TL;DR. The long/short ratio number on Binance or Coinglass looks simple — "1.42" or "0.87" — but there are four different formulas behind it and they tell different stories. Top Trader Position Ratio (whale dollars), Top Trader Account Ratio (whale heads), Global Account Ratio (retail crowd) and Taker Buy/Sell (real-time flow). Read the wrong one and you flip the signal. This guide walks through all four, the BTC distributions for each, and how to stack the ratio with funding and open interest.
1. What the long/short ratio actually measures.
Binance perpetual pages and the CoinGlass Long/Short Ratio dashboard both display a "Long/Short Ratio." The literal meaning is "how much more something is long than short" — but the "something" differs by venue and method: account count, dollar value of positions, or 5-minute taker flow. The same coin at the same moment can read 1.2 / 2.4 / 0.8 across the three. When a reading looks wrong, check whether you mixed variants before blaming the number. New to Binance perps? Read our funding rate guide first.
2. The four mainstream long/short ratios.
| Name | Sample | Weighting | What it tells you |
|---|---|---|---|
| Top Trader Position Ratio | Top 20% Binance accounts by size | Dollar-weighted | Where whale money is actually pointing |
| Top Trader Account Ratio | Top 20% Binance accounts by size | One account, one vote | Whale headcount (easier to spoof) |
| Global Account Ratio | All Binance accounts with positions | One account, one vote | Retail sentiment thermometer |
| Taker Buy/Sell Ratio | Last 5 minutes of taker volume | Dollar-weighted | Real-time flow direction |
Top Trader Position Ratio is the highest-signal variant — top 20% of accounts by size, weighted by USD notional, refreshed every 5 minutes (Binance Top Trader Long/Short Position Ratio official page). One whale's $5M long outweighs 100 retail $50k longs. Hard to spoof because spoofing costs real capital. Binance app perps shows this by default; on Coinglass you must pick "Top Trader Long/Short Ratio (Positions)" manually.
Top Trader Account Ratio samples the same 20% but counts heads. Easier to spoof. Useful only for divergence with Position Ratio. Global Account Ratio is the retail thermometer — retail leans long overall, but the level drifts (its median was about 1.24 on Binance BTCUSDT over 5–26 September 2026, with a low of 0.85); only P95+ behaves contrarian. Use percentile, not raw number. Taker Buy/Sell Ratio is the only one measuring flow, not positions — last 5 minutes of aggressive market orders. Useful as live confirmation: a rising candle on taker 0.9 is a bull trap; a falling candle on taker 1.4 is usually a fake-out.
3. Top Trader vs Global: why the gap.
Open Coinglass during a hot rally and the four ratios often disagree. A common combination: the Global Account Ratio runs well above its usual range (retail headcount leaning long), the Top Trader Account Ratio is only mildly long (whales, counted by heads, less one-sided), the Top Trader Position Ratio sits near 1.0 (whale dollars roughly balanced), and Taker Buy/Sell slips below 1.0 (aggressive selling outpacing aggressive buying). Read together: retail is excited, whales are not committing real money, and short-term buying is fading. That says the rally is running low on fuel, not that the top is in; often it is just a pullback before new highs. A real top usually also needs open interest at extremes and funding pinned high for days.
Why the spread? Each formula samples a different population. Position ratio captures dollars in a small whale set; account ratio captures heads in the same set; global ratio captures retail headcount; taker captures aggressive flow. So when you see a single "long/short ratio" figure in a CoinDesk or The Block headline, ask which variant before reacting.
4. How to read the number in four steps.
For any L/S chart, run this loop. (1) Which variant? Binance app defaults to Top Trader Position; the web client lets you toggle; Coinglass keeps your last pick. Forum screenshots usually do not label — always ask. (2) Where does this sit historically? Normal ranges differ by variant and drift with the market, so there is no fixed threshold table worth memorising. From Binance's public futures data, between 5 and 26 September 2026 BTCUSDT's Top Trader Position ratio stayed between 1.83 and 2.52 (median about 2.12), while the Global Account ratio ran between 0.85 and 1.83 (median about 1.24). Apply the often-quoted "2.0+ is extreme" rule for Top Trader Position and every day of those three weeks was extreme.
So pull the 90-day distribution for the specific symbol and variant (Coinglass has it; Binance's own endpoint only goes back 30 days) and place today's reading inside it: P25–P75 is normal, P90+ is hot, only P95+ is extreme. Altcoins run much wider than BTC, so never borrow BTC's ruler. (3) Are variants aligned or diverging? Aligned (all long or all short) = consensus = contrarian flavor, but risky to fade. Divergent (whales vs. retail) = the trade worth watching; whales historically have a small edge (background: Investopedia: long-short strategy primer). (4) Does it line up with funding and OI? Save that for the three-way stack in section 7.
