Change market, time or detector
All times are UTC. Chart history only changes the display. Historical comparisons use the 30 days before the selected start and measure their own outcome horizon.
AIXBTUSDT fell into Sep 16
−11%
50 historical matches for the proposed conditions. This is an exploratory investigation. The combined trendline, compression and microstructure hypothesis still needs validation.
Explore the move and its lead-up
Hourly candles. The marked lines are the minutes the readings came from: 1 minute, 15 minutes, 1 hour, 4 hours and 24 hours before the selected start. These five samples test the lead-up only, not the full move. Use the minute inspector to explore during or after the move.
Explore historical state
Current-minute reads are free after sign-in. Pro and Research include historical-minute inspection and comparison with an hour earlier.
Past readings use stored inputs where available. Recent minutes may use the live buffer and are labelled unpinned. Missing minutes remain visible.
Read the available analytical inputs at one minute, including measurements over the preceding windows. Compare with an hour earlier, then take a reading into Research to test it. Dates are UTC.
These additional reads are separate from the five pre-start samples. Observations during or after the selected move describe what had already happened; they are not advance evidence.
Selected time: 2026-09-13 18:00 UTC. During the selected move: hindsight. Choose Read this minute to request it. Missing minutes remain unavailable.
Follow positioning and flow through the move
Read 23 sampled minutes from the selected start through one hour after the selected end. The peak hour has extra 15-minute samples. The end marks the peak hour, not the exact peak tick. These are separate snapshots, not a continuous tape or an advance-warning test.
Historical access applies. Reads preserve each source revision; missing values stay unavailable. Contract OI, turnover and direct spot basis are not supplied by this table. Published premium is a different measure from direct basis.
No additional reads have run.
The OI score measures absolute USD-notional change, not signed contract growth. Zero 15-second liquidation intensity can coexist with substantial hourly liquidations. Binance liquidation amounts are observed publications, not complete venue totals.
Pre-start evidence: five separate snapshots
Geometry and microstructure, on the same clock
Read left to right: 24h → 4h → 1h → 15m → 1m before the selected start. These are five snapshots, not a continuous tape. Neutral readings stay visible; unavailable values are never zero.
Descending resistance
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| Time since resistance breakhours · Highlight: 12 to 720 (inclusive) | 184 | 204 | Read failed | Read failed | 208 |
| Resistance line spandays | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
| Resistance touchescount | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
| Touches before resistance breakcount · Highlight: ≥ 3 | 6 | 6 | Read failed | Read failed | 6 |
| Line ATR contractionmultiple | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
| Resistance line statelabel | none | none | Read failed | Read failed | none |
| Distance to broken resistancemultiple · Highlight: -0.5 to 0.5 (inclusive) | -0.1315 | -0.1664 | Read failed | Read failed | -0.2728 |
| Distance to resistancemultiple | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
| Support shapelabel | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
Candle compression
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| last 15-minute candle range versus its baselinemultiple · Highlight: ≤ 0.5 or ≥ 2 | 1.135 | 2.156 | Read failed | Read failed | 0.4174 |
| last four-hour candle range versus its baselinemultiple | 0.4505 | 1.053 | Read failed | Read failed | 1.183 |
| last one-hour candle range versus its baselinemultiple | 1.229 | 1.945 | Read failed | Read failed | 0.6061 |
Order book
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| Order-book imbalanceratio · Highlight: ≤ -0.5 or ≥ 0.5 | -0.4365 | 0.3813 | Read failed | Read failed | 0.6514 |
| normalized spreadratio | 0.09782 | 0.1958 | Read failed | Read failed | 0.09711 |
| order-book imbalance rankpercentile | 0.2735 | 0.0622 | Read failed | Read failed | 0.7349 |
Trade flow
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| Informed-trading pressureratio | 0.3412 | 0.3846 | Read failed | Read failed | 0.3467 |
| Net taker flow over the trailing four hoursratio | 0.05241 | 0.03542 | Read failed | Read failed | -0.05519 |
| Net taker flow over the trailing hourratio | 0.03391 | 0.1826 | Read failed | Read failed | 0.2157 |
| large-trade share of recent flowratio | 0 | 0 | Read failed | Read failed | 0 |
| taker-buy share of recent flowratio · Highlight: ≤ 0.35 or ≥ 0.65 | 0.5917 | 0.6798 | Read failed | Read failed | 0.741 |
| VPIN regimelabel | normal | normal | Read failed | Read failed | normal |
| Turnover against its recent hoursratio | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
