Price after hourly price and OI states
For one exact perpetual contract: each time an hour entered a new price and open interest state, what did the recorded price do over the next 1, 4 and 24 hours? A description of observed history only.
Contract and range
Bybit PYPLUSDT · last 7d
Contract bybit:perp:PYPLUSDT. Hourly buckets from 2026-10-12 01:00 UTC to 2026-10-19 02:00 UTC; the newest bucket closed at 2026-10-19 02:00 UTC, at least 10 minutes ago. Source: exchange public data recorded by SwoleCharts (price: bars_1h, open interest: market_1h). Hour-by-hour state history · Compare two recorded hours
168 of 168 hours classified; 8 state entries found.
Price change after entering each state
| State entered | After | n | Higher | Median | 25th pct | 75th pct |
|---|---|---|---|---|---|---|
| Price up, exposure up | 1h | 2 | insufficient observations | |||
| 4h | 1 | insufficient observations | ||||
| 24h | 1 | insufficient observations | ||||
| Price down, exposure up | 1h | 3 | insufficient observations | |||
| 4h | 3 | insufficient observations | ||||
| 24h | 1 | insufficient observations | ||||
| Price up, exposure down | 1h | 2 | insufficient observations | |||
| 4h | 2 | insufficient observations | ||||
| 24h | 2 | insufficient observations | ||||
| Price down, exposure down | 1h | 1 | insufficient observations | |||
| 4h | 1 | insufficient observations | ||||
| 24h | 1 | insufficient observations | ||||
n is the number of selected, non-overlapping entries. Higher is the share of those whose recorded close was strictly above the entry close; unchanged closes count in the denominator. With fewer than 5 selected entries the distribution is not shown.
Entries, exclusions and selection
| State entered | After | Entries | Run start unknown | Not yet matured | Recording gap | Missing price | Eligible | Overlapping, skipped | Selected n |
|---|---|---|---|---|---|---|---|---|---|
| Price up, exposure up | 1h | 2 | 0 | 0 | 0 | 0 | 2 | 0 | 2 |
| 4h | 2 | 0 | 0 | 0 | 0 | 2 | 1 | 1 | |
| 24h | 2 | 0 | 0 | 0 | 0 | 2 | 1 | 1 | |
| Price down, exposure up | 1h | 3 | 0 | 0 | 0 | 0 | 3 | 0 | 3 |
| 4h | 3 | 0 | 0 | 0 | 0 | 3 | 0 | 3 | |
| 24h | 3 | 0 | 0 | 0 | 0 | 3 | 2 | 1 | |
| Price up, exposure down | 1h | 2 | 0 | 0 | 0 | 0 | 2 | 0 | 2 |
| 4h | 2 | 0 | 0 | 0 | 0 | 2 | 0 | 2 | |
| 24h | 2 | 0 | 0 | 0 | 0 | 2 | 0 | 2 | |
| Price down, exposure down | 1h | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 |
| 4h | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | |
| 24h | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 |
Entries counts hours that entered the state from a different known state. Run start unknown counts hours in the state whose previous hour was unavailable or before the window; they are not entries. Each horizon then drops entries whose outcome hour has not completed, touches a recording gap, or lacks a recorded price, and skips entries that overlap the previous selected one.
Selected entries, newest first
Up to 20 most recent selected entries per state and horizon, for checking against the state history.
Price up, exposure up
| Entry hour (UTC) | From | After | Entry close | Later close | Change |
|---|---|---|---|---|---|
| 2026-10-15 15:00 | Below threshold | 1h | 56.400 | 56.360 | -0.07% |
| 2026-10-15 13:00 | Below threshold | 1h | 55.930 | 56.160 | +0.41% |
| 2026-10-15 13:00 | Below threshold | 4h | 55.930 | 57.020 | +1.95% |
| 2026-10-15 13:00 | Below threshold | 24h | 55.930 | 57.890 | +3.50% |
Price down, exposure up
| Entry hour (UTC) | From | After | Entry close | Later close | Change |
|---|---|---|---|---|---|
| 2026-10-15 10:00 | Below threshold | 1h | 55.610 | 55.580 | -0.05% |
| 2026-10-14 20:00 | Below threshold | 1h | 56.030 | 56.120 | +0.16% |
| 2026-10-14 13:00 | Below threshold | 1h | 55.960 | 55.890 | -0.13% |
| 2026-10-15 10:00 | Below threshold | 4h | 55.610 | 56.160 | +0.99% |
| 2026-10-14 20:00 | Below threshold | 4h | 56.030 | 56.180 | +0.27% |
| 2026-10-14 13:00 | Below threshold | 4h | 55.960 | 55.620 | -0.61% |
| 2026-10-14 13:00 | Below threshold | 24h | 55.960 | 55.930 | -0.05% |
Price up, exposure down
| Entry hour (UTC) | From | After | Entry close | Later close | Change |
|---|---|---|---|---|---|
| 2026-10-17 01:00 | Below threshold | 1h | 56.460 | 56.490 | +0.05% |
| 2026-10-12 15:00 | Below threshold | 1h | 56.030 | 55.980 | -0.09% |
| 2026-10-17 01:00 | Below threshold | 4h | 56.460 | 56.410 | -0.09% |
| 2026-10-12 15:00 | Below threshold | 4h | 56.030 | 56.100 | +0.12% |
| 2026-10-17 01:00 | Below threshold | 24h | 56.460 | 56.420 | -0.07% |
| 2026-10-12 15:00 | Below threshold | 24h | 56.030 | 56.080 | +0.09% |
Price down, exposure down
| Entry hour (UTC) | From | After | Entry close | Later close | Change |
|---|---|---|---|---|---|
| 2026-10-16 21:00 | Below threshold | 1h | 56.700 | 56.610 | -0.16% |
| 2026-10-16 21:00 | Below threshold | 4h | 56.700 | 56.460 | -0.42% |
| 2026-10-16 21:00 | Below threshold | 24h | 56.700 | 56.430 | -0.48% |
How the outcomes are measured
Read the guide to hourly price and OI states and what followed for worked examples and limits.
States are the ones on the hourly state history: each hourly bucket compared with the one before by price change and price-adjusted open interest change. An entry is an hour in one of the four directional states whose previous hour had a different known state, including below threshold. It is known at that hour's close, so the entry never depends on how long the state later lasted.
The change after k hours is the recorded close k hours later divided by the entry hour's close, minus one. It is measured only when the later hour has completed, every hourly price in between is recorded and positive, and the engine was recording throughout. Excluded entries are counted, never filled from a neighbouring hour, interpolated or replaced with a live quote.
Within one state and horizon, entries are taken in time order: the earliest is kept and any entry that starts before the kept one's later close is skipped, whatever its return. The same hours still feed several states' and horizons' rows, so the rows are not independent of each other. Quantiles interpolate linearly between sorted returns (position q × (n − 1)).
This is a description of a short recorded history of one contract, not a forecast, signal or backtest. A handful of entries over 7 or 30 days says little about other periods, and the states describe what moved, not who traded.