5. Three real cases — Aug 2024 crash, Nov 2024 ATH, Mar 2024 melt-up.
BTC/USDT daily: open $58,161, intraday low $49,000 (−15.8%), close $54,018. According to CoinDesk's same-day report, liquidations across crypto futures topped $1 billion, with roughly 87% of liquidated traders on the long side, as the yen carry-trade unwind cascaded. Here is what the indicator can and cannot tell you in a move like this. It shows you which side is crowded: an elevated Global Account Ratio (retail chasing longs) next to a low Top Trader Position Ratio (whales leaning short) is the classic whale-vs-retail divergence, and when an outside shock arrives (here, the yen unwind) the crowded long side is the one that gets liquidated in a chain. What it cannot tell you is when the shock comes or how far price falls; a divergence can sit there for days while price keeps rising. Treat divergence as more useful than the absolute level, and as a reason to cut leverage, not a precise timing trigger.
BTC daily close rose from $69,496 (Nov 1) to $104,464 (Dec 15), a +50.32% leg (ETH +57.64%, SOL +35.02%). What to look for in a leg like this: if the dollar-weighted Top Trader Position Ratio keeps sliding while the Global Account Ratio climbs on media buzz, that is the classic distribution signature, whales handing inventory to late buyers. Lesson: a dollar-weighted ratio stuck near the bottom of its own history during a rally suggests the "real" demand engine may no longer be running.
2024-03-10 close $68,956 → 2024-03-14 high $73,777 (the all-time high at the time) → 2024-03-19 close $61,937 (−16% from the high in 5 sessions). The pattern to watch in a move like this: Global Account Ratio (heads) in a high percentile while Top Trader Position Ratio (dollars) sits in a low one — a scissor between retail enthusiasm and whale caution. On altcoins the scissor tends to be wider than on BTC. The wider the scissor, the more fragile the top. Combine with the liquidation map.
6. Four common misreadings.
"L/S ratio above 2 means short." Check the variant first. Top Trader Position Ratio at 2.0 is usually a trend-following signal: whales are 2:1 long with real money, so do not fade it.
"Use L/S ratio as a direct entry trigger." It is a thermometer, not a button. Entry needs price action (support/resistance, volume, pattern) and position sizing. The ratio is one filter, not the trigger.
"Coinglass alone is enough." Coinglass aggregates and may default to the wrong variant. Use Binance app as raw reference, Coinglass for multi-venue cross-check (OKX, Bybit).
"L/S ratio and open interest are the same." Different dimensions. L/S says direction; open interest says leverage water level. A 1.0 ratio on all-time-high OI is a coin flip on direction at record-fragile leverage — the smallest catalyst sets off cascading liquidations.
7. L/S ratio + funding + open interest — the three-way stack.
On its own the L/S ratio is a weak signal; stacked with funding and OI it tells you far more. The matrix:
| Global L/S | Funding | OI | State | Bias |
|---|---|---|---|---|
| ≤ P25 | Negative | Low | Capitulation | Scale into longs (with technicals) |
| ~ P50 | ±0.01% | Median | Normal chop | Trend-follow |
| ≥ P90 | +0.05% + | All-time-high | Long leverage maxed | Reduce, tighten stops, no new longs |
| ≥ P95 | +0.08% sustained 24h | ATH | Mania | Half-size, consider probing shorts |
| ≤ P25 | +0.03% | — | Squeeze setup | Short squeeze risk |
Three dimensions, three jobs: L/S = direction crowding; funding = sentiment cost (see funding rate guide); OI = leverage water level. All three at extremes = the most dangerous moment; all three at floors = reversal odds accumulating. A routine worth keeping: once a week, pull 7 days of three-way data, mark "three-way P90+" and "P10−" windows, then walk the price chart against them. Keep it up and you build a real water-level intuition. SOP in our weekly indicator checklist.
8. FAQ.
Is the long/short ratio a contrarian indicator?
Depends on the variant. Global Account Ratio (retail headcount) does behave contrarian at extremes. Top Trader Position Ratio (whale dollars) is usually trend-following. Do not treat all variants alike.
What is the difference between Top Trader and Global ratios?
Top Trader samples the largest 20% of Binance accounts; Global samples all accounts. The first reflects "money," the second reflects "headcount." They diverge often, and historically whales have a small edge when they disagree with retail.
Position ratio or account ratio — which is better?
Position ratio. Account ratio gives one vote per account regardless of size; position ratio weights by USD notional, which is closer to how money actually moves price. Binance shows both — start with position.
Is Taker Buy/Sell the same as long/short?
No. Taker is real-time aggressive flow; L/S is open position state. Taker is water velocity, L/S is water level. Used together they confirm each other.
Why do Binance and Coinglass disagree?
Binance shows Binance-only data; Coinglass often aggregates multiple venues or shows a different variant by default. Match symbol, time, and formula before comparing.
Is a ratio of 2.0 high?
Depends on the asset and the variant, and there is no fixed threshold. From Binance's public futures data, between 5 and 26 September 2026 BTCUSDT's Top Trader Position ratio stayed between 1.83 and 2.52 while the Global Account ratio ran between 0.85 and 1.83 — two variants, very different numbers, and both drift with the market. Pull the 90-day distribution for that pair and variant, see which percentile today's reading sits in (P90+ hot, P95+ extreme), then judge.
Can the L/S ratio predict tops and bottoms?
Alone, weakly. Stacked with funding (P90+) and OI (all-time-high) it identifies leverage-maxed regimes where pullback risk is meaningfully elevated. See section 7.
What to read next.
Crypto assets are volatile and not suitable for every investor. This page is editorial analysis, not financial advice.