Long/Short Ratio (LSR) is the count of accounts holding longs divided by the count of accounts holding shorts on a perpetual futures contract. An LSR above 1 means most traders are positioned long; below 1 means most are short. The level on its own does not predict reversals, but extreme readings almost always sit close to the moments when the crowd is most one-sided and most exposed.
What it is — the ratio of accounts long to accounts short. Exchanges compute it from live positions and refresh every 5–15 minutes.
Why it matters — LSR shows the direction of the crowd's bias — something neither Open Interest nor funding rate can show by construction. Extreme readings cluster near local tops and bottoms.
How to use it — only with confluence. LSR alone is never a trade, it is a stretch gauge. See how Moami's AI team reads LSR inside a position's context.
and short counts balanced
on BTC and ETH perps
— short-squeeze risk
across major venues
What's inside
- What Long/Short Ratio actually is
- How LSR is calculated
- Three flavors of LSR — and why it matters
- 5 signals you can read at a glance
- 4 price-and-LSR scenarios
- 3 ways disciplined traders use it
- Common mistakes
- LSR vs funding and Open Interest
- When LSR lies
- Pre-trade checklist
- Frequently asked questions
What Long/Short Ratio actually is
Long/Short Ratio is the number of accounts currently holding a long position on a contract divided by the number of accounts holding a short. If BTC perp shows an LSR of 2.1, there are 2.1 long accounts for every short account. This is not a balance of capital — it is a balance of participants. The metric answers "how many heads are on each side", not "how many dollars".
This is the core difference between LSR and Open Interest. OI counts contracts, and every contract has both a long and a short side, so long-OI and short-OI are mechanically equal. You cannot say "longs outnumber shorts" through OI at all. LSR looks at heads instead of contracts, and that is what makes it a direct read on crowd sentiment.
Every major venue publishes it. Binance ships Global Long/Short Ratio Accounts, Top Trader Accounts Long/Short Ratio and Top Trader Positions Long/Short Ratio. Bybit and OKX expose similar slices. Aggregators stitch them into a single chart. The same event often produces different LSRs across venues — and that dispersion frequently leads price.
How LSR is calculated
The formula is almost trivial: LSR = (count of accounts long) / (count of accounts short). Exchanges snapshot open positions every 5–15 minutes and publish the ratio via their API. Binance exposes topLongShortAccountRatio, Bybit calls it account-ratio, OKX uses long-short-account-ratio. The number is recomputed for each contract independently.
One subtlety: a single account can be long on one pair and short on another. The exchange records each per-symbol net position separately. If a trader holds +1 BTC and −0.5 BTC on the same contract, the venue records them as long for that symbol. A trader long BTC and short ETH counts on both ratios — one each, in the relevant pair.
Another limitation: LSR only counts accounts with an open position. In a range, when most capital sits in cash, LSR can hover near 1.0 even when interest is depressed. Read it side by side with Open Interest: LSR up on rising OI means new longs are pouring in; LSR up on falling OI means shorts are closing rather than longs entering.
Different venues calculate LSR differently. Binance publishes an account-based ratio — heads — and separately a position-based ratio — total notional long divided by total notional short. They are two different metrics. Account-based shows mood, position-based shows where capital sits. They often diverge, and that divergence is a signal on its own.
Three flavors of LSR — and why it matters
Most venues publish several LSR variants simultaneously. Knowing the differences is half of using the metric properly.
| Type | What it counts | What it shows |
|---|---|---|
| Global LSR (Accounts) | All accounts with an open position, by head count. | Retail sentiment in aggregate. The loudest "crowd" reading. |
| Top Trader LSR (Accounts) | Top 20% of accounts by net equity, by head count. | Where profitable traders sit. Often leads price. |
| Top Trader LSR (Positions) | Notional size of positions held by top 20% accounts. | Where large capital sits. Slow but heavy signal. |
| Group divergence | Top traders on one side, retail on the other. | The strongest single setup — top traders are usually right. |
→ Scroll the table if it doesn't fit
5 signals you can read at a glance
- The level. Around 1.0 is neutral. 1.5–2.0 is a working bias. Above 2.5 on BTC, or below 0.5, is extreme and demands attention.
- Group divergence. Top Trader LSR > 2 while Global LSR < 1 means professionals and retail are on opposite sides. This is the densest single LSR setup.
- Rate of change. LSR jumping from 1.2 to 2.4 in an hour is a panic entry. A slow climb over a day is healthy accumulation.
- Price divergence. New price high while LSR falls = large capital quietly distributing. Sideways price with LSR drifting up = the crowd loading exposure ahead of a move.
- Cross-venue dispersion. LSR 2.3 on Binance, 1.4 on OKX — the two venues are seeing different flow. Typically the venue with larger OI ends up right.
