Everyone sees the pricealmost nobody sees which side the whales just switched to.
Case № 20260903 · 03.09.2026 08:20 UTC · BTC$77,886 / ETH$2,403 · analysis: Fable 5.1 by Anthropic
📅YESTERDAY
The code closed the three scenarios opened on 31.08. "The corridor holds under max pain" (weighed at 45%) — confirmed: price never left the range for the whole term. "Burn-up through the short fuel" (32%) — not confirmed: the high stopped at 79,110 against the required 79,650. "ETH slides into the long fuel" (23%) — not confirmed, by two dollars: the low was 2,396 against a 2,394 level.
Separately, and uncomfortably: yesterday's "Burn-up through the short fuel" was cancelled at the boundary for a second day running — this time at 76,463 against a 76,500 boundary. We drew a conclusion and widened that boundary today; why exactly — in the review at the end of this issue.
⚡THE GIST
• On Hyperliquid perps the whales have been pulling BTC toward long for a second day, and the tempo now overrides the week: the daily shift is almost one and a half times the weekly one. The side is still short, but it's melting. In ETH the same group turned inside the window — the week pulled toward short, the last day already toward long.
• The options ceiling the whales sold yesterday got bought through today: a structure appeared in the Deribit book that buys a move above the very boundary sold a day earlier. The week's quiet accumulation points the same way — the biggest growth in open interest gathered on calls just above spot.
• Spot stands $614 below the upper edge of the market makers' break-even corridor. That's neither a ceiling nor a forecast: it's where the dealers' hedge thins out, and the same push produces a wider swing.
• ETH is more fragile than BTC today, and the number shows it: twelve times more fuel lies beneath the market than above it, and it lies exactly where the dealers' hedge disappears. In BTC the fuel advantage runs the other way — upward.
• The funds are braking for a second week: spot BTC ETFs took in $121.8M over the week against $1,797.8M the week before — almost fifteen times less. The freshest fund figures are for 02.09; they haven't reported today yet.
• 📊 Our Polygon — Pendulum: standing position, $0 on the day, +$8,641 in total since 11.05.
• 📊 Our Polygon — signals: best — Dust Strategy V2 +$92,969 at a 5% win rate · worst — Wheel Trail V8 −$7,088. Table with trade counts at the bottom.
🐋WHALES: WHAT CHANGED — three arenas
Futures (Hyperliquid). The group of big accounts stood short in both assets and still does — but the motion inside the position changed side. In BTC the week pulled moderately toward long, and the last day added almost one and a half times more than the whole week before it: the tempo is accelerating, and accelerating upward. ETH is the mirror image — the week pulled toward short, the day turned it. Leverage in the group eased a little at the same time: 58 overleveraged whales under watch against 60 yesterday, and 72% of their fuel still stands long.
Options (Deribit). Two positions, and today for the first time this week they look the same way — both up. First, the quiet accumulation: over the week the biggest growth in BTC open interest gathered on calls just above spot, two thirds of it on the September expiry, stacked in small lots outside the block tape. Second, the blocks themselves: yesterday a whale sold the ceiling; today a new structure in the book buys a move above it. One and the same boundary went from sold to bought within a day.
Money (the Hyperliquid border). Over the seven days to 02.09 $130.8M crossed into the venue and $174.4M crossed out — a net −$43.7M. It's mostly CASH moving, not coins: $150.9M of USDC left across the bridge, and 88% of that landed on exchange deposit addresses — money leaving derivatives for spot venues. Coins through Unit, meanwhile, stand level: BTC came in and left in almost equal amounts, ETH came in four times more than it left. The week's two biggest operations were deposits of $15.0M each into private wallets with no public label.
This is positioning, not a direction forecast.
🧭CROSS-CHECK — five camps of witnesses
What the funds are doing. The only camp that measures purchases of the coin itself rather than positions: how much money institutions put into spot ETFs. Over the week 27.08–02.09 BTC took in $121.8M against $1,797.8M the week before. ETH — $387.2M against $893.4M. Both sums are positive; both are sharply smaller than the previous week. Funds report with a lag; the freshest data is for 02.09, nothing for today yet. That last trading day in ETH actually came in as a $48.1M outflow, most of it from BlackRock's ETHA. This contradicts the whales: the perp whales are turning back toward long while institutional money brakes for a second week.
