Everyone in trading knows prices trend and then they revert. It's one of the oldest patterns in finance, documented to death. Knowing it exists buys you nothing though. The real work is figuring out when the conditions actually favor a reversion trade, and then structuring it so the pile of small frequent wins covers the occasional trend continuation that tries to take your head off.
What actually pulls price back
Reversion comes from three forces working together. There's fundamental anchoring, where an asset has some notion of fair value from earnings or utility or network effects, and prices that stretch too far from it attract buyers or sellers who lean the other way. There's position unwinding, where the traders who drove a move start taking profits, and that alone reverses part of the action. And there's liquidity behavior, where providers step in to sell into spikes and buy into crashes, which dampens the extremes.
All three are strongest when the original move came from sentiment or technicals rather than a real change in the fundamentals. A 20% rally on FOMO with no catalyst is far more likely to unwind than a 20% rally on a genuine earnings beat or a protocol upgrade that actually made the asset worth more. That distinction is the whole game, and it's the part most people skip.
Measuring the tendency
The simplest statistical read on reversion is the Hurst exponent. Below 0.5 means the series tends to reverse, above 0.5 means it tends to continue, and 0.5 is a random walk. You can compute it for any price series with R/S analysis or DFA.
The catch is that the Hurst exponent drifts over time, and an asset can be mean-reverting on one time scale and trending on another. Bitcoin's a clean example. It has historically trended on weekly and monthly frames but behaved more mean-reverting on daily ones. So the same asset wants a different strategy depending on how long you hold, which is easy to forget when you've decided it's a "trending" or "reverting" name.
When to actually enter
A typical entry is a move of two or more standard deviations from the recent mean paired with fading momentum, so price is still stretched but the rate of increase is slowing. Layering in volume helps a lot. If the extreme move happened on declining volume, it's more likely to revert, because it was thin liquidity pushing price around rather than real conviction behind it.
Bollinger Band extremes, RSI overbought and oversold levels, and distance from the 20-day moving average all get at the same idea from slightly different angles. Stacking two or three of them cuts false signals versus leaning on any single one. In Blockcircle I've found the volume filter does more work than people expect, since it's the cheapest way to separate a real exhaustion from a move that's still building.
Keeping the losses small
The obvious risk is that the move keeps going instead of reverting. A two-sigma move can turn into a four-sigma move, and fighting that can be ruinous. Stops are non-negotiable, but they set up a real tension. Too tight and you get shaken out right before the snap-back. Too wide and the one continuation event eats months of accumulated gains.
What works for me is pairing price stops with time stops. If a reversion trade hasn't moved your way inside a set number of periods, say five daily candles for a swing, you exit regardless of price. It frees up capital stuck in trades where the reversion just isn't showing up, even when price never touched your stop. Dead trades cost you the opportunity, not just the drawdown.
Why it wants to be a basket
Reversion works better as a portfolio than as a string of individual picks. Keep a watchlist of candidates and rotate into the most stretched ones as they set up, and you spread out the idiosyncratic risk of any single trade going wrong. The win rate on one trade might sit around 55 to 60%, which feels shaky on its own. But running several positions at once, with the winners quietly covering the losers, smooths the equity curve in a way no single trade sequence ever will. Build the process around that and the individual trade stops feeling so heavy.