They held the short through a 6% rally and started cutting it on a quiet weekend nightthe asymmetry is in the timing, not the size.
Case #20260920 · 20.09.2026 05:47 UTC · BTC$80,507 / ETH$2,577 · analysis: Opus 5 by Anthropic
📅YESTERDAY
Saturday's 19.09 expiry settled at 08:00 UTC — and the market-maker break-even corridor failed in both assets at once. Both settled above the corridor by roughly 3.3%. From max painBTC diverged 5.6% and ETH6.3% — that is how the registry counts it, measuring the levels at settlement time. Against the morning levels we published yesterday the gap was half that: 3.9% in BTC and 3.0% in ETH. It is the first day since 4 September when both assets broke the corridor; across the whole registry since February there have been 22 such days.
Case registry: none new over the day, 725 closed with a score, same as yesterday. One old case a whale took apart himself, without waiting for expiry.
⚡THE GIST
• The whales cut their BTC short for the first time in a week: the net tilt slimmed from $423M to $366M, with $57M moving to the long side over the day.
• In ether the move is the opposite: the whales' long side slimmed by $66M, and the short-to-long tilt grew from 2.2 to 2.5 times.
• Insurance costs something again: the premium over the actual move in BTC is +3.5 against +0.9 yesterday, and the extra charge for downside protection is gone — skew−0.7, with calls slightly dearer than puts.
• The crowd and the whales converged: the BTC gap narrowed by 15 points over three days, to 21.9 — that already counts as an event by our threshold.
• Verdict — break ⚡ in bitcoin: the direction of positioning changed for the first time in a week, even though ether still pulls short.
• 📊 Our Polygon — Pendulum: standing position, $0 on the day, +$9,642 in total since 11.05.
• 📊 Our Polygon — signals: best — Dust Strategy V2 +$86,219 at a 5% win rate · worst — Wheel Trail V8 −$8,787. Table at the bottom.
🐋WHALES: WHAT CHANGED — three arenas
Arena one — Hyperliquid futures.Bitcoin and ether went different ways. In BTC the short side slimmed over the day from $661M to $628M while the long side grew from $238M to $262M: the ratio fell from 2.8 to 2.4 times (how we track whale positions). The cutting started in the second half of Saturday and ran all night. In ETH it is the other way round: the short side slimmed by $48M, but the long side by $66M, so the tilt grew from 2.2 to 2.5 times. Overleveraged whales number 62 against 59 yesterday, with longs at 55% of them. Not one sits at the edge of a forced close this time — yesterday there was one.
Arena two — Deribit options. The weekend: the block market is close to asleep, and the quiet footprint changed sides. Over two days the snapshot recognised only 22 structures in BTC and 6 in ETH against 84 and 40 the day before (what a block is). The biggest in BTC is a structure on puts around the current price: the whale keeps the premium if the market doesn't slide lower. Next to it two structures on a capped rise. Quiet accumulation in small lots outside the blocks stays call-side above the market in BTC, but the week's third-biggest footprint is already a put below the market, and nearly all of it sits in the quarterly. In ETH the footprints go both ways: calls above the market and puts below it. The registry took in no new cases over the day.
Arena three — money at the exchange border. The week ended with an inflow, and it is mostly ether. Over seven days $102.5M of ether and $74.9M of USDC cash came into the exchange, while $70.1M of BTC left against $24.8M brought in. With all coins together, $216.3M came in against $129.9M out. The border snapshot refreshes once a day and stops on the morning of 19.09, so Saturday itself is not in it yet.
This is positioning, not a direction forecast.
🧭CROSS-CHECK — five camps of witnesses
What the funds are doing. Spot ETFs are the only camp that measures purchases of the coin itself, not a leveraged position. Funds don't trade on weekends: the latest figures are Friday's, 18.09, when $433M went into the BTC funds and $144M into the ETH funds. The week of 14–18.09 closed almost flat in BTC, +$6M against −$288M the week before; in ETH it came to −$140M against +$224M. This camp is silent: there will be no new figures until Monday, and Friday's buying is older than Saturday's move.