Funding and positioning
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| Funding minus premium indexsigned_fraction | 0.00005 | 0.00005 | Read failed | Read failed | 0.00005 |
| Minutes until funding settlesminutes | 360 | 120 | Read failed | Read failed | 361 |
| Premium index change over 1hsigned_fraction | 0 | 0 | Read failed | Read failed | 0 |
| Premium index rank over 30dpercentile | 0.6597 | 0.6549 | Read failed | Read failed | 0.6535 |
| Published premium indexsigned_fraction | 0 | 0 | Read failed | Read failed | 0 |
| funding raterate | 0.00005 | 0.00005 | Read failed | Read failed | 0.00005 |
| global long ratioratio | 0.6401 | 0.6408 | Read failed | Read failed | 0.6404 |
| hourly change in open contractssigned_fraction | 0.0008598 | 0.004706 | Read failed | Read failed | 0.0008954 |
| normalized funding rateratio | 0.01667 | 0.01667 | Read failed | Read failed | 0.01667 |
| open-interest velocity rankpercentile | 0.1952 | 0.1645 | Read failed | Read failed | 0.1493 |
| top-trader long ratioratio | 0.726 | 0.7246 | Read failed | Read failed | 0.7286 |
| top-versus-crowd long skewratio_difference | 0.0859 | 0.0838 | Read failed | Read failed | 0.0882 |
Liquidations
| Reading | -24h | -4h | -1h | -15m | -1m |
|---|---|---|---|---|---|
| Cascade clusteringratio · Highlight: ≥ 0.5 | 0.5285 | 0.529 | Read failed | Read failed | 0.529 |
| Liquidation intensityratio | 0 | 0 | Read failed | Read failed | 0 |
| Liquidations in the trailing hourusd | 0 | 0 | Read failed | Read failed | 0 |
| Longs liquidated in the trailing hourusd | 0 | 0 | Read failed | Read failed | 0 |
| Shorts liquidated in the trailing hourusd | 0 | 0 | Read failed | Read failed | 0 |
| Which side the trailing hour of liquidations fell onratio | Unavailable | Unavailable | Read failed | Read failed | Unavailable |
Highlighted cells meet the printed fixed rule. This does not establish rarity, statistical significance or a cause of the move. Book readings describe the best quotes, not full-depth liquidity; liquidation readings describe what the venue reported.
Highlighted readings and exact search conditions
Highlighted readings meet fixed rules, not a statistical-significance test. The observed value and proposed search condition are different: a condition rounds a value or selects an interval for a repeatable search.
Could not read 15 minutes and 1 hour before the move. The engine did not answer for those minutes, so nothing is claimed about them. Reload to try again.
- -24h2026-09-12 18:00 UTC
At 24 hours before the move, 9 of the 65 readings met a highlight rule.
Compression · 4h ≤ 0.46Sweep + reclaim · 15m bullishSweep + reclaim · 1h bullishCascade clustering ≥ 0.52the distance from price back to the broken descending resistance -0.5 to 0.5how many times the broken descending resistance had been tested ≥ 3how long ago the descending resistance broke 12 to 720how many times the broken short descending resistance had been tested ≥ 3how long ago the short descending resistance broke 12 to 720Show the full readings
READING BEFORE THE MOVE SEARCHES FOR Compression · 4h feature.candle_compression_4h 0.4505 ≤ 0.46 Sweep + reclaim · 15m feature.candle_sweep_reclaim_15m bullish bullish Sweep + reclaim · 1h feature.candle_sweep_reclaim_1h bullish bullish Cascade clustering feature.cascade_clustering_index 0.5285 ≥ 0.52 the distance from price back to the broken descending resistance feature.desc_resistance_broken_dist_atr -0.1315 -0.5 to 0.5 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 6 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 184 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 4 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 332 12 to 720 - -4h2026-09-13 14:00 UTC
At 4 hours before the move, 9 of the 65 readings met a highlight rule.
Book imbalance percentile ≤ 0.063Compression · 15m ≥ 2.1Cascade clustering ≥ 0.52the distance from price back to the broken descending resistance -0.5 to 0.5how many times the broken descending resistance had been tested ≥ 3how long ago the descending resistance broke 12 to 720how many times the broken short descending resistance had been tested ≥ 3how long ago the short descending resistance broke 12 to 720Taker buy share · 15m ≥ 0.67Show the full readings
READING BEFORE THE MOVE SEARCHES FOR Book imbalance percentile feature.book_imbalance_pctrank 0.0622 ≤ 0.063 Compression · 15m feature.candle_compression_15m 2.156 ≥ 2.1 Cascade clustering feature.cascade_clustering_index 0.529 ≥ 0.52 the distance from price back to the broken descending resistance feature.desc_resistance_broken_dist_atr -0.1664 -0.5 to 0.5 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 6 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 204 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 4 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 352 12 to 720 Taker buy share · 15m feature.taker_buy_ratio_15m 0.6798 ≥ 0.67 - -1m2026-09-13 17:59 UTC
At 1 minute before the move, 9 of the 65 readings met a highlight rule.
