ASRA · Measurement methodology
A null result can reflect an absent signal, a blind sensor or a misaligned test. Positive controls made those possibilities separable.
The progression moved from detection failure to sensor repair, then to native-horizon target alignment and selection repair. Synthetic controls remained healthy; the completed V3 test did not demonstrate incremental historical return transfer within its frozen validation domain.
Failure and diagnosis
A synthetic positive control exposed detection failure. The null readout could no longer support an interpretation of “no signal” without first ruling out a measurement blind spot.
That invalidated a specific interpretation of the failed test. It did not invalidate every directional hypothesis or establish a market signal.
Separating these stages changes what a failure means and which part of the research design needs repair.
Completed synthetic repair
Can the measurement system detect a controlled nonlinear signal?
Use positive controls to distinguish detection failure from absence of signal, then repair the sensor.
Synthetic positive controls · Failed sensor/control setup vs repaired setup
Assessment: Tested synthetic settings; exact experiment version not supplied.
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.
Detection recovery is limited to the documented test settings. No real-market extension is established by this record.
Does target alignment and selection repair recover power beyond the sensor repair?
Completed measurement repair
Measurement repair
Does target alignment and selection repair recover detection power?
Align targets to their native horizons and repair selection separately from sensor detection.
Target alignment and selection repair; test domain not supplied · Previous vs repaired target and selection architecture
Assessment: Test domain, experiment version and evidence class await source review.
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.
The separate synthetic scheduler study does not establish the data domain of this completed repair.
The reviewed progression supports power recovery after target and selection repair. It does not isolate an adaptive scheduler's contribution. The separate completed scheduler test now narrows that architecture decision.
Completed synthetic assessment
NATIVE_FIXED_SELECTION_SUFFICIENT
Does adaptive question ordering add meaningful value beyond simpler fixed native-horizon selection?
Registered synthetic comparison of adaptive scheduling, fixed native-horizon selection and the Forced reference, with reported raw and adjusted results, paired uncertainty and validity guards.
Synthetic methodology research · primary 60bps setting · S1 adaptive scheduler vs S2 fixed native-horizon selection; Forced reference
Model: S1 · Baseline: S2
Assessment: Completed registered synthetic experiment; calendar assessment dates not supplied.
Primary 60bps synthetic setting. The conservative paired interval excludes the meaningful positive-increment margin; this is not proof of exact equivalence.
The upper bound of the conservative paired interval is below this meaningful adaptive-increment margin.
Raw and adjusted detection retention relative to the Forced reference in the primary 60bps setting.
Registered trials with reported raw and adjusted detection, a conservative paired 95% interval, a prespecified meaningful-increment margin, null guards and presentation-order checks. The adjustment method is not supplied in the reviewed update.
| Measure | Observed |
|---|---|
| Conservative paired 95% interval | [−2.47, +1.69] pp |
| Registered trials | 1,536 |
| Trials per cell | 256 |
| Forced detection · raw and adjusted | 242 / 256 |
| S1 adaptive detection · raw | 237 / 256 |
| S1 adaptive detection · adjusted | 236 / 256 |
| S2 fixed selection · raw and adjusted | 237 / 256 |
| Undefined primary tests | 0 |
| Scientific presentation-order effects | 0 |
All null guards passed.
The registered synthetic test found no meaningful adaptive incremental value beyond fixed native-horizon selection, supporting a simpler research architecture within the tested scope.
The paired interval excludes the prespecified meaningful-increment margin. The result supports simplification within the registered synthetic scope. It does not establish exact equivalence or make this methodology test a prerequisite for the earlier independent risk assessment.
Completed retrospective validation
MEASUREMENT_REPAIR_DOES_NOT_TRANSFER_TO_HISTORICAL_BTC
Does the repaired fixed-native measurement stack add historical one-hour BTC return information beyond the frozen OHLCV baseline?
Frozen native 1h objective: P0 OHLCV_RIDGE → P1 ACTIVITY_RIDGE → P2 NONLINEAR_OHLCV. P2 uses the R3 completed return_1h × return_6h product and ridge alpha 1, with canonical tie handling and adaptive ordering diagnostic only.
Frozen fixed-native V3 validation with registered paired bootstrap and Holm gates. The return-lane holdout was not opened; this validation is not independent holdout confirmation.
Native 1 hour · Historical BTC return validation · P1 ACTIVITY_RIDGE and P2 NONLINEAR_OHLCV vs P0 OHLCV_RIDGE
Model: P1 / P2 · Baseline: P0
Assessment: [2024-01-01, 2025-07-30) · end exclusive · retrospective validation
ACTIVITY_RIDGE did not lower pooled historical return forecast MSE versus the frozen OHLCV_RIDGE baseline.
NONLINEAR_OHLCV did not lower pooled historical return forecast MSE versus the frozen OHLCV_RIDGE baseline.
Both challengers failed the registered paired bootstrap and Holm gates. The locked return-lane holdout remained unopened.
| Measure | Observed |
|---|---|
| Common native hourly validation targets | 13,819 |
| P1 positive MSE folds | 3 / 5 |
| P2 positive MSE folds | 2 / 5 |
| P0 pooled rank IC | ≈0.050895 |
The baseline diagnostic was not identically zero. Failure of incremental challenger transfer is not evidence that every BTC return predictor contains no information.
Both challengers failed the registered paired bootstrap and Holm gates.
Zero finalists. The locked return-lane holdout was not opened.
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.
Secondary shadow economics did not rescue the failed predictive gate. They did not determine candidate selection or finalist ranking, did not open the holdout, and do not establish policy utility; market exposure was high.
Next question · Return / methodology
Next methodology question
How small a conditional return effect can the repaired research stack reliably detect under BTC-like historical feature distribution and temporal dependence?
NEXT QUESTION only. No study has started and no sensitivity result is reported.
Inspect small-signal sensitivityNEXT QUESTION only. The study has not started; no design, threshold, result or power value is supplied.
Source availability
Research records are maintained in the private research archive. The sensor repair is explicitly synthetic. The native-horizon repair's exact test domain and evidence class await source review, so it carries a neutral measurement-repair role without an inferred scientific badge.
Public experiment artifacts, generated-data settings and immutable code links are not available. The completed scheduler test supplies its own synthetic context and verified study identifier; its evidence class does not transfer to the earlier native-horizon repair.
Candidate Generator V3 is a separate retrospective historical validation. Its completed-study identifier and validation interval do not establish an independent assessment or access to the locked return-lane holdout.
Research record maintained in the private research archive. The reviewed claims are reported here; a public report and manifest are not available.
Code: No verified public repository or immutable code link is available for this study.
Assessment scope: Tested synthetic settings; exact experiment version not supplied.
Research record maintained in the private research archive. The reviewed claims are reported here; a public report and manifest are not available.
Code: No verified public repository or immutable code link is available for this study.
Assessment scope: Test domain, experiment version and evidence class await source review.
Research record maintained in the private research archive. The reviewed claims are reported here; a public report and manifest are not available.
Code: Verified completed-study identifier supplied with the reviewed result. No public code destination is available.
Completed study identifier: b45e2a3fa2daf4b51967059d964393536a41f106
Assessment scope: Completed registered synthetic experiment; calendar assessment dates not supplied.
Research record maintained in the private research archive. The reviewed claims are reported here; a public report and manifest are not available.
Code: Verified completed-study identifier supplied with the reviewed result. No public code destination is available.
Completed study identifier: 1dc9e529b9729e770691e36a2ad937c0009f6fdb
Assessment scope: [2024-01-01, 2025-07-30) · end exclusive · retrospective validation