Research is a work in progress. Estimated release: 1 December 2026. Our goal is to make it a search engine for trading setups in recorded crypto markets.
Binance USDT-MARBUSDT rose 87% on 2 September 2026
Picked after the rise, on one market.
Next stepWhat happened
Price opened the 14:00 hour at 0.1109 and reached a highest trade of 0.2068 in the 08:00 hour; the tested readings were taken at 13:59.
What stood out at 13:59
- Order book tilt0.6458Tested: 0.64 or higher
- Containment (15m)outsideTested: is outside
- Liquidation bunching0.5272Tested: 0.52 or higher
13:59 always meets its own rules.
What the record says
More often than any ARB minute, in this sample
Price at least 2% higher after 30 minutes
111 matches in 43,201 minutes checked.
Picked after the rise, on one market.
Evidence
Does it hold on other dates?
Nothing runs until you press Run.
Exactly what was counted3 conditions · 111 of 43,201 minutes · saved text
- Order book tilt 0.64 or higher. Size at the best bid against the best ask; above 0 means more bids, below 0 more asks.Book imbalance · top levels ≥ 0.64 · feature.book_imbalance
- Containment (15m) is outside.Containment · 15m outside · feature.candle_containment_15m
- Liquidation bunching 0.52 or higher. How bunched in time recent liquidations were; higher means more bunched, and it stays below 1.Cascade clustering ≥ 0.52 · feature.cascade_clustering_index
Read at 1 minute before the move.
Read at 13:59 UTC on 2 September, one minute before the selected 14:00 start.
ARB only, 10 August to 8 September 2026 (UTC, ending 9 September 00:00); all 3 readings had values on 43,201 of those minutes.
Reference: every ARB minute over the same dates, matches included; it is not a matched control group.
Outcome: the close exactly 30 minutes later, not the highest or lowest price along the way.
The 14:00 start was picked automatically after the move.
ARBUSDT stored readings end 2026-09-09 00:00 UTC (exclusive). Prices and tick replay have separate coverage.
The feature manifest spans the lead-up. Individual minutes or inputs can still be unavailable.
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.
Saved study q#4a84c956 @ dataset 418a74e4, counted 9 Sep 2026 09:58 UTC.
Saved summary text, as written 9 Sep 2026
Written automatically when this study was saved and kept as written; where a line differs from the counts above, the counts govern.
Minutes like this appeared 111 times in the 43,201 minutes searched (0.3%). What followed was close to a coin flip: 52 rose and 59 fell over the next 30 minutes (111 of the 111 matches have a complete 30 minutes forward).
After 4 hours this setup fell 2% or more in 19 of 111 complete windows (17.1%), against 12.2% with no condition (5,258 of 42,961): 1.40x.
This reports the time frame with the biggest gap from any minute, chosen after counting. At least 16 of its 111 windows overlap the rise.
Best point inside 4 hours reached +5% or more in 14 of 111 complete windows; worst point fell 5% or more in 4 of 111.
About 3.6 a day, on 20 of 31 days: spread through the record rather than one event.
It counts the single minute on 9 September as a day.
The other reading times, as written
From 15 minutes before
Minutes like this appeared only 19 times in the 43,168 minutes searched. What followed was more often higher: 12 of the 19 matches with a complete 30 minutes forward rose (63%).
12 rose and 7 fell over 19 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.
About 0.6 a day, on 14 of 31 days: spread through the record rather than one event.
From 1 hour before
Minutes like this appeared only 1 time in the 43,201 minutes searched. What followed was more often higher: 1 of the 1 matches with a complete 30 minutes forward rose (100%).
1 rose and 0 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.
About 0 a day over 31 days, but 2026-09-02 alone holds 100% of them: this is close to a single episode.
Picked after the rise, on one market; the 111 matches came on 20 of 30 days, 16 during the rise.
When the matches happened20 of 30 days
- 19 Aug2 matches
- 20 Aug7 matches
- 21 Aug6 matches
- 22 Aug8 matches
- 23 Aug8 matches
- 24 Aug4 matches
- 25 Aug7 matches
- 26 Aug8 matches
- 27 Aug10 matches
- 28 Aug8 matches
- 29 Aug10 matches
- 30 Aug1 match
- 1 Sep1 match
- 2 Sep7 matches
- 3 Sep3 matches
- 4 Sep3 matches
- 5 Sep6 matches
- 6 Sep3 matches
- 7 Sep7 matches
- 8 Sep2 matches
No matches on the other 10 days.
