Scan to get Binance App
Download APKCatch single-candle high-conviction reversals on Binance perpetuals — master bullish and bearish Belt Hold recognition and execution.
The Belt Hold (Japanese sumo term Yorikiri, literally 'grab the opponent's belt and push them out of the ring') is a single-candle high-conviction reversal pattern that comes in two flavours: Bullish Belt Hold and Bearish Belt Hold. The core recognition rule is: the open equals the candle's extreme — bullish belt hold opens at the low (no lower wick / 'shaved bottom'), bearish belt hold opens at the high (no upper wick / 'shaved top').
The pattern was first documented in Japan's Sakata Five Methods and brought to Western audiences by Steve Nison in Japanese Candlestick Charting Techniques. It is a strong single-bar reversal signal but is weaker than 2~3 candle combinations (engulfing, harami, three outside, three inside, etc.); it requires trend location, volume and a second confirmation candle to be tradeable — otherwise it is easily mistaken for an ordinary marubozu.
On Binance Futures the Belt Hold typically appears: 1) at the end of a downtrend on key support (bullish); 2) at the end of an uptrend on key resistance (bearish); 3) on candle-period boundaries that produce gap-like open behaviour. Mastering Belt Hold lets you identify reversals at the very first candle of the move, at the cost of accepting single-bar signal risk — always wait for the next candle to confirm.
When the open sits exactly on the low, buyers dominated from the very first tick — bears couldn't even probe lower. This 'open-strong' rhythm usually means institutional long positioning was completed overnight/pre-session and unleashed at the open. Appearing at downtrend support, it commonly marks institutions stepping in and cascading short-stop runs.
A standalone bullish belt hold is not enough. Wait for K2: 1) K2 opens inside K1's body; 2) K2 closes above K1's close; 3) K2 does NOT break K1's low (the belt-hold open). All three required. Highest win-rate setup on Binance Futures: daily/4H bullish belt hold + K2 bullish break + RSI rising out of oversold + extreme negative funding rate — four-factor confluence pushes win rate to 80%+.
Open = high means sellers dominated from the first tick — late longs were trapped immediately. At uptrend resistance this typically reflects institutional distribution at the top, short positioning pre-built overnight, with the dump triggered as the new period opens.
K2 must: 1) open inside K1's body; 2) close below K1's close; 3) NOT break above K1's high (the belt-hold open). High win rate setup: daily/4H bearish belt hold + K2 bearish continuation + RSI falling out of overbought + extreme positive funding — four-factor confluence ~80%+.
Belt Hold has slightly different meaning in gapped markets (US/A-share equities) versus 24/7 continuous markets like Binance perpetuals.
Binance perps trade continuously, so there is no traditional opening gap, but candle-period boundaries (UTC midnight rollover for 4H/1D) still behave as a 'new candle open'. A belt hold that prints right on a period boundary is strongest because the previous candle closed and the new one immediately opens at an extreme — no probing, strong capital consensus.
When the new candle opens with a clear gap from the prior close (rare on Binance perps but happens around major news), the belt hold doubles in strength. Bullish gap belt hold = up-gap + bullish belt hold; bearish gap belt hold = down-gap + bearish belt hold. This is the precursor of a classic 'island reversal' in equities; on Binance Futures it usually comes from funding settlements, Binance announcements or macro releases.
The critical element is where the open prints. Best case: bullish belt hold opens at prior low / EMA200 / Fibo 0.618; bearish opens at prior high / EMA200 / Fibo 0.618 extension. This 'key level + extreme open' double anchor is how real belt holds are separated from coincidental candles.
Belt Hold is often confused with the Marubozu ('shaved head and feet'). They are related but distinct.
| Attribute | Belt Hold | Marubozu |
|---|---|---|
| Open position | Must equal the extreme (bull=low, bear=high) | No requirement |
| Body / range | ≥80% (small wick allowed on opposite side) | ≥95% (almost no wicks) |
| Opposite wick | Small wick allowed (e.g. small upper wick on bullish) | Both sides shaved |
| Core meaning | One side dominates the open | One side dominates the entire bar |
| Reversal weight | Location-sensitive (strongest at trend ends) | Primarily a continuation signal |
| Win rate | 60-70% as reversal (with confirmation) | 55-65% as continuation |
Key rule: whether the open equals the extreme is the deciding factor. A perfect marubozu is always a belt hold, but a belt hold is not always a perfect marubozu. Belt hold = 'open dominance'; marubozu = 'whole-bar dominance'. On Binance Futures, belt holds at trend ends have the strongest reversal meaning, while marubozu inside a trend acts as continuation.
Key rule: R/R ≥1:2, ideal 1:3. Multi-timeframe confluence (daily trend end + 4H pattern + 1H entry) lifts win rate from ~65% to 80%+.