What the options market says. This camp measures one thing: what insurance costs on Deribit. BTCskew−0.7 against +1.6 yesterday: the extra charge for downside protection is gone, and calls now cost slightly more than puts (what skew is). In ETH the tilt is deeper, −1.4. Insurance itself got dearer: the premium over the actual move on the monthly window is +3.5 in BTC and +3.0 in ETH — yesterday both sat near zero. BTC two-day vol is in the 26th percentile of its own hourly history since February against the 13th yesterday (what implied vol is). Our fear-and-greed index stands at 89 out of 100 against 94 in yesterday's snapshot. This confirms the whales: the options side stopped paying for a fall exactly as the whales began cutting their bitcoin short.
What the futures say. This camp measures who pays whom to hold a position. BTC funding on Hyperliquid is +0.010% in eight-hour terms, the same on Binance; in ETH+0.010% against +0.013% yesterday (what funding is). Longs pay shorts in both assets. This confirms the whales: demand for a leveraged long is holding, and in bitcoin the whales are moving the same way.
What the crowd is doing. This camp measures how retail accounts stand against the whales. On Bybit perps 52% of BTC accounts stand long and 65% in ETH, both almost unchanged from yesterday. The gap between the crowd and the whales in BTC is 21.9 points against 25.6 yesterday: over three days it narrowed by 15 points — more than its usual move, so the beacon counts that as an event. In ETH the gap is 35.1 points, in the 87th percentile of the month. This confirms the whales: the two sides are converging from both ends — the whales are cutting short and retail isn't adding long.
Any signs of stress. This camp measures whether anyone has already been forced out of the market. Among the top Hyperliquid whales, not one forced closure over the day. On OKX swaps $126k was liquidated — 1% of the week's volume. This camp is silent: the market was repriced without a single cascade.
Again: this is positioning, not a direction forecast.
🌡ALTSEASON — the classic and our barometer
The classic altseason index stands at 43.2 out of 100 against 35.6 yesterday — the highest since 19 August, but this is still bitcoin territory: the alts caught up while BTC itself slid.
Our barometer watches money, not prices: open interest, funding and retail come from Bybit perps, the whales from the Hyperliquid registry. Alts' share of open interest is 43.3% — $5.10B of $11.78B, almost unchanged.
Alt contracts with funding hotter than bitcoin's make up just 3% of open interest against 74% yesterday. The bar here is BTC's own funding, and it jumped over the day: it is the bitcoin long that now costs more.
Retail in the top-10 coins stands 65.1% long; the hottest are DOGE and XRP, at about 77% each.
Whales in alts are structurally short: 30.9% long across $1.41B of positions, the biggest of them HYPE. That gap itself is permanent, the signal is in its CHANGE — over the day it barely moved.
Full barometer: indiciadesk.com/en/altseason
⚖️VERDICT OF THE DAY
break ⚡ — the direction in which the whales' bitcoin tilt had been moving changed for the first time in a week: the net short shrank by $57M while the long side grew for a third day running. Alongside that the options market stopped paying extra for downside protection, and our two independent sources of measurement on the whales — perps and options — are looking the same way for the first time in days. Witnesses: options, funding and the crowd side with the whales; the funds and stress are silent.
What this does NOT mean. This is not a claim that bitcoin goes up. $57M was cut from a $423M tilt — one day and 13% of the position; ether meanwhile pulls the other way. The test will be Friday's quarterly on 25.09: 189,467 contracts in BTC and 768,075 in ETH.
🔒In today's full RADAR:
— who exactly was cutting the short and who was adding: two addresses of one house haven't moved their position for a third day, while one player flipped sides in bitcoin over the day;
— the case a whale took apart himself after 83 days in play — what he bought in June and with what move in price he walked away.
