POINT-IN-TIME PERPETUAL FUTURES RESEARCH
Do the ranked Top 5 actually separate from the rest?
This is a descriptive research log, not a backfilled chart study. Each market snapshot was frozen before the later +1h outcome was known, then matured candidates were compared by absolute post-scan movement.
Current sample
Top 5 vs remaining ranked candidates
| Horizon | Top 5 n | Top 5 median |move| | Rest n | Rest median |move| | Median ratio |
|---|---|---|---|---|---|
| +1h | 110 | 1.54% | 286 | 0.69% | 2.23× |
| +4h | 110 | 3.60% | 286 | 1.43% | 2.51× |
| +8h | 100 | 6.16% | 260 | 2.24% | 2.75× |
| +12h | 80 | 7.03% | 208 | 2.93% | 2.40× |
Longer horizons have fewer matured observations. These ratios describe separation inside this candidate set; they are not probabilities, directional predictions, win rates or trading returns.
Chronological split check
The 22 snapshots were split chronologically in half rather than randomly. At +1h, Top-5 median absolute movement was 1.35% in the early half and 1.59% in the later half. The remaining ranked candidates were 0.69% and 0.68% respectively.
This does not establish a durable out-of-sample trading edge. It is a basic stability check showing that the observed ranking separation was not concentrated only in the first half of this small sample.
What Money Radar ranks
- Funding rate normalized to an 8-hour equivalent when settlement intervals differ.
- Short-term open-interest change from frozen observations.
- Closed 15-minute price structure.
- Relative volume versus the previous 20 completed 15-minute bars.
- Mark/index basis.
The score is designed to prioritize which perpetual markets deserve research attention first. It does not decide long or short direction.
Get the next verified Top 5 by emailOne email field. Get selected verified anomaly briefs and the growing point-in-time outcome report. No exchange login, wallet or API key required.
Or open the free Top 5 screener →Method limits
Current evidence is descriptive and sample-limited. The candidate universe is defined by the screener methodology, longer-horizon samples are smaller, and absolute movement does not measure whether a hypothetical trade would have been profitable. Continued frozen snapshots are needed before stronger claims would be justified.