Book imbalance · top levels ≥ 0.65Compression · 15m ≤ 0.42Cascade clustering ≥ 0.52the distance from price back to the broken descending resistance -0.5 to 0.5how many times the broken descending resistance had been tested ≥ 3how long ago the descending resistance broke 12 to 720how many times the broken short descending resistance had been tested ≥ 3how long ago the short descending resistance broke 12 to 720Taker buy share · 15m ≥ 0.74Show the full readings
READING BEFORE THE MOVE SEARCHES FOR Book imbalance · top levels feature.book_imbalance 0.6514 ≥ 0.65 Compression · 15m feature.candle_compression_15m 0.4174 ≤ 0.42 Cascade clustering feature.cascade_clustering_index 0.529 ≥ 0.52 the distance from price back to the broken descending resistance feature.desc_resistance_broken_dist_atr -0.2728 -0.5 to 0.5 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 6 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 208 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 4 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 356 12 to 720 Taker buy share · 15m feature.taker_buy_ratio_15m 0.741 ≥ 0.74
Historical comparison
Exact setup countedBook imbalance · top levels ≥ 0.65 and Compression · 15m ≤ 0.42 and Cascade clustering ≥ 0.52 · Read at 1 minute before the move.These are the printed conditions only. The selected detector family does not change this automatically proposed combination. Edit it before testing the full recipe.
The selected move lasts 71 hours. This comparison measures 72 hours after each historical match. Changing it uses the existing result.
Descriptive comparison; an edge has not been established.
| Outcome | Matching setup | Same-market reference | Difference |
|---|---|---|---|
| Finished at least 2% higher | 17 / 38 (44.7%)12 unavailable | 10,976 / 37,813 (29.0%)5388 unavailable | +15.7 pp |
| Finished at least 2% lower | 15 / 38 (39.5%)12 unavailable | 14,572 / 37,813 (38.5%)5388 unavailable | +0.9 pp |
| Reached +5% at any point | 21 / 39 (53.8%)11 unavailable | 23,205 / 38,386 (60.5%)4815 unavailable | -6.6 pp |
| Reached -5% at any point | 16 / 39 (41.0%)11 unavailable | 20,400 / 38,386 (53.1%)4815 unavailable | -12.1 pp |
Reference: eligible minutes in the same market and scan period, without the proposed conditions. The 30-window reporting floor is not a significance test. Overlapping windows and shared market events can reduce independent evidence. Conditions were selected after seeing this move; confirmation needs unseen data.
Multiple recorded sectors: sector_ai, sector_crypto. Choose one explicitly.
Starting defaults and coverage
Starting defaults: 2026-08-21 to 2026-09-20 UTC; finishes down 2% or more after 1 hour. Review or change the markets, dates, conditions and outcome before running. Current recorded classification and sector tags, not point-in-time membership. Historical membership may differ. Instrument history does not guarantee every setup input or outcome is present; actual study exclusions remain authoritative.
Choose the study population explicitly: This market only · All recorded crypto perpetuals
Watch the ignition replay Free replay covers the latest fully archived UTC day on 6 major markets. Pro reaches 90 days; Research reaches the whole recorded order book, which begins 15 May 2026. Earlier days are searchable but hold no book, so they cannot be replayed. Access is checked when you open a replay.
Evidence and counting
Read at 1 minute before the move.
Of the 50 fires, 29 came before the move began and 10 happened while it was already under way. A fire during the move sees it; it does not telegraph it.
About 1.6 a day, on 18 of 31 days: spread through the record rather than one event.
This is not a similarity search. Each count matches the exact readings above and nothing else.
Picked in hindsight: the counts include every other time these readings appeared, not just this move.
Fires often arrive as clusters of consecutive minutes, and forward windows overlap. A cluster is one episode, not many independent trials.
If these counts read weaker than the move suggested, that is the tool working: the lead-up of a picked move always looks special until it is counted everywhere else.
Also counted
Book imbalance percentile ≤ 0.063 and Compression · 15m ≥ 2.1 and Cascade clustering ≥ 0.5249 matches, 30 rose, 19 fell
Read at 4 hours before the move.
After 30 minutes this setup fell 2% or more in 0 of 49 complete windows (0.0%), against 1.0% with no condition (428 of 42,046): 0.00x.
Best point inside 30 minutes reached +5% or more in 0 of 49 complete windows; worst point fell 5% or more in 0 of 49.
Both are true: these readings sat before this fall, and the other times they appeared, what followed more often rose. That disagreement is the finding: a lead-up that looks special is not the same as one that usually comes before falls.
Of the 49 fires, 27 came before the move began and 12 happened while it was already under way. A fire during the move sees it; it does not telegraph it.
About 1.6 a day, on 18 of 31 days: spread through the record rather than one event.
Compression · 4h ≤ 0.46 and Sweep + reclaim · 15m bullish and Sweep + reclaim · 1h bullish2 matches, 0 rose, 1 fell
Read at 24 hours before the move.
0 rose and 1 fell over 1 complete 30m windows. Too few completed windows for a directional generalization. No horizon cleared the floor of 30 complete windows on both the setup and the reference, so no lift is claimed.
Of the 2 fires, 2 came before the move began; none were during it.
About 0.1 a day over 31 days, but 2026-09-10 alone holds 50% of them: this is close to a single episode.
Every night the record counts what fired and what usually followed: the daily brief.
No account? Replay a real move tick by tick, free: TUT short squeeze · 9 Aug 2026.