ANALYST access → indiciadesk.com/en/agent
💎DEEP
🔒WHALES IN DETAIL 💠
Portraits from the Hyperliquid registry, snapshot 04:46 UTC. Nicknames aren't treated as identification.
The market's biggest loss.ETH SHORT $265.4M, leverage 5×, entry $2,297, uPnL −$29.0M against −$32.8M yesterday, liquidation at $3,439 — 33% above spot. The coin count hasn't changed for a third day: the size and the loss move only with price. The same address in BTC. SHORT $169.7M, leverage 5×, entry $72,731, uPnL −$16.2M, liquidation at $122,658 (+53%). Not a single new coin here either. A second address with the same pattern.ETH SHORT $185.2M and BTC SHORT $109.7M, both at 5× leverage, uPnL −$19.5M and −$12.4M, liquidations at $3,691 and $139,718. Together the two addresses carry $77.0M of unrealised loss against $85.6M yesterday, and margin across the two accounts grew from $159.9M to $178.6M. Who flipped. The address holding a 30× BTC short of $54.1M from $79,991 stood long a week ago; over the day it added 34BTC. Liquidation at $152,496 — 90% above spot, so the buffer is wide. The same address holds an ETH SHORT of $56.3M at 25× from $2,407 and grew it further over the day. Who was cutting. The address with shorts in both assets cut its ETH short to $55.9M over the day, from $62.9M yesterday, but in BTC it went the other way and grew the position to $95.2M at 20× leverage — average entry $79,105, liquidation at $91,661, only 14% above spot. Account $16.7M. The biggest bitcoin long. LONG $107.2M, 1,333 BTC at 3× leverage from $78,057, uPnL +$3.1M, account $123.1M — a second day without changes. The long out of the rally went into the red. LONG $86.0M, 1,070 BTC from $80,606 at 20×: yesterday it was $434k in the green, now $238k in the red. Liquidation at $63,832 (−21%), account slimmed from $8.7M to $8.0M. The ether longs.ETH LONG $78.1M at 20× from $2,134, uPnL +$13.4M — the market's biggest gain. The risky ETH LONG at 25× shrank over the day from 33,875 to about 25,900 coins, to $66.8M; the average entry rose from $2,514 to $2,524, so the position was also being added to before the cut. uPnL +$1.4M, liquidation at $2,522 — only 2% below spot, account slimmed from $5.4M to $3.7M. Fragility.62 overleveraged whales against 59 yesterday, $1.09B of positions together, 55% of them longs. Not one sits near a margin call this time.
Cases in the registry — none new over the day; one old one a whale took apart himself. Case #388 · BTC multi-leg structure: at the end of June the whale bought 25 calls and 25 puts at $52,000 and sold 50 calls at $75,000 for Friday 25.09.2026 — 100 contracts. On 19.09 the registry recorded the position dismantled: open interest at its strikes fell 41%, and price travelled +36.8% from entry. The whale left six days before expiry, on day 83.
Профіль виплат на експірацію 2026-09-25. Кит заплатив $272k премії. Беззбитковість: $41 104 і $62 896 і $87 104.
Case #1091 · ETH bear call spread: 1,000 calls at $2,660 sold, 1,000 calls at $2,700 bought, expiry tomorrow, Monday 21.09.2026 — 2,000 contracts, entered at $2,618. Net premium about 2.1ETH received (≈$5.5k). Spot $2,577 sits 3% below the sold strike, so for now the structure works for the whale.
Профіль виплат на експірацію 2026-09-21. Кит отримав $5k премії. Беззбитковість: $2 665.
Case #1092 · BTC call diagonal: 100 calls at $70,000 bought for Friday 25.09.2026, 100 calls at $72,000 sold for Friday 30.10.2026 — 200 contracts, entered at $80,978, second day. Both legs are deep in the money, net premium about 1.12 BTC paid (≈$91k).
