Guide
How to identify a crypto trend reversal, and how often the signal is wrong
A trend reversal is identified after it has begun, not before: price closes on the other side of a line that follows the trend. Every method, a moving average, a channel of recent highs and lows, a volatility band, a momentum measure, is a different way to draw that line, and each trades speed against false alarms. None can tell a reversal from a pullback at the moment it fires. In a seven-year test of one such rule on crypto, 58% of the trades that reversal signals opened ended at a loss. This page compares the four common methods, shows what the record says, and sets out how to use a reversal signal given that it is usually wrong.
Reversal or pullback: you cannot know at the time
Every reversal starts as a pullback. A fall of a tenth inside an uptrend that recovers is a pullback; the same fall that keeps going is the start of a downtrend. On the day, the two look identical, and no indicator separates them, because the difference lies in prices that have not happened yet.
So a reversal signal is not a prediction that the trend has ended. It is a rule for when to stop assuming it continues. Its job is not to be right. Its job is to make being wrong cheap and being right, occasionally, large.
Four ways to define a reversal, compared
These are the families behind nearly every trend indicator. The table describes how each works. It does not rank them: which one did best depends on the market and the years chosen, and anyone who names a winner has usually chosen the years.
| Method | Calls a reversal when | Good at | Weak at |
|---|---|---|---|
| Moving average | Price closes below an average of the last N closes, or a faster average crosses below a slower one. | Simple, and visible on every chart. One setting to understand. | Slow: the longer the average, the later the call. The lookback is fixed, so it does not widen when the market gets wilder, and price chopping around the average flips it repeatedly. |
| Channel breakout | Price closes below the lowest low of the last N candles (a Donchian channel). | Fully objective, with no smoothing: the level is a price that actually traded. | In a volatile market the channel is wide, so a large part of the gain is given back before the exit. Entries come late for the same reason. |
| Volatility band | Price closes beyond a line set a multiple of the average true range (ATR) away, as in Supertrend or a chandelier exit. | The distance scales with volatility: wider when the market is wild, tighter when it is calm. The line only moves in the trend's favour. | Flips often in a sideways market, each time at a small loss. The multiplier trades speed against false alarms, and no setting removes the trade-off. |
| Momentum | A momentum measure turns before price does: a MACD signal-line cross, or price making a new high while RSI does not (a divergence). | Can turn earlier than anything based on price alone. | Earliest also means most false alarms. In a strong trend momentum can fade for a long time while price keeps going, and reading a divergence takes judgement. |
All four share one property: the faster a method calls a reversal, the more often it calls one that is not there. Tuning the settings moves a method along that line; it does not take it off. What separates people who use these tools well from people who do not is rarely the choice of indicator. It is whether they fixed one in advance and kept to it. The general guide to trend trading covers the strategies built on each family. The best-known moving-average signal, the golden cross, has a guide of its own.
What the record says about reversal signals
TideTrade's rule is of the third kind, a volatility band: the Supertrend indicator with period 10 and multiplier 3. Its pre-registered study ran the rule on the 20 most-traded crypto assets from 2019 to 2026. Every round trip in it was opened by one reversal signal and closed by the next:
A reversal signal from this rule was wrong more often than it was right. The method paid in that data only because the losses were small and a minority of the wins were very large. Over the period its worst fall was 46% against 85% for holding the same basket; in the four large falls it avoided 41 to 95% of each fall and missed 38 to 72% of the rebound that followed. The study's own summary: in past data the rule had smaller falls; its return was similar overall and lower in recent years.
These figures describe one rule in one market over one period. Of the other three methods the site has measured one, a moving average, in a pre-registered comparison on the same data. Its outcome, in the wording published: In past data Supertrend(10, 3) and the closest-matched moving average were not clearly different in how far they fell. Supertrend's return was higher. The closest match by trade count was a 250-day average, which still traded more often. Channel breakouts and momentum measures have not been measured here, and the site makes no claim about them. Past data is not a forecast.
When a reversal counts: completed candles
A candle that is still forming can cross a line and cross back before it closes. On TideTrade's own data about half of the flips seen on a forming candle reverse before it ends. A reversal read before the close is therefore a coin toss presented as news. Daily crypto candles close at 00:00 UTC; a reading taken after that is the one that goes on the record.
The useful early warning is not the forming candle but the distance to the line. An asset whose close is within half an ATR of its line could change state on the next candle; TideTrade shows these as slack water. That is a statement about distance, not a signal.
How to use a signal that is usually wrong
- Fix the method before you look. Choose one definition of a reversal and its settings in advance. Changing the method after a signal you dislike is the same as having no method.
- Read it on completed candles. A line crossed in the middle of a candle is often uncrossed by the close. Decide what counts as a close and wait for it.
- Make being wrong cheap. Since most signals fail, the size of each position matters more than the timing of any one signal. Spread across several assets and never borrow to hold a trend.
- Take every signal. The few trips that pay for the rest cannot be told apart in advance from the many that do not. Skipping the ones that feel wrong removes the method's whole edge.
- Use the longer candle for context. A weekly reading flips less often and later than a daily one. Knowing both says whether a daily flip is happening inside a larger uptrend or a larger downtrend, without telling you which will win.
The Tide Rules turn these into five mechanical rules for one particular line.
Where to see reversals as they are recorded
On TideTrade a reversal is called a flip. The scanner shows each of 140 crypto assets with its state, its distance from the line and when it last flipped, and can list what flipped in the last day, week or month. The ledger holds every flip with the days held and the result to the next flip, losing ones included, with a daily hash so that an edit would show. The same states are available as data through the free API.
Questions people ask about trend reversals
- Can a crypto trend reversal be predicted?
- Not reliably. Every method identifies a reversal after it has begun, when price closes on the other side of a line that follows the trend. In the study of TideTrade's rule 58% of the 682 round trips opened by a reversal signal lost money.
- Which indicator is best for spotting a trend reversal?
- None is best in every market. Moving averages are simple but slow and do not adapt to volatility; channel breakouts are objective but give back a large part of the gain before they exit; volatility bands such as Supertrend adapt to volatility but flip often in sideways markets; momentum measures turn earliest and give the most false alarms. Fixing one method in advance and keeping losses small matters more than which one.
- How do I tell a reversal from a pullback?
- At the moment of the signal you cannot: the difference lies in prices that have not happened yet. A pullback that keeps falling becomes a reversal, and one that recovers becomes a whipsaw. A rule treats both the same way and relies on the whipsaws being cheap.
- What is a whipsaw?
- A signal that reverses soon after it fires, so the trade it opened closes at a small loss. In the study of TideTrade's rule 12% of round trips were lost within ten days, and the average losing trip lost 14%.
- Should I act on a reversal before the candle closes?
- No. A forming candle can cross a line and cross back before it closes; on TideTrade's own data about half of the flips seen on a forming candle reverse before it ends. Daily crypto candles close at 00:00 UTC, and a reading taken after the close is the one that counts.
- Which timeframe shows a trend reversal best?
- Longer candles flip less often and later; shorter ones flip sooner and more often. TideTrade reads daily and weekly candles, both only after they close, so a daily flip can be seen against the weekly state.
TideTrade is decision support, not financial advice. Cryptoassets are high-risk; trend rules lose money in sideways markets and past states do not predict future prices. See the disclaimer. Figures are from the published, pre-registered study.