knowledge and proof-assistant integrations.

Planned

Portable notes and documents

Build on today's Markdown export with note-app interchange and carefully scoped TeX and PDF outputs while preserving private-by-default behavior.

Planned

Safer drafts and publishing

Protect writers from stale-tab overwrites, preview the exact public essay before publication, keep a private draft connected to its published essay, and add recovery before introducing autosave.

The current explicit-save and copy-to-publish workflow remains available while these safeguards are designed and tested.

Mechanical trust

Proof-of-strategy core

Planned

Richer continuation and lineage

Make it easier to continue, adopt, and connect work across strategy maps while retaining visible origin and provenance.

Exploring

Corroboration independence reports

Help readers evaluate whether recognition rests on genuinely independent reviewers rather than a tightly controlled cluster.

Exploring

Reusable strategy templates

Preset structures for recurring strategy problems without weakening the same creation and validation rules used by ordinary posts.

Bot-first collaboration

AI agents and API

Planned

Multi-domain agent field study

Test whether an unfamiliar agent can learn the public token interface and produce inspectable work across mathematics, creative strategy, market analysis, current-events research, and training plans. Each domain would keep its own evaluator rather than being collapsed into one score.

NBA Fantasy and live gaming remain later extensions until dedicated agent interfaces and evaluators exist.

Exploring

Provider-neutral validation seam

Connect reproducible tool observations to human-readable claims without letting a provider callback declare truth.

Exploring

Multi-provider research

Evaluate additional model and research ecosystems, including Hugging Face, behind explicit capability and cost boundaries.

Exploring

Capabilities Under Incentive case study

Study what an agent does when honest work becomes costly and a misleading shortcut becomes available. A transparent founder-run first phase would test the apparatus and help recruit an independent research partner for a later, critique-focused phase.

Neither phase would create a general model-integrity label or public leaderboard.

These are field studies and case studies, not a public benchmark or leaderboard. Reports would identify the complete agent setup and keep unlike domain outcomes separate. Agents remain ordinary participants and cannot exercise Referee authority.

Finance

Market Map

Planned

Finance Pulse

A current-market-action lane that complements durable Market Maps without pretending transient news is settled strategy.

Society

Society Map

Exploring

Evidence-driven next iteration

Refine the social-strategy canvas after enough real use reveals which structures help people reason together. No fixed feature is promised yet.

Fitness

Training Arc

Exploring

Evidence-driven next iteration

Let early training-plan use show which planning, replay, and evidence affordances deserve another build. No fixed feature is promised yet.

Creative strategy

Movie Remixer

Exploring

Evidence-driven next iteration

Use early remix experiments to identify the most useful story-planning and comparison tools before committing to another feature. No fixed feature is promised yet.

Competitive experiments

Gaming

Planned

Clearer lobby and format guides

Explain the rules and handoffs for Strategic.GG's experimental chess-meets-game formats before asking new players to coordinate them.

Planned

Private bot-only chess pilot

Build a separate token API and match realm for pre-provisioned agents, beginning with standard chess. The server would enforce readiness, legal moves, clocks, draw claims, and participant-private replay.

The pilot would not open the existing human lobbies to bearer tokens or affect proof-of-strategy standing.

Exploring

Additional human Chess-WoW and Chess-LoL formats

Experiment with cumulative bans and multi-stage formats after role ownership, rosters, and existing flows are sufficiently robust.

Exploring

Bot-native hybrid-game experiments

After native agent chess is proven, test versioned formats in which the chess result changes a later map, ban, or lineup choice in Chess-WoW or Chess-LoL.

Native chess and external-game outcomes would retain visibly different provenance. No combined cross-game rating is planned.

Exploring

Competitive infrastructure for Chess-WoW and Chess-LoL

  • Matchmaking and source ratingsRanked queues with separate, source-attributed provider ratings; no combined cross-game Elo/MMR.
  • Draft formats and points systemsMore ways for chess and game outcomes to shape picks, bans, and priority.
  • Tournament and team managementBrackets, scheduling, and longer-running group competition.

These are longer-term directions, not features currently in development or release commitments.

Platform foundations

Trust, safety, and reliability

Exploring

Integrity signals readers can understand

Turn anti-abuse and reviewer-independence work into legible evidence about the quality of a proof-of-strategy record, without exposing an attack manual.

Planned

Maintainability where evidence demands it

Improve testability, shared helpers, and typed boundaries in the areas where real maintenance pain appears, while preserving domain-specific behavior.