Case #948 · ETH condor for the quarterly: 2,750 puts at $1,700 bought, 2,750 puts at $1,900 sold, 3,000 calls at $2,800 sold, 3,000 calls at $3,000 bought, expiry Friday 25.09.2026 — 11,500 contracts, entered at $2,462, day 26. Net premium about 36.8ETH received (≈$91k). Spot sits inside the winning zone of $1,900–$2,800, five days from expiry.
Профіль виплат на експірацію 2026-09-25. Кит отримав $91k премії. Беззбитковість: $1 867 і $2 830.
Case #1080 · BTC risk reversal: 100 calls at $70,000 sold and 100 puts at $70,000 bought, expiry Friday 30.10.2026 — a synthetic short from $70,000, premium about 8.22BTC received (≈$624k), fifth day. From the $75,860 entry spot has moved 6.1% against this position, and over the day the gap narrowed a little.
Профіль виплат на експірацію 2026-10-30. Кит отримав $624k премії. Беззбитковість: $76 236.
ΔOI over the week, both ways:BTC+3,658 calls at $82,000 and +2,922 at $85,000 against +2,444 puts at $74,000, 83% of which sit in the quarterly and the 02.10 expiry. ETH+18,204 calls at $2,800 and +13,486 at $2,600 against +15,289 puts at $2,500.
🔬Who exactly
Both of the biggest shorts are Abraxas Capital Mgmt (Heka Funds): for a third day the firm hasn't moved its positions in coins, and its unrealised loss shrank over the day from $85.6M to $77.0M as price slid. Margin across the two accounts is $178.6M.
The address that flipped from long to short in bitcoin and grew its ether short is Fasanara Capital: BTC SHORT $54.1M at 30× leverage and ETH SHORT $56.3M at 25×, account $54.1M.
The address that cut its ether short but grew the bitcoin one to $95.2M with liquidation 14% above spot (1,182 BTC against 1,077 yesterday) is self-labelled VBVIT; nicknames aren't treated as identification.
The risky 25× ether long, cut by about 8,000 coins over the day, belongs to Machi Big Brother: account $3.7M against a $66.8M position.
The biggest bitcoin long, the long out of the rally and the market's biggest gain are addresses without a public label.
Money by address. In the available snapshot the day's biggest transfers are small: $1.0M of ether from Binance and the same again from Wintermute. Over the week the coin channel sent players over $60M of BTC, while ether went the other way, into the exchange.
🔒THE WEEK'S DYNAMICS 🔬
Perps.BTC: net $366M short (long $262M against short $628M), $57M toward long over the day — a reversal inside a weekly window that still shows $50M toward short. ETH: net $429M short (long $285M against $714M), $19M toward short over the day, $154M over the week.
Gamma flip.BTC across all expiries $71,234 against $71,620 yesterday, spot 13% above it (what the flip is). ETH$2,248 against $2,308, spot 15% above. Over the week the call wall rose by $2,000 and the put wall by $5,500: dealers are carrying their protection up behind price.
The price of fear. On the weekly window the premium is minus 2.7 (DVOL35.7 against the actual weekly HV38.4) against minus 5.4 yesterday — the gap halved in a day. Options pricing is catching up with the actual move.
Changing of the guard. Over the week the whales assembled 241 confirmed structures: 63 bullish, 65 bearish, 30 on volatility, 83 on range — a fourth day with more bearish than bullish, but the edge is now inside the margin of error.
🔒CROSS-CHECK IN DETAIL 💠
BTC levels. Everything below comes from Deribit options open interest. The market-maker break-even corridor is $78,000–$81,000 against $72,000–$81,500 yesterday — three times narrower, and spot stands in its top quarter; by our study across 392 expiries price stays inside such a corridor in roughly 82% of cases (what the corridor is). Max pain on the combined two-exchange open interest is $80,000, on Deribit alone also $80,000 (what max pain is) — practically at spot.