Of the 111 matches, 84 came before the rise began, 16 during it and 11 after it.
Neighbouring minutes overlap, so these are not 111 separate trials.
Picked after the rise, on one market; the 111 matches came on 20 of 30 days, 16 during the rise.
Other time frames30 minutes to 7 days
The answer uses 30 minutes, the time frame this study counted by default. Each row counts the same matches at a later close, in both directions. The bars show how often price was 2% or more higher at each time frame: first the matches, then any ARB minute.
- 30 minutes
2 of 111 (1.8%) were 2% or more lower; any ARB minute 821 of 43,171 (1.9%).
Every match has a 30-minute close.
At least 16 of these windows overlap the rise.
- 1 hour
5 of 111 (4.5%) were 2% or more lower; any ARB minute 1,790 of 43,141 (4.1%).
Every match has a 1-hour close.
At least 16 of these windows overlap the rise.
- 4 hours
19 of 111 (17.1%) were 2% or more lower; any ARB minute 5,258 of 42,961 (12.2%).
Every match has a 4-hour close.
At least 16 of these windows overlap the rise.
- 24 hours
29 of 109 (26.6%) were 2% or more lower; any ARB minute 11,342 of 41,761 (27.2%).
2 without a 24-hour close yet, left out.
At least 16 of these windows overlap the rise.
- 72 hours
48 of 99 (48.5%) were 2% or more lower; any ARB minute 13,809 of 38,881 (35.5%).
12 without a 72-hour close yet, left out.
At least 1 of these windows overlap the rise.
- 7 days
23 of 80 (28.7%) were 2% or more lower; any ARB minute 8,709 of 33,121 (26.3%).
31 without a 7-day close yet, left out.
At least 30 of these windows overlap the rise.
The saved summary picked, after counting, the time frame with the biggest gap from any minute: 4 hours.
After 4 hours this setup fell 2% or more in 19 of 111 complete windows (17.1%), against 12.2% with no condition (5,258 of 42,961): 1.40x.
At least 16 of those windows ran through the rise itself, so it partly counts the rise as an outcome.
Along the way (the highest and lowest price inside 4 hours, not the close):
Best point inside 4 hours reached +5% or more in 14 of 111 complete windows; worst point fell 5% or more in 4 of 111.
Picked after the rise, on one market; the 111 matches came on 20 of 30 days, 16 during the rise.
Readings at five times before the rise24 hours to 1 minute before · 2 other times tried
Highlights follow fixed rules, not significance tests. Proposed conditions round values or use ranges; inspect them before running.
24 hours before · 1 Sep 14:00 UTC · 8 of 58 stood out
Order book lopsidedness 0.91 or higher4-hour range against usual 2.1 or higherHow many times the broken descending resistance had been tested 3 or higherHow long ago the descending resistance broke between 12 and 720Touches on the broken resistance 3 or moreTime since the resistance broke between 12 and 720 hours15-minute price change -0.014 or lowerDistance from the day's average price -0.022 or lowerShow the full readings
Reading Value at that minute Stored rule Book imbalance percentile feature.book_imbalance_pctrank 0.9143 ≥ 0.91 the last four-hour candle range versus its baseline feature.candle_compression_4h 2.136 ≥ 2.1 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 5 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 310 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 3 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 21 12 to 720 Return · 15m feature.ret_15m -0.01464 ≤ -0.014 VWAP distance · day feature.vwap_dist -0.02264 ≤ -0.022 4 hours before · 2 Sep 10:00 UTC · 6 of 58 stood out
Order book tilt -0.75 or lowerLiquidation bunching 0.53 or higherHow many times the broken descending resistance had been tested 3 or higherHow long ago the descending resistance broke between 12 and 720Touches on the broken resistance 3 or moreTime since the resistance broke between 12 and 720 hoursShow the full readings
Reading Value at that minute Stored rule Book imbalance · top levels feature.book_imbalance -0.7522 ≤ -0.75 Cascade clustering feature.cascade_clustering_index 0.5372 ≥ 0.53 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 5 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 330 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 3 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 41 12 to 720 1 hour before · 2 Sep 13:00 UTC · 10 of 58 stood out
Order book tilt -0.87 or lower15-minute range against usual 0.31 or lowerContainment (15m) is insideLiquidation bunching 0.52 or higherHow many times the broken descending resistance had been tested 3 or higherHow long ago the descending resistance broke between 12 and 720Touches on the broken resistance 3 or moreTime since the resistance broke between 12 and 720 hoursVolatility rank 0.024 or lowerDistance from the day's average price -0.031 or lowerShow the full readings
