The risk candidate passed the strong-baseline challenge described in the reviewed research brief.
Passing a development comparison does not itself establish independent confirmation.
Inspect risk baseline challengemeanydeany / Research program
ASRA
AI Systematic Research Architecture
Generate hypotheses. Repair measurement. Confirm what survives.
Research engine
Failure changes the measurement and narrows the claim.
ASRA organizes research around testable questions, explicit comparisons and visible rejection. A promising candidate can fail a control, survive only within a narrow domain, or remain unresolved.
- 01Generate
Candidates and explicit market questions
- 02Measure
Aligned targets, sensors and positive controls
- 03Falsify
Alternatives, costs, nulls and failure criteria
- 04Repair
Measurement and selection when controls fail
- 05Confirm
Robustness and frozen independent assessment
- 06Translate
Separate policy tests and bounded utility claims
The full research sequence
- Idea
- Candidate generation
- Measurement
- Falsification
- Repair
- Robustness
- Independent assessment
- Policy research
Failed controls → return to measurementDiagnose the sensor, target alignment and selection. A failed test can require a repaired measurement before the market claim is interpretable.
Rejected or unresolved → retain the recordUnsupported candidates stop here or remain open questions. This map describes the process; it does not imply every candidate reaches confirmation.
Strongest confirmed finding
Independent Risk Forecast Confirmation
INDEPENDENTRISK INFORMATION CONFIRMED
Context adds short-horizon BTC risk information beyond historical volatility persistence: H+C improves both primary losses against HAR-basis on the frozen independent assessment.
1 hour · BTC risk forecast · H+C vs HAR-basis
Model: H_PLUS_C · Baseline: B4B_HAR_BASIS
Assessment: [2025-08-01, 2026-07-30) · end exclusive
- Log-MSE improvement
- +12.58%H+C vs HAR-basis
Lower forecast loss versus HAR-basis; this is not a return metric.
- QLIKE improvement
- +11.65%H+C vs HAR-basis
Lower forecast loss versus HAR-basis; this is not a return metric.
- Chronological segments positive in both primary losses
- 4 / 4Consistency across the chronological assessment
Consistency across four chronological segments, not four independent replications.
Completed translation test
A forecast and a policy answer different questions.
Better risk forecasts did not automatically become a better risk policy.
Frozen primary P3 · pooled H+C vs HAR policy: 2.070% volatility improvement and 2.016% ES5 improvement, below the 3.0% preregistered minimum.
Only the confirmed-forecast-era diagnostic passed the joint era rule. Required multi-era consistency failed.
Forecast information survived; the preregistered policy-utility claim did not.
FORECAST_INFORMATION_WITHOUT_POLICY_UTILITY
Measurement and development
What the repairs and comparisons established
Positive controls exposed a measurement blind spot; a repaired nonlinear sensor recovered detection within the tested synthetic settings.
Recovery in synthetic controls does not establish historical BTC nonlinear alpha or market predictability.
Inspect nonlinear sensor recoveryNative-Horizon Selection
Measurement repair
Native-horizon target alignment and selection repair recovered methodological power in the supplied research progression.
Do not attribute repaired power specifically to adaptive scheduling. The completed repair's evidence class and test domain remain unresolved.
Inspect native-horizon selectionCompleted synthetic methodology
Native Scheduler Robustness V1
The registered synthetic test found no meaningful adaptive incremental value beyond fixed native-horizon selection, supporting a simpler research architecture within the tested scope.
NATIVE_FIXED_SELECTION_SUFFICIENT
Completed retrospective transfer test
ASRA Candidate Generator V3
The repaired fixed-native Activity and nonlinear challengers did not improve historical one-hour BTC return forecasts over the frozen OHLCV baseline. Zero finalists; the locked return holdout remained unopened.
Scientific progression
The record includes failed claims and simpler methods.
Two research lanes track how findings changed the next question. These are separate studies with distinct validation domains, not one sequential statistical experiment. Synthetic measurement health, historical return transfer, independent risk information and policy utility remain separate claims.
Return / methodology
Measurement repair, synthetic robustness and retrospective historical transfer each retain their own assessment domain.
Directional hypotheses
Directional evidence was weak or cost-fragile.
Narrow the claim; retain qualified historical EMA work separately. No blanket invalidation follows.
Inspect this stage- SYNTHETICINVALIDATED
Synthetic positive-control failure
A synthetic control exposed detection failure.
The interpretation of a null readout as no signal was invalidated: a measurement blind spot could produce the same readout.
Inspect this stage Measurement decomposition
Measurement stages were examined separately.
Isolate sensor detection, target alignment and selection; no separate data domain is asserted for the diagnosis.
Inspect this stage- SYNTHETICCOMPLETED
Nonlinear sensor repair
Synthetic positive-control detection recovered.
Recovery in synthetic controls does not establish historical BTC nonlinear alpha or market predictability.
Inspect this stage - COMPLETED
Native-horizon selection repair
Native-horizon target alignment and selection repair recovered methodological power in the supplied research progression.
Do not attribute repaired power specifically to adaptive scheduling. The completed repair's evidence class and test domain remain unresolved.
