When events were actually recorded6 × 10MIN · UTC
Each bar counts stored market-state entries during a 10-minute interval. Missing sources are never interpolated.
No observed events available to compare
VETO / ALPHA FACTORY / EXPERIMENT 001
Put hypotheses to work.
Let evidence decide.
One click runs a reproducible BTC/ETH/SOL batch on public hourly perpetual candles. Two frozen protocols, separate validation metrics, and no fake claims of alpha.
AF-01 / MISSION STATUSREADY TO RUN
01
SOURCEPublic historical feed
02
FREEZEPrecommitted models
03
TRAINTrain-only selection
04
HOLDOUTFuture unseen bars
05
REVIEWIndependent human review
Local research only · no API keys · no orders
PUBLIC SOURCE
FIXED ASSETSBTC · ETH · SOL 1H PERPETUAL
The batch has not been run yet.
Run the frozen batch to evaluate two protocols per asset without placing orders.
DISCOVERY 011 · TRAIN / OOSRESEARCH 009 · COSTED WALK-FORWARDQUARANTINED
Execution authority remains closed. Research does not access real accounts or authorize trading.
FINANCIAL FRONTIER / MR-02 · MULTI-ASSET OBSERVATION
Three markets.
One observable record.
Independently read the BTC, ETH and SOL market-state journals. Compare actual recorded transitions and their frequency — never rebrand observation frequency as a price forecast.
READ-ONLY OBSERVATION0 / 3sources with enough recent observations
MR-02 / BTCCONNECTING
What states were recorded?
READY—
OBSERVE—
WAIT—
VETO—
NO DATA—
Median inter-transition interval—
Last recorded state—
This journal records state changes, not every moment in the market. Counts are NOT time-weighted READY/WAIT duration. Trading-edge analysis requires separate prices, execution costs and independent replication.
FINANCIAL FRONTIER / EXPERIMENT MR-03
Coincidence or pattern?
Count same-minute UTC WAIT-state arrivals in BTC, ETH and SOL. Contrast with 12 precommitted circular time shifts — a descriptive negative control, not proof of predictability.
EXPERIMENT STATUSCONNECTINGObservational research only
BTC— WAIT minutes
ETH— WAIT minutes
SOL— WAIT minutes
Observed overlaps vs shifted controls12 FIXED PHASE SHIFTS
Validating sources…
MR-03 / PRECOMMITTED CONTROLInvestigation rules
WINDOW60 × 1 MIN UTC
EVENTWAIT ARRIVAL / BINARY
CONTROLS12 CIRCULAR SHIFTS
MINIMUM5 WAIT MIN / ASSET
SCIENTIFIC AUTHORITYNOT VALIDATED
This compares co-occurring state events, not trading strategies. Shifts preserve within-market bursts but don't establish independence. Excess overlap is neither a p-value, validated alpha, nor a causal effect.
AF-02 / RESEARCH AUTOMATION
Automation must leave evidence.
A six-hourly research job is configured. This timeline shows only actual GitHub Actions runs; scheduling does not prove that any research worker actually executed.
SCHEDULE03:17 · 09:17 · 15:17 · 21:17UTC / GitHub best-effort
COMPLETED JOBS—In recent history; not scientific proof
RUNNER STARTUP FAILURE—Platform startup, not trading outcome
TRADING AUTHORITYDISABLEDRead-only · artifacts only
Checking real workflow runs…
01
Research Protocol
02Discovery 011 maths, train selection, separate OOS check.
Paper Observatory
03Public paper observatory with no auto-handoff from research.
Evidence Health
Inspect recorded observation quality and completeness.
Research outputs are not validated trading alpha; modelled returns are not executed trades or guaranteed gains.