Volatility
Volatility drives every cost of LPing: the more a price moves, the more an LP loses to arbitrage and the more often a range has to be re-centred. rangebound uses the most cautious of several measurements.
The measurements
All yearly rates.
| Measure | What it is |
|---|---|
| RV 1h, 1d, 5d | Realised volatility of the listed underlying from 5-minute returns during US market hours |
| RV 20d | Realised volatility from the last 20 daily closes |
| IV 30d | Implied volatility of the underlying's at-the-money options about 30 days out |
| On-chain RV 24h, 7d | Realised volatility of the token's own pool price, around the clock |
Session realised volatility ignores the overnight gap, so outside US hours it describes the last session. On-chain realised volatility fills that gap: it comes from the pool itself, 24/7, so an overnight sell-off in the token raises it even while the stock is closed.
The volatility in use (σ)
σ = the largest of: IV 30d, RV 5d (session), on-chain RV 7dTaking the largest makes the costs conservative. The source used is shown next to σ wherever it appears, for example σ 34% · IV.
One-day readings don't set σ. Instead, when the last day's realised volatility is more than twice σ, the pool gets a Vol spike flag.
IV against RV
IV / RV = IV 30d ÷ RV 5dAbove 1.5, options are pricing a bigger move than the stock has made recently, often ahead of earnings or news, and the pool gets an IV ≫ RV flag.
Where you see it
The pool page's Volatility section shows every measure as a bar, highlights the one in use, and gives the resulting LVR for a full-range and a ±2% position.
Caveats
- Thin pools read high on-chain. A trade pushes the price and arbitrage pulls it back, which looks like volatility. Five-minute sampling reduces this, but the thinnest pools can still read somewhat high.
- No live reference overnight. When the US market is closed, there is no traded underlying to check the token against.
- Market data is unofficial. Listed-market data comes from public sources, not an exchange feed.