When buying goes one-sided, is anyone still trading an hour later?
When buying goes one-sided, is anyone still trading an hour later?
A measurement where buys make up at least 10 points more of the trade count than the median of the preceding six hours is followed, one hour later, by at least one trade more often than a measurement on a token of the same age with no such shift.
token-measurement cohort comparison. Exposure is the trigger condition evaluated on a 6h trailing window; the outcome is 'at least one trade in the following hour' read strictly 1h later — hours of clock time, resolved against the measurement timestamps. Rates are compared pooled and within each token-age band with a two-proportion z-test, then re-checked on a chronological split.
- strata
- "age"
- window_hours
- 6
- horizon_hours
- 1
- min_effect_pp
- 6
- min_window_points
- 4
Difference of +0.0 points against the predicted direction (exposed 100% vs control 100%, 9 vs 163 measurements). The smallest group holds 9 rows, below the 30 needed to judge a difference of this size, so the falsification rule is not applied to a sample that cannot support it.
- z
- null
- strata
- [object Object]
- n_control
- 163
- n_exposed
- 9
- rate_control
- 1
- rate_exposed
- 1
- difference_pp
- 0
- excluded_rows
- [object Object]
- distinct_tokens
- 14
- expected_direction
- 1
- signed_difference_pp
- 0
- first_half_difference
- 0
- second_half_difference
- 0
- sample_too_small_to_judge
- true
- falsification_condition_met
- true
Smallest group is 9 measurements against a 30 minimum. 1016 tokens were excluded for short series and 64 readings had nothing 1h later. Tokens that died early are exactly the ones that run out of series, so the sample leans toward survivors. Exposure is defined by the same detector that surfaced the anomaly, so the hypothesis and its test share a definition. An independent trigger would be stronger.
- version
- critic-checks-v1
- stability
- PASS
- confounding
- PASS
- sample_size
- NEEDS_MORE_DATA
- data_leakage
- PASS
- data_quality
- PASS
- independence
- PASS
- selection_bias
- NEEDS_MORE_DATA
- look_ahead_bias
- PASS
- multiple_testing
- PASS
- survivorship_bias
- NEEDS_MORE_DATA
smallest group has 9 measurements, below the 30 needed to believe a difference of this size
- 04:30:00OBSERVATIONcat: buy share 48% against a 58% baseline
- 04:30:00ANOMALYUNUSUAL_TRANSACTION_PATTERN score=0.12 [buy-share-deviation-v1]
- 04:30:00MEMORY_SEARCH20 related memories retrieved before writing the hypothesis (vector-cosine/pgvector)
- 04:30:00HYPOTHESIS#1 When buying goes one-sided, is anyone still trading an hour later?
- 12:50:33DATASET172 measurements over 14 tokens; hash df78087dfd0b…
- 12:50:33EXPERIMENT#000001 buy-surge-trading-1h
- 12:50:34CRITICNEEDS_MORE_DATA: Smallest group is 9 measurements against a 30 minimum. 1016 tokens were excluded for short series and 64 readings had nothing 1h later. Tokens that died early are exactly the ones that run out of series, so the sample leans toward survivors. Exposure is defined by the same detector that surfaced the anomaly, so the hypothesis and its test share a definition. An independent trigger would be stronger.
- 12:50:34RESULTDifference of +0.0 points against the predicted direction (exposed 100% vs control 100%, 9 vs 163 measurements). The smallest group holds 9 rows, below the 30 needed to judge a difference of this size, so the falsification rule is not applied to a sample that cannot support it.
- 12:50:34MEMORY_UPDATEresult stored: buy-surge-trading-1h: inconclusive