Today, Sunday 20.09. The expiry is thin: BTC 3,436 contracts, put wall $80,000 on 203, call wall $82,000 on 367 (what the walls are). ETH 26,119 contracts, max pain$2,580, put wall $2,460 on 1,793, call wall $2,700 on 1,370.
Tomorrow, Monday 21.09.BTC 2,546 contracts, max pain$80,500; ETH 34,178, max pain$2,560 — the densest call wall stands at the same level, 3,302 contracts.
ETH levels. The corridor is $2,520–$2,620, spot a little above its middle; max pain$2,580.
The quarterly, Friday 25.09.BTC: 189,467 contracts, max pain$72,000 on Deribit and $73,000 on the combined open interest — 10% below spot; densest are the $70,000 put (8,612) and the $70,000 call (10,916). ETH: 768,075 contracts, max pain$2,200, put $2,100 on 36,844, call $3,000 on 36,606.
Fuel. Below BTC$50.7M of long positions in the $76,370–78,781 band against $39.5M of shorts above the market in the $81,997–84,408 band — over the day both sides grew and the tilt narrowed to 1.3x. In ETH$85.6M of long fuel below the market against $43.9M of short fuel above it. This is about the RANGE of a move for the same push, not about a higher probability of that move (what the fuel map is).
Gamma vacuum. In BTC it is empty above at $86,000–$88,000 and below at $74,000–$75,000. In ETH it is dense above spot and empty below at $2,500–$2,550.
Stress test ±2%. On a 2% fall both assets run into their densest put wall — the move is damped. On a 2% rise BTC passes the $82,000 call wall, and the damping on the way up weakens.
OI flow.BTC put/call across all expiries: month 0.56 · week 0.56 · now 0.56 — stable; ETH0.53 · 0.55 · 0.52 (what put/call is).
Two options venues. On the same expiries up to 30 days BTC put/call is 0.55 on Deribit against 1.15 on Bybit; ETH0.61 against 0.89 — over the day both Bybit sides turned less put-heavy. We don't know who owns the Bybit positions, so this is neither "whales" nor "retail" — just a second venue with a different lean.
IV by tenor.BTC 2d P30/C34 · 5d P34/C36 · 40d P36/C35, percentiles since February 26th, 23rd and 17th; yesterday the two-day sat in the 13th. ETH 2d P43/C46 · 5d P47/C51 · 40d P50/C51, percentiles 29th, 31st and 28th. BTCDVOL35.7 against 30-day realised 32.2 — premium +3.5; ETHDVOL51.1 with HV 48.1, premium +3.0 (what the premium is).
💎SCENARIOS IN DETAIL 💠
The code's verdicts for yesterday. From 18.09: the squeeze up through the short fuel was confirmed — a high of $81,840 against the $77,900 trigger; the corridor was not confirmed, price sat inside the band only 17% of the time. The three scenarios from 19.09 run until tomorrow. The week since 13.09: 21 scenarios — 6 confirmed, 2 partially, 3 not, 7 cancelled at their cancellation levels, 3 still running. All-time: squeeze up 10 of 30, corridor 9 of 26 plus 6 partial, slide into the long fuel 5 of 19 — the directional types do worse than a coin toss.
The corridor holds — 45%. ⏱ 2 days, to Tuesday 22.09. BTC stays between the fuel bands, $78,800–$82,000. Cancellation level: a touch of $76,350 — the bottom of the long-fuel band.
Squeeze up through the short fuel — 30%. ⏱ 2 days, to Tuesday 22.09. Trigger: a touch of $82,000 — the bottom of the band of short positions. Cancellation level: a touch of $78,800.
Slide into the long fuel — 25%. ⏱ 2 days, to Tuesday 22.09. Trigger: a touch of $78,800 — the top of the band of long positions. Cancellation level: a touch of $82,000.
This is how we weigh the odds — not a promise; the code will compute the verdict from spot history.
