Wait for the bar to close
After a completed five-minute bar. Features use only information available then.
MICRO TRADING LAB / FIVE-MINUTE RESEARCHFast decisions.
Fast decisions.
Slow permission.
A five-minute US-equity challenger built to enter only after a completed bar, pay modeled execution costs, stop at the session boundary, and earn every gate before it can call itself deployable.
- Bar interval
- 5 min Completed core-session bars
- Latency
- 1 bar No same-bar entry
- Holding window
- 1 / 3 bars Same session only
- Evaluated gates
- 0 / 9 Nine pending; all are mandatory
Research boundary: Research and paper trading only. Historical gates cannot approve live deployment without independent forward paper evidence.
FROZEN OPERATING CONTRACT
What the challenger is allowed to do.
“Micro” means five-minute US-equity bars here, not CME Micro E-mini futures. The first release is deliberately narrow so the execution assumptions can be tested rather than hand-waved.
Pay latency and costs
Next open after one full latency bar, with spread, slippage, impact, and a 2× stress test.
Stay small and long-only
Maximum planned notional $1,000; no shorting in v1.
Respect the tape
Planned participation is capped at 1% of decision-time volume.
Never hold overnight
09:30-16:00 America/New_York; extended hours and overnight positions are disabled.
Stop the session
A 0.5% frozen-account daily loss stops later entries.
EVIDENCE LEDGER
Nine gates. No partial approval.
Engineering checks can validate the machinery, but they cannot manufacture market evidence. Until a symbol-specific protocol is frozen and real intraday data is evaluated, every gate remains not evaluated.
| Gate | Requirement | Status |
|---|---|---|
| 01DATA QUALITY | Minimum session/bar history, valid OHLCV, unique ordered offset-aware timestamps, and exact agreement with a source-named calendar whose SHA-256 was frozen in the protocol. | NOT EVALUATED |
| 02NO LOOKAHEAD | Feature prefix invariance and purged temporal boundaries pass. | NOT EVALUATED |
| 03OOS SUPPORT | At least 3 outer folds, 100 executed OOS trades, and 20 traded sessions. | NOT EVALUATED |
| 04NET EXPECTANCY | OOS mean daily return is positive and its 95% stationary-block-bootstrap lower bound is above zero. | NOT EVALUATED |
| 05RISK ADJUSTED RETURN | OOS net daily Sharpe is at least 0.50. | NOT EVALUATED |
| 06DRAWDOWN | OOS maximum drawdown is no worse than -5%. | NOT EVALUATED |
| 07COST STRESS | OOS total return remains positive at 2x modeled execution costs. | NOT EVALUATED |
| 08FOLD STABILITY | At least 60% of eligible outer folds have positive net return. | NOT EVALUATED |
| 09FORWARD PAPER EVIDENCE | At least 30 post-freeze paper sessions and 200 closed paper trades reconciled to a protocol-pinned calendar and a hash-chained stream of every frozen-model decision with a provider-signed pre-entry receipt; full fills must pass execution-bar participation checks, and the 95% block-bootstrap lower bound on mean daily return and total return under a 2x decision/execution-bar modeled-cost floor must both be positive. | NOT EVALUATED |
DEPLOYMENT LADDER
What is done, and what is still missing.
- 01Protocol frozen
Decision clock, targets, execution costs, risk controls, and all nine gates are written down.
- 02Engineering validation
9 of 9 focused pipeline tests pass.
- 03Real intraday fit
Pending a licensed bar source, authoritative exchange calendar, frozen symbol, venue, and evaluation window.
- 04Independent forward paper record
30 post-freeze sessions and 200 closed paper trades are required with signed pre-entry receipts.
- 05Deployment permission
Locked unless every gate passes. A backtest can never unlock this stage by itself.
This page does not hide a direction, position size, or model score. Those outputs do not exist yet because the protocol is intentionally unfitted. When a real release exists, the public status artifact will be promoted only after its protocol hash and all nine gate results reconcile; any inconsistency falls back to NO-GO.