Inspect this stage - SYNTHETICCOMPLETED
Scheduler robustness
The registered synthetic test found no meaningful adaptive incremental value beyond fixed native-horizon selection, supporting a simpler research architecture within the tested scope.
Supports simplification within the tested synthetic scope, not exact equivalence, historical BTC alpha or execution authority.
Inspect this stage - SYNTHETICCOMPLETED
Fixed-native architecture sufficient
The completed scheduler study supports simplification within its registered synthetic robustness domain.
This architecture decision does not itself establish historical return transfer.
Inspect this stage V3 historical transfer test
The repaired fixed-native Activity and nonlinear challengers did not improve historical one-hour BTC return forecasts over the frozen OHLCV baseline. Zero finalists; the locked return holdout remained unopened.
The repaired Activity / nonlinear challengers did not add historical return information beyond the frozen OHLCV baseline in the tested domain. This does not establish that all BTC return information is absent.
Inspect this stageNo incremental historical transfer demonstrated
Zero finalists; the locked return-lane holdout remained unopened.
No historical holdout confirmation or universal return-unpredictability claim follows.
Inspect this stage
Next question · Return / methodology
Small-Signal Sensitivity CalibrationNEXT QUESTION only. No study has started and no sensitivity result is reported.
Risk
Risk baseline challenge, independent forecast confirmation and retrospective policy translation use separate claim and assessment contexts.
Risk candidate discovery
A candidate short-horizon risk information claim emerged.
Development discovery needed a demanding baseline comparison.
Inspect this stage- DEVELOPMENTCOMPLETED
Strong baseline challenge
The risk candidate passed the strong-baseline challenge described in the reviewed research brief.
Passing a development comparison does not itself establish independent confirmation.
Inspect this stage - INDEPENDENTCOMPLETED
Independent forecast confirmation
Both primary losses improved; all four chronological segments were positive in both losses.
Risk information, not directional alpha, a trading signal, sizing approval or execution authority.
Inspect this stage Policy translation tested
The completed policy study evaluated the frozen primary P3 mapping.
Both primary pooled risk improvements missed the preregistered minimum.
Inspect this stagePolicy utility not confirmed
Forecast information survived; the preregistered policy-utility claim did not.
Only CONFIRMED_FORECAST_ERA_DIAGNOSTIC passed the joint era rule; required multi-era consistency failed.
Inspect this stageTurnover / opportunity-cost mechanism isolated
H+C retains better risk ranking and lower turnover; much of the relative fee-net return advantage is associated with lower turnover.
These observations do not override the failed P3 rule or establish a completed causal decomposition.
Inspect this stage
Next question · Risk
Turnover & Opportunity-Cost DecompositionNEXT QUESTION only. Mechanism observations belong to the completed policy study.
Current / next
Completed findings, then proposed questions
Current work focuses on event-level BTC market microstructure, passive execution, and low-latency market-data systems. Earlier ASRA, forecasting, and historical strategy studies remain frozen under their original evidence boundaries.
Current direction / Market microstructure
BTC Market Microstructure and HFT
Do observable order-flow and liquidity states change fill probability and post-fill markouts enough to support robust, cost-aware passive-execution payoff?
Planned measurement layer: event-level trades, quotes and order-book updates; local book reconstruction; OFI, microprice, spread, depth and trade imbalance; then 1s, 5s, 10s, 30s and 60s markouts.
Current build / Systems engineering
C++ market-data and deterministic replay core
Build an event-driven core that captures, validates, reconstructs and replays trades and order-book updates with explicit timestamp and sequence discipline.
Raw WebSocket capture → sequence validation → local order book → deterministic replay → microstructure features → fill-probability, adverse-selection and markout analysis in Python.
Evidence protocol
A result carries its context and its boundary.
Evidence class describes the assessment. Work state describes whether a study is completed, underway, proposed, blocked or invalidated. A completed study can leave a specific claim unconfirmed.
Independent assessment is separate from candidate development. It does not imply replication by another institution. Synthetic controls establish what the tested measurement can detect; they do not establish historical market alpha.
- A frozen claim assessed separately from model and candidate development.
- Evidence used during hypothesis, model, baseline or specification development.
- Generated data, positive controls or controlled methodological simulations.
- Historical analysis with its selection and evaluation history disclosed separately.
- Observations accrued after a specified freeze or start under a recorded protocol.
- IN PROGRESS
- The study is underway. No completed result is reported.
- BLOCKED
- A specified objective cannot proceed under a documented binding constraint.
- INVALIDATED
- A particular interpretation lost support after a defined validity check failed.
Source availability
Research record maintained in the private research archive. The reviewed claims are reported here; a public report and manifest are not available.
No verified public repository or immutable code link is available for this study.
Each detail page discloses its own assessment scope, source availability and code lineage. Original thesis and seminar-paper PDFs remain available in the academic archive.
Read the original papersHuman responsibility and separate systems
AI supports candidate generation, implementation and review. meanydeany remains responsible for research questions, evidence standards and published claims.
Existing lab and BTC pages document research infrastructure and architectural lineage. Operational telemetry is read-only systems evidence.
Inspect systems and research infrastructure