Chronological comparison
The research pipeline separates training, validation, and held-out periods. Duplicate records and outcomes unavailable at each cutoff are excluded.
SANCIRE LABS
Sancire Labs / WorkSignal Research
From observation to tested signal.
An experimental system that records market observations, replays them against history, and tests whether a signal survives costs, liquidity, and new data.
Examine the system01 / System
SL / SIG / 02A pattern has to survive more than a backtest.
The system keeps observation, historical replay, execution assumptions, and forward evidence separate. A promising chart is a question to test, not an answer.
Stage 01 / 04
Freeze what was visible.
Capture the candidate, its features, source, and entry-time context before the eventual outcome is known.
OutputA time-stamped observation without hindsight
02 / Engineering
Reviewed 14 September 2026Separate the pattern from the evidence.
Explicit accounting
Paper exits track partial realizations, costs, stale prices, and liquidity assumptions. Results from older accounting versions cannot qualify the corrected model.
Frozen experiments
Registered rules retain their identity and start time. Future evaluation counts unresolved cases as well as closed outcomes, so quick winners cannot hide slower failures.
03 / Current boundary
As built todayThe experiment stays an experiment.
Current work is designed to improve the quality of evidence, not to manufacture certainty or imply financial performance that has not been demonstrated.
- 01
Results are paper-only. The system has not established real order execution, cash-out capacity, or live-trading readiness.
- 02
The latest retained assessment does not establish a profitable edge. New qualifying outcomes under the corrected accounting model are still required.
- 03
No live deployment, wallet authority, automated trading permission, or guarantee of future performance is implied.
From Sancire Labs
