Skip to content

feat(tools): add Options GPS for view-based options decisions#6

Open
bitloi wants to merge 2 commits intoentrius:mainfrom
bitloi:feature/options-gps-issue-2
Open

feat(tools): add Options GPS for view-based options decisions#6
bitloi wants to merge 2 commits intoentrius:mainfrom
bitloi:feature/options-gps-issue-2

Conversation

@bitloi
Copy link

@bitloi bitloi commented Feb 26, 2026

Summary

Options GPS turns a trader's view into one clear options decision. The user provides symbol, market view (bullish / bearish / neutral), and risk tolerance (low / medium / high). The tool uses the Synth API for 1h and 24h prediction percentiles, option pricing, and volatility, then outputs three strategy cards: Best Match, Safer Alternative, and Higher Upside, each with chance of profit, max loss, tail risk, and invalidation/review guidance.

The pipeline: forecast fusion (1h/24h → Aligned/Countermove/Unclear), strategy generation (long call/put, debit/credit spreads, iron condor, butterfly), payoff engine (PoP, EV, tail risk), view-first ranking with guardrails, and a 4-screen CLI (View Setup, Top Plays, Why This Works, If Wrong) plus a JSON decision log.

Technical Document

The 1-page technical document is at tools/options-gps/README.md. It covers:

  • What the tool does and the problem it solves
  • Architecture: forecast fusion, strategy generator, payoff engine, ranking, guardrails
  • Synth API integration: get_prediction_percentiles, get_option_pricing, get_volatility
  • Usage instructions and test commands

Related Issues

Closes #2 — Options GPS: High-Level System Design

Testing

  • Tests pass in mock mode (python -m pytest tools/options-gps/tests/ -v)
  • Manually tested with mock data (BTC, neutral, medium risk — full 4 screens + decision log)
  • Tests added in tools/options-gps/tests/ (fusion, strategies, payoff, ranking, guardrails, confidence, Synth client)

Unit tests: 23 passed. Manual: CLI run (BTC, neutral, medium) — all screens + decision log; see demo video for full run.

Checklist

  • 1-page technical document at tools/options-gps/README.md
  • Tool lives in its own subfolder under tools/
  • Tool uses synth_client.SynthClient for all Synth API access
  • Code follows project style guidelines
  • Self-review completed

Demo

- Forecast fusion (1h/24h) -> Aligned/Countermove/Unclear
- Strategy generator: long call/put, debit spreads, iron condor
- Payoff engine: PoP and EV from Synth percentiles
- Ranking: Best Match, Safer Alternative, Higher Upside
- Guardrails: no-trade on conflict or high vol
- CLI 4 screens: View Setup, Top Plays, Why This Works, If Wrong
- 16 tests (pipeline + Synth client); README per submission requirements
…ardrails

- Strategy families: bull put credit spread, bear call credit spread, long call butterfly
- Tail risk metric (worst 20% avg loss) and loss profile per card
- Forecast confidence from percentile dispersion; no-trade when confidence < 0.25
- Hard guardrails: max-loss cap by risk level, spread-quality checks
- Strategy-specific invalidation triggers and reroute rules
- View-first fit scoring (user view primary, fusion bonus secondary)
- Enforced card semantics: Safer must beat Best on PoP or max_loss; Upside must beat on EV
- Structured decision/audit log (JSON) after Screen 4
- ASCII payoff visualization in Screen 3
- 23 tests, all passing
@e35ventura
Copy link
Collaborator

you're missing demo video

@bitloi
Copy link
Author

bitloi commented Feb 27, 2026

I just added the demo video @e35ventura

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Options GPS: High-Level System Design

2 participants