4 price-and-LSR scenarios
| State | What it shows | What usually happens next |
|---|---|---|
| Price ↑ + LSR ↑ | Fresh longs are pouring in to confirm the move. | Healthy trend, but watch LSR not pushing past 2.5 — overheat starts there. |
| Price ↑ + LSR ↓ | Price rises while the crowd closes longs or shorts. | Distribution by larger capital. Often precedes reversal in 3–10 days. |
| Price ↓ + LSR ↑ | Price drops, retail piles into longs — "catching the falling knife". | Liquidation cascade ahead. On lower timeframes, fast continuation down. |
| Price ↓ + LSR ↓ | Capitulation: crowd flips and shorts the bottom. | Short-squeeze setup. Any positive catalyst triggers a chain of covers. |
3 ways disciplined traders use it
-
Counter-trend trade on extremes
Wait for Global LSR above 2.5 (or below 0.5) sustained for at least two hours. Confirm the bias on funding rate matching the crowded side (positive when longs are overloaded). Enter against the crowded side with a 2–4% take and a stop just past the most recent extreme. Size so a stop costs no more than 1% of equity.
-
Follow top-trader LSR divergence
When Top Trader LSR diverges from Global LSR by 30%+ for several hours, take the top-trader side. This is a slow signal that plays out over hours and days, not minutes. Take-profit at the next structural level, stop beyond the daily-frame entry zone of where the top traders accumulated.
-
Trend confirmation through LSR consolidation
In a strong trend, healthy LSR does not stretch to extreme. If price rises while LSR holds 1.3–2.0 for several days, that is a sign of calm accumulation without panic. Use these moments to add to existing longs, not as a standalone entry trigger.
Practical tip. Compare the current LSR to its 30-day percentile, not to abstract "norms". On SOL, 2.7 might be routine; on BTC it is already an overheat signal. The pair-specific context outweighs round numbers.
Let the AI team read LSR for you
Paste any open or planned position — Moami's AI analysts compare current LSR against its historical range, cross-reference funding rate, Open Interest and the liquidation map, and return an explanation in under 30 seconds — never a prediction.
Common mistakes
Treating LSR as position size. Account-based LSR counts heads, not dollars. A single whale holding a $50M long adds one count — the same as a hundred retail traders with micro positions. On altcoins the distortion is huge.
Shorting just because LSR is high. Strong trends can hold LSR above 2.5 for weeks. Bitcoin in early 2024 held an LSR around 2.8 while still rallying another 20%. Without funding-rate and price-structure confirmation, this becomes roulette.
Ignoring the top-trader vs retail split. Global LSR and Top Trader LSR tell two different stories. Reading only one is throwing away the most useful part of the metric.
Mixing data from different exchanges. LSR calculation differs across Binance, Bybit and OKX in their top-trader cutoff and account base. Comparing absolute values cross-venue without normalisation is charting an imaginary metric.
LSR vs funding and Open Interest
- Direction of the participant skew — something OI cannot show by construction.
- A retail-vs-top-trader split — proxy for "smart money vs the crowd".
- 5–15 minute refresh — much faster than the 8-hour funding clock.
- Price divergence — the earliest public tell of an upcoming reversal.
- Capital size per side — that's Open Interest and position-based LSR.
- Liquidation levels — that's the liquidation map.
- Holding cost — that's funding rate.
- Precise timing — even a textbook divergence may resolve tomorrow, not now.
When LSR lies
On thin altcoins. If a token has only a few thousand active accounts, an LSR of 3.5 may mean a dozen new traders on one side. Reading it as "market sentiment" gets you in trouble.
Right after a mass liquidation. A cascade clears one side and LSR jumps abruptly, but it reflects the event that already happened, not the structure that exists now. Wait 1–2 hours for it to stabilise.
During major catalysts. ETF approvals, regulatory shocks, listings — fundamentals override any sentiment metric. Fading a real catalyst on an LSR signal is the most expensive way to learn the lesson.
When exchanges quietly change the math. Once or twice a year venues update their top-trader thresholds or account base. Historical charts before and after the change can be incomparable. Check exchange changelogs if you rely on long backtests.
Pre-trade checklist
Before relying on LSR for a decision, run the list. If a single line is unchecked, stay on the sidelines.
- Stable for 2+ hours. A single LSR spike is noise. A stable reading or divergence over several hours is a structural signal.
- Percentile, not round number. Current value should sit in the top or bottom 10% of the 30-day range. A naked level without context is meaningless.
- Top Trader vs Global. Check both charts. Alignment = thin signal; divergence = the strongest standalone LSR setup.
- Confluence. At least one independent signal: extreme funding, a dense liquidation cluster, or weakening 4H momentum.
- Cross-venue check. Skew visible on 2+ exchanges with serious OI. A single-venue spike is local, not systemic.
- Clear invalidator. Know the price level that kills the thesis before you enter. Without one, an LSR trade becomes hope against a position paying you to be wrong via poor R:R.
Editorial note. Moami AI provides probabilistic analysis — explanations and scenarios, never predictions. LSR is one of dozens of signals our AI analysts cross-reference with risk, liquidity and price action to help you make decisions on data.