Buying the dip in crypto: a tested rule
Price has fallen 1.5 times its usual hourly move below its highest price of the last 4 hours. It's the classic "buy the dip", scaled to each coin's own swings. Traders also call it buying the dip.
On ETH over the past year, price reached the target before going wrong in 54 of every 100 of these (171 in all). Buying at a random time with the same target and stop did it 46 times in 100.
Proven on BTC and ETH. Still being tested on HYPE, SPCX and ZEC.
It comes up about 3 a week on each coin.
Get it free on your phone ›How to spot it
- Start from the coin's highest price of the last 4 hours.
- The alert comes at the first 15-minute close that is 1.5 times the coin's usual hourly move below that high, so the dip is sized to the coin: bigger on a wild day, smaller on a quiet one.
- One alert per dip: the next waits until price is back above the high it fell from, or 7 days have passed.
How the alert is meant to be used
Each alert gives three prices: where the candle closed (the alert price), the line it's wrong under, and the target. The record on this page assumes buying at the alert price, then whichever comes first within 24 hours:
- Target: +0.5% from the alert price, within 24 hours.
- Wrong if a 15-minute candle closes under the dip's lowest price so far. A wick under the line doesn't count; a close does.
- Neither within 24 hours: counted as neither.
The edge is a few more times in 100 than random timing: it helps with timing, not with size. A strong trend can swamp it either way.
A real one: ETH on 10/21
How often it works, by coin
| Coin | Result | Found | Reached the target first | Random buy, same exits |
|---|---|---|---|---|
| BTC | Proven | 175 in a year | 39% | 31% |
| ETH | Proven | 171 in a year | 54% | 46% |
| HYPE | Still being tested | 37 in 90 days | 65% | 49% |
| SPCX | Still being tested | 32 in 90 days | 41% | 43% |
| ZEC | Still being tested | 202 in a year | 66% | 61% |
"Reached the target first": of every 100 of these, how many reached the target before a 15-minute close under the line. "Random buy": the same coin bought at random times of the same hour, volatility and month, with the same target and stop. Proven means it beat random timing over a full year (Binance candles, 8/2025 to 9/2026, the rule fixed in advance). HYPE and SPCX have 90 days of data so far. How we test
Recorded live since 10/16
Since 10/16, after the test data ends, every one is recorded the moment its candle closes: 20 on 5 coins so far. Of the 20 that ended, 13 reached the target, 7 went wrong and 0 did neither in 24 hours. Too few yet to judge on their own. The full track record
| When | Coin | Alert | How it went | |
|---|---|---|---|---|
| 10/29 12:15 | HYPE | 93.94 | reached the target | chart |
| 10/28 14:30 | HYPE | 80.80 | reached the target | chart |
| 10/28 10:00 | HYPE | 79.77 | reached the target | chart |
| 10/28 08:45 | SPCX | 143.97 | reached the target | chart |
| 10/21 17:45 | BTC | 84,205 | closed under the line | chart |
| 10/21 17:30 | ZEC | 1,311.2 | reached the target | chart |
| 10/21 13:15 | ETH | 2,489.6 | closed under the line | chart |
| 10/21 11:00 | HYPE | 81.01 | reached the target | chart |
What the alert says
Down 1.34% from the 4h high (2,498.2), 1.5× its usual hourly move. Wrong if a 15-minute candle closes under 2,456.6. Target 2,477.4 (+0.5%).
Questions
Is buying the dip a good strategy in crypto?
A fixed drop on its own, like "down 2%", did no better than buying at a random time in our tests. A dip sized to each coin's usual move did a little better on BTC and ETH: On ETH over the past year, price reached the target before going wrong in 54 of every 100 of these (171 in all). Buying at a random time with the same target and stop did it 46 times in 100.
How big a dip is it?
1.5 times the coin's usual hourly move below its 4-hour high. The usual hourly move is the average size of its hourly candles (traders call it ATR), so the same rule fits a calm day and a wild one.
Is the dip the bottom?
No. The target is a small bounce, +0.5% within 24 hours, and the alert is wrong if a 15-minute candle closes under the dip's lowest price so far. It's about timing a bounce, not calling the low.
Which coins does the dip from the 4-hour high work on?
Proven on BTC and ETH: there it beat buying at a random time over a full year. Still being tested on HYPE, SPCX and ZEC.
Where is the dip from the 4-hour high wrong?
If a 15-minute candle closes under the dip's lowest price so far. Only a close counts: a wick under the line and back doesn't.
How often does it happen?
On 15-minute candles, about 3 a week on each coin.
Free, no sign-up. The price, the line it's wrong under and the target, the moment the candle closes.