🔭WHAT TO WATCH
• Monday: the first fund figures after the weekend and the return of the block market — over two days the snapshot recognised only 22 structures in BTC against 84 the day before (💠).
• The 20× BTC short with liquidation 14% above spot: the address was building it all through Saturday (🔬).
• The premium over the actual move: from zero to +3.5 in BTC in a day — if it keeps rising, last week's cheap insurance is over (💠).
🆓TAIL
📊POLYGON
(our own signals at real prices, not a backtest)
⚙️Pendulum · Wheel(standing position, real prices)
+$0 on the day · +$9,642 in total · since 11.05.2026
⭐ Top-5 signals($1,000 per signal; "total" is the sum of all trades, not the return on a single thousand):
signal trades win last total since
Dust Strategy V2 435%−$1,125+$86,219 21.03
Dust Strategy V1 349%−$1,125+$33,656 20.03
Volatility Convergence V3 3040%−$759+$12,440 26.05
Skew 2.0 V2 4137%−$775+$10,606 26.05
Fear Flash V2 2232%−$418+$3,023 16.05
Deepest in the red: Wheel Trail V8 −$8,787 · Wheel Trail V6 −$7,718 · Wheel Trail V7 −$6,179 — three variants of one signal that hunts a reversal and pays for every attempt while the trend runs. Not a single trade closed in the top five over the day: at the weekend the signals stay quiet. The most profitable signal wins only 5% of its 43 trades. Win rate isn't money.
📅PUBLIC LEDGER
(computed by code at the settlement price)
The market-maker corridor. Across 392 expiries since February the market-maker break-even corridor held the settlement price 320 times — 82%, half a point less than yesterday. Saturday's expiry broke it in both assets to the upside, by roughly 3.3%. Decision: we lean on the corridor knowing that roughly one expiry in five it doesn't hold — and we publish every such case the same day.
Max pain. An exact match with the settlement in 36% of those same 392 expiries. Yesterday the two assets diverged from it by 5.6% and 6.3% as the registry counts it, and by 3.9% and 3.0% against the morning levels we published the day before: on a thin expiry the levels shift within half a day, so the settlement is measured against new ones. Decision: we don't chase max pain; a thin weekend expiry doesn't hold its point at all.
The walls. The call wall held the settlement in 49% of expiries, the put wall in 41%: roughly a coin toss. Today's Sunday expiry carries 3,436 contracts in BTC — the walls on it mean nothing. Decision: a wall without an expiry and a contract count isn't published.
Whale structures by type (our Deribit registry, 725 closed cases). Call diagonal — the type of two of the day's three BTC blocks: 51 closed, −$1.55M in total. Bull put spread — the type of the third: 80 closed, 21% came in, −$526k in total. Condor — the type of the big quarterly case: 36 closed, 69% came in, −$126k in total. All closed cases together: −$5.81M. Decision: win rate isn't money, and we don't copy whale structures — we track where they stand.
Polygon. Pendulum +$9,642 since 11.05 with $0 on the day. Not a single new trade in the top five over the day — the weekend. Decision: we publish both tallies daily with the same denominator, including the days when nothing happened.
📏Scale calibrator
• What protection costs. Our fear-and-greed index — 89 out of 100 against 94 yesterday. We compute it from the price of volatility and the option skew on Deribit, not from news, which is why it diverges from the well-known index.
• Cheap or expensive for this market itself.BTC two-day vol — 26th percentile of the hourly history since February, five-day — 23rd, forty days out — 17th: cheaper than usual, but no longer at the floor.
• Which way protection tilts.BTCskew−0.7: calls cost slightly more than puts — yesterday it was +1.6, on the put side. Terms: indiciadesk.com/en/glossary/
Not investment advice and not a recommendation to trade. Derivatives trading carries a high risk of loss; INDICIA DESK is a research publisher, not an adviser licensed by MAS, the SEC or the FCA — decisions and their consequences are yours.