What Tisiver ranks
Tisiver is a free perpetual-futures research dashboard. Signals screens crypto, stock, commodity, FX, index and bond perpetuals with an adjustable market-activity score. These are derivative contracts, not ownership of the underlying stocks or commodities. The separate rv60m machine-learning watchlist ranks crypto perpetuals by expected next-hour volatility. The two methods answer different questions: what is active now, and where does the model expect more movement next?
Data sources and freshness
Tisiver analyzes publicly available derivatives-market data. Coverage depends on the method and the available snapshot; a market appearing in Signals does not mean it is covered by the ML model. Each workspace shows its own snapshot timestamp and availability status.
Snapshots are normally published roughly every five minutes, and the workspaces check for new data every 30 seconds. A refresh request does not guarantee a newer snapshot. Check the displayed timestamp and any stale or unavailable message before comparing observations.
Four trend-signal weights
Signals combines turnover, funding, momentum and open interest. Turnover measures trading volume relative to open interest; open interest measures the size of outstanding positions. Funding and momentum contribute by magnitude, so a high score does not imply upward price direction. Each component is scaled and capped before the weighted average is calculated.
The sliders change how much each component contributes. Raising one weight can reorder the board without changing any market data. Treat the score as a transparent screening preference, not a probability of profit. Keep the same time window and weights when comparing observations.
Worked example: turnover versus open interest
Illustration only: two markets each trade $10 million over 24 hours. Market A has $5 million of open interest, while Market B has $20 million. Their volume-to-open-interest ratios are 2 and 0.5 respectively. A has higher turnover even though B has more outstanding positions.
A turnover-heavy view and an open-interest-heavy view can therefore favor different markets. The final order also depends on funding, momentum and the selected weights. This example explains the inputs; it is not a live ranking or a trading recommendation.
How to read the rv60m model
rv60m estimates realized volatility over the next 60 minutes for crypto perpetuals. The volatility measure uses the square root of the sum of twelve squared five-minute log returns. Squaring the returns makes both rising and falling moves contribute. A price can return to its starting point and still have substantial realized volatility.
The published score expresses relative rank within a snapshot, not the predicted percentage price change or the chance of a profitable trade. A score of 0.9 is not a 90% probability of a price rise. Check the snapshot timestamp, model version and current coverage; predictions can be wrong and coverage can change.
Compare methods without averaging their ranks
Signals describes current market activity using chosen weights. rv60m estimates future volatility. A market can rank highly in one and lower in the other without either display being broken. Compare keeps the methods separate; a missing model rank can mean missing coverage rather than low expected volatility.
All information is for research, not investment advice. No rank or score accounts for every execution cost or guarantees a return. Use the explanations and the individual inputs to form questions for further research.