Ψ_T Evidence Observatory · Finance Layer v2.6

Card 01 — BTCUSDT Drawdown Candidate

This evidence card records a BTCUSDT-specific drawdown candidate trace. The current status is robust exploratory evidence. The card supports candidate-level evidence intake for Cycle X, but it does not represent final proof, causal proof, trading-performance validation, or product validation.

PASS robust_exploratory_evidence BTCUSDT only Drawdown-oriented trace Not final proof

Core Finding

BTCUSDT alert windows separated from calm-regime baseline.

BTCUSDT alert windows separated from calm-regime baseline primarily through larger forward drawdown magnitude over 48–72 hour windows. The evidence was robust across tested sensitivity settings, but remains exploratory because the event sample is still limited.

Interpretation boundary

This card supports a BTCUSDT drawdown-oriented candidate trace at exploratory level. It should not be generalized to other assets or treated as production-grade proof without further replication.

Status Summary

Evidence classification for this card.

Status

PASS

Card-level status from the v2.6 finance-layer audit.

Evidence level

robust_exploratory_evidence

Strong enough for candidate evidence intake, not final proof.

Cycle X role

candidate_a input

May be used as source-backed candidate input with limitations preserved.

Corrected Alert-vs-Calm Separation

Event windows outperformed calm baseline.

Corrected alert-vs-calm separation shows that BTCUSDT event windows carried stronger drawdown-success behavior than calm-regime baseline.

Event success

85.37%

35 / 41 event windows.

Matched pressure success

77.24%

95 / 123 matched pressure windows.

Calm success

67.48%

83 / 123 calm windows.

Event over calm lift

1.2651

Event success divided by calm success.

Event minus calm

17.89%

Absolute event-success advantage over calm baseline.

p(event vs calm)

0.027257

Corrected event-vs-calm separation test.

Odds ratio

2.6490

Event odds relative to calm baseline.

Bootstrap P(lift > 1)

99.22%

Bootstrap probability of positive lift.

Bootstrap lift band

1.0549 – 1.5000

Bootstrap lift interval from the evidence card.

Drawdown Mechanism Support

Event windows carried larger forward drawdown magnitude.

Event mean drawdown

0.0392

Matched pressure mean drawdown

0.0282

Calm mean drawdown

0.0238

Event median drawdown

0.0242

Calm median drawdown

0.0166

Mean magnitude lift

1.6498

Event over calm mean drawdown magnitude lift.

Mean magnitude difference

0.0154

Event minus calm mean drawdown magnitude.

Permutation p

0.000800

Permutation test for mean difference.

Robustness / Sensitivity

The candidate survived tested sensitivity settings.

Valid sensitivity rows

Positive rows

27 / 27

100.00% positive event-over-calm rows.

Strong rows

18 / 27

66.67% strong event-over-calm rows.

Primary event mean drawdown

0.0453

Primary calm mean drawdown

0.0290

Primary mean lift

1.5627

Primary permutation p

0.002999

Primary bootstrap P(lift > 1)

99.10%

Primary bootstrap lift band

1.0870 – 2.1576

Evidence Chain

Audit chain leading to v2.6 card status.

The card is not based on a single isolated check. It follows a staged evidence chain from isolation audit to robustness/sensitivity audit.

Version	Role	Finding	Interpretation
v2.2	Isolation audit	`btc_alert_over_calm_candidate`	BTCUSDT alert windows first separated from calm-regime baseline in OHLCV-compatible success rate.
v2.3	Hit-level calm patch	`btc_alert_over_calm_candidate_confirmed_after_hit_patch`	Corrected calm hit-column mapping confirmed the candidate was not a mapping artifact; drawdown_hit carried the separation.
v2.4	Drawdown mechanism audit	`btc_drawdown_mechanism_candidate_supported`	Event windows showed higher mean and median drawdown magnitude and heavier tails than calm baseline.
v2.5	Robustness / sensitivity audit	`btc_drawdown_candidate_robust_exploratory_evidence`	The event-over-calm mean drawdown lift stayed positive in all sensitivity rows and strong in two-thirds of rows.

Limitations

Exploratory evidence boundaries.

This card must preserve its limitations. It can support candidate-level evidence intake, but it should not be interpreted as final proof.

Scope limitations

Event sample remains limited at 41 BTCUSDT alert windows.
Evidence is asset-specific to BTCUSDT.
It should not be generalized to ETH, SOL, SP500, or XAU without separate cards.
The strongest expression appears in 48–72 hour windows.
24 hour windows are weaker and less stable.

Data and interpretation limitations

The card is based on OHLCV-compatible drawdown evidence.
Crypto funding-rate and open-interest pressure were unavailable in this test chain.
The status is robust exploratory evidence, not production-grade causal proof.
This card does not validate trading performance or product performance.

Cycle X Intake Mapping

How this card maps into cycle_x_0001.

This card can be used as the first actual structured input for candidate_a. Because limitations remain, the safest attachment outcome is source_backed_with_limitations.

{
              "candidate_id": "candidate_a",
              "assigned_source_id": "src_0001",
              "assigned_group": "group_a",
              "attachment_id": "attach_candidate_a_btcusdt_v2_6_0001",
              "attachment_type": "manual_structured_research_note",
              "structured_audit_output_reference": "BTCUSDT Drawdown Candidate Evidence Card v2.6",
              "source_reference": "psi_t_v2_6_btc_evidence_card_gpt.md",
              "internal_trace_reference": "finance_layer_v2_6_btcusdt_drawdown_candidate",
              "attachment_outcome_recommendation": "source_backed_with_limitations",
              "limitations_to_preserve": [
                "limited_event_sample_41_alert_windows",
                "asset_specific_btcusdt_only",
                "strongest_expression_48_72h_windows",
                "funding_open_interest_unavailable",
                "paired_candidate_b_not_attached_yet",
                "not_final_proof"
              ],
              "next_required_step": "Attach candidate_b comparable structured input before src_0001 / src_0002 population."
            }

Decision

BTCUSDT can be labeled robust exploratory evidence.

BTCUSDT can now be labeled robust exploratory evidence for a drawdown-oriented candidate trace. This does not mean final proof. It means the BTCUSDT candidate survived isolation, hit-level repair, drawdown mechanism audit, and robustness/sensitivity stress tests.

Recommended next step

v2.7 — BTC Forward Replication / New-Data Holdout Plan. The next research step is to pre-register BTC drawdown criteria and test them on future unseen alert windows.

Claim Boundary

This card is not a proof claim.

This evidence card is a bounded exploratory research artifact. It does not prove Ψ_T, does not establish causal certainty, does not validate predictions as a product claim, and does not validate trading performance. Its correct role is candidate-level evidence intake and forward-replication preparation.