Bullish belt hold sweet spot: RSI on K1 at 25-35 (oversold), reverting to 40-55 on K2 close. Bearish mirror: K1 RSI 65-75 (overbought), K2 breaks below 50. RSI exiting extremes aligned with belt-hold direction is one of the strongest Binance Futures reversal combos.
Bullish belt hold: MACD histogram printed its largest green bar on K1 and starts shrinking; K2 area approaches or just crossed a golden cross. Bearish mirror: max red bar → shrinking → dead cross. MACD is the gold-standard read for whether reversal momentum is real.
K1 (the belt hold itself) must trade 1.5~2.5× average; K2 confirmation candle ≥1.2×. Low-volume belt holds usually morph into fake reversals — win rate collapses without volume.
Highest probability setups appear where belt hold prints at: 1) daily EMA200; 2) Fibo 0.618/0.786 retracement or 1.272 extension; 3) prior swing high/low; 4) trendline of the daily/4H structure; 5) extreme funding rate (longs >0.08% or shorts <-0.05%). Pattern + key level + indicator + confirmation candle = four-factor confluence, the safest reversal entry on Binance Futures.
In 2026 after a 5-day BTC decline, the 4H near EMA200 printed a textbook bullish belt hold: K1 open = low, +4.1% body covering 87% of range, no lower wick, minimal upper wick, volume 2.3×. RSI rebounded 27→48; MACD histogram clearly shrinking. K2 +2.3% closed above K1 close without breaking K1 open, volume 1.4×. Long entered at K2 close, stop 0.7% below K1 open. BTC subsequently rallied 9.5%; all three target tranches hit, R/R 1:3.2.
After an 8-day ETH rally into a prior high, the daily printed a bearish belt hold: K1 open = high, -5.3% body covering 90% of range, no upper wick, volume 2.4×. RSI fell 75→52; MACD widened into a dead cross. K2 -2.8% closed below K1 close without breaking above K1 open, volume 1.5×. Short entered at K2 close, stop 0.8% above K1 open. ETH dropped 11.5%, all three target tranches hit.
Common DNA: 1) pattern at trend end on a key level (EMA200, prior high/low); 2) strict K1 criteria — open=extreme, body ≥80% range, ≥1.5× volume; 3) K2 continuation without breaking K1 open; 4) RSI/MACD synchronously confirming; 5) entry only after K2 close — never chase K1 mid-bar.
Mastering Belt Hold plus strict risk control lets you catch reversals at the very first candle, with K2 confirmation and multi-factor confluence as the trade-off for that early read. Combined with engulfing, harami, three outside, three inside, hammer, shooting star and Three Line Strike from this series, you build a complete Binance Futures candlestick trading system.
No. Belt Hold requires the open to equal the candle's extreme (open=low for bullish, open=high for bearish) and allows a small wick on the opposite side. Marubozu requires near-zero wicks on both sides (body ≥95% of range). Every perfect marubozu is a belt hold, but not every belt hold is a perfect marubozu. Belt hold = 'open dominance', marubozu = 'whole-bar dominance'.
Not recommended. Standalone belt holds win only 55-65% of the time due to fake setups. Wait for K2: bullish requires K2 close above K1 close without breaking K1 open; bearish requires K2 close below K1 close without breaking K1 open. With K2 confirmation, win rate lifts to 75%+.
Strict definition: open=low (truly no lower wick). In practice a tiny lower wick (≤5% of body) is tolerated — the core is 'buyers dominated from tick one, no probing lower'. If the lower wick is significant (>10% of body) it's closer to a hammer or long-legged doji, not a belt hold.
Binance perps have no traditional gap, but candle-period boundaries (UTC midnight, 4H/1D rollovers) function as a 'new candle open'. Belt holds at these boundaries carry the strongest meaning. Event-driven moves (funding settlements, Binance announcements, macro data) can produce gap belt holds with doubled strength.
Daily and 4H combined. Below 1H, fake belt holds dominate because the open extreme gets probed away easily. Daily signals are rare but very precise. Swing traders use daily+4H confluence; intraday traders use 4H pattern + 1H entry, treating K2 close as the trigger.
Critical. K1 must trade 1.5~2.5× average volume to count as a valid belt hold — one side dominating the open requires real capital backing it. Low-volume belt holds are almost always false signals. K2 confirmation also needs ≥1.2× average for a high-quality entry.
Three failure modes: 1) K2 fails to continue direction (immediately breaks back through K1 open); 2) K2 closes against the expected direction; 3) The pattern appeared inside a range, not at a clear trend end. The moment K2 breaches K1 open, exit at the preset stop (K1 open ±0.5~1%). Never average down hoping it works.
Download BNApp and register Binance with code BNAPP to enjoy maximum fee rebates.
Register Now · Download APK · BNAPP
Register with code BNAPP for permanent fee rebates