Reading Value at that minute Stored rule Book imbalance · top levels feature.book_imbalance -0.8788 ≤ -0.87 Compression · 15m feature.candle_compression_15m 0.3053 ≤ 0.31 Containment · 15m feature.candle_containment_15m inside inside Cascade clustering feature.cascade_clustering_index 0.5287 ≥ 0.52 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 5 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 333 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 3 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 44 12 to 720 Realized vol percentile feature.realized_vol_pctrank 0.02325 ≤ 0.024 VWAP distance · day feature.vwap_dist -0.03198 ≤ -0.031 15 minutes before · 2 Sep 13:45 UTC · 10 of 58 stood out
Order book tilt -0.83 or lowerContainment (15m) is outsideLiquidation bunching 0.52 or higherHow many times the broken descending resistance had been tested 3 or higherHow long ago the descending resistance broke between 12 and 720Touches on the broken resistance 3 or moreTime since the resistance broke between 12 and 720 hoursOpen interest change rank 0.99 or higher15-minute price change 0.018 or higher1-hour price change 0.025 or higherShow the full readings
Reading Value at that minute Stored rule Book imbalance · top levels feature.book_imbalance -0.8343 ≤ -0.83 Containment · 15m feature.candle_containment_15m outside outside Cascade clustering feature.cascade_clustering_index 0.5272 ≥ 0.52 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 5 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 333.8 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 3 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 44.75 12 to 720 OI velocity percentile feature.oi_velocity_pctrank 0.9988 ≥ 0.99 Return · 15m feature.ret_15m 0.01833 ≥ 0.018 Return · 1h feature.ret_1h 0.02525 ≥ 0.025 1 minute before · 2 Sep 13:59 UTC · 8 of 58 stood out
Order book tilt 0.64 or higherContainment (15m) is outsideLiquidation bunching 0.52 or higherHow many times the broken descending resistance had been tested 3 or higherHow long ago the descending resistance broke between 12 and 720Touches on the broken resistance 3 or moreTime since the resistance broke between 12 and 720 hours1-hour price change 0.022 or higherShow the full readings
Reading Value at that minute Stored rule Book imbalance · top levels feature.book_imbalance 0.6458 ≥ 0.64 Containment · 15m feature.candle_containment_15m outside outside Cascade clustering feature.cascade_clustering_index 0.5272 ≥ 0.52 how many times the broken descending resistance had been tested feature.desc_resistance_broken_touches 5 ≥ 3 how long ago the descending resistance broke feature.desc_resistance_hours_since_break 334 12 to 720 how many times the broken short descending resistance had been tested feature.desc_resistance_short_broken_touches 3 ≥ 3 how long ago the short descending resistance broke feature.desc_resistance_short_hours_since_break 44.98 12 to 720 Return · 1h feature.ret_1h 0.02205 ≥ 0.022
Two other reading times we tried 19 and 1 matches
15 minutes before (13:45): order book tilt -0.83 or lower, containment (15m) is outside, open interest change rank 0.99 or higher.
Book imbalance · top levels ≤ -0.83 and Containment · 15m outside and OI velocity percentile ≥ 0.99
19 matches: price was 2% or more higher after 30 minutes in 0 of 19, and 2% or more lower in 0 of 19.
Of the 19 matches, 14 came before the rise began, 2 during it and 3 after it.
1 hour before (13:00): order book tilt -0.87 or lower, 15-minute range against usual 0.31 or lower, volatility rank 0.024 or lower.
Book imbalance · top levels ≤ -0.87 and Compression · 15m ≤ 0.31 and Realized vol percentile ≤ 0.024
1 match: price was 2% or more higher after 30 minutes in 0 of 1, and 2% or more lower in 0 of 1.
Of the 1 match, 1 came before the rise began, 0 during it and 0 after it.
Fewer than 30 results, so no rate is printed.
The saved study leads with 13:59 because it is the only set with at least 30 matches before the rise.
Picked after the rise, on one market; the 111 matches came on 20 of 30 days, 16 during the rise.
Hourly candles around the risefull context, lead-up, move
The 14:00 start was picked after the move, and the hour before was already rising.
Hourly candles, UTC. 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.
Other ways to test itsame dates
- This market only
- The 21 layer 2 sector markets · No result for these markets has been measured.
- Every recorded market
Same 3 conditions on ARB, 11 July to 9 August 2026.
Each opens Review the test.
Running needs a free account and uses your allowance; you see the cost first.