What the options market says. This camp measures one thing: how much the market is willing to pay for insurance. Our fear and greed index stands at 85 out of 100 — the same as yesterday and the day before, a third day without motion. BTC weekly vol is meanwhile cheaper than in 93% of the hours since February, while the market's actual movement runs stronger than what option prices assume. This contradicts the whales: the whales are rebuilding, and insurance is priced as if nothing were happening.
What the futures say. Here you can see whether the crowd pays for the right to stay long (what funding is). BTC funding settled at +0.0001% against +0.0107% yesterday — the cost of holding a long practically vanished within a day. In ETH it remains, +0.0033%; the basis in both assets is near zero. This camp is silent: long pressure hasn't reversed into short, it simply switched off.
What the crowd is doing. Retail perp accounts on Bybit stand long: 55% in BTC and 66% in ETH. The Hyperliquid whales stand the opposite way in both assets, and that gap by itself is structural — it always exists. The event is that over the day it began to narrow, and from the whale side, not the crowd's: retail didn't change its positions. This confirms the whales: they're the ones moving.
Any signs of stress. Stress is measured by how many positions the market closes by force. The last day gave 8% of the weekly liquidation volume — an even background, no acceleration; on Hyperliquid one large position was blown out over the day, mostly on the long side. Overleveraged whales are fewer: 58 against 60 yesterday. This camp is silent: the fuel is there, the ignition isn't.
Again: this is positioning, not a direction forecast.
🌡ALTSEASON — the classic and our barometer
The classic altseason index stands at 26.3 out of 100: bitcoin-season territory, money sits in the first coin.
Our barometer measures the same thing from inside Hyperliquid derivatives. Alts' share of perp open interest is 41.7% of $10.5B, almost unchanged on the day. Hot funding — the share of the market where longs pay noticeably — covers 48% against 54% yesterday: the heat is fading.
Retail in the top-10 coins stands 68.1% long, densest in DOGE and XRP at 77% each. Whales in alts, by contrast, are only 36.1% long across $1.15B of positions, and their biggest one is HYPE.
Whales are structurally short in alts almost always, so the signal isn't the tilt itself but the CHANGE in the gap: over the week their long share rose from 30.7% to 36.1%. The gap is narrowing from their side — just as in BTC. Details and charts: indiciadesk.com/en/altseason
⚖️VERDICT OF THE DAY
⟳ continuation. For a second day the whales are rebuilding in one direction on two arenas at once, and the rest of the market hasn't reacted: insurance stands still, the cost of a long vanished, the funds braked. The fragile point turned out not to be where everyone is looking: today ETH is the vulnerable one, not BTC.
Forecasts closed by the code over the week: 20 — 8 confirmed · 8 not · 4 cancelled at the boundary.
What this does NOT mean. It doesn't mean a promise of a move up. A positioning tilt is about whose side costs more to hold and what swing the same push produces — not about direction.
💎SCENARIOS
The corridor holds — 35% · Burn-up through the short fuel — 42% · ETH slides into the long fuel — 23%.
This is how we weigh the odds — not a promise; trigger and cancellation levels are in the full version.
🔒In today's full RADAR:
— the portrait of the fund sitting in the deepest paper loss of our whale registry and not cutting the position for a second week;
— the structure of the day: a whale built a position on tomorrow's expiry that cost him practically nothing.
ANALYST access → indiciadesk.com/en/agent · 3 days free
◦ The biggest short in the registry: BTC$148.6M at 5× leverage, entry $71,362, uPnL −$12.2M. Liquidation at $120,669, 55% above spot — margin isn't pressing at all. The position already stood a week ago and has only grown since.
◦ The same account also holds an ETH short of $150.3M from $2,159: uPnL −$15.2M, the deepest in the whole registry. Liquidation at $3,705, +54%. No pressure — stubbornness.
◦ The registry's biggest gain is a long: ETH$48.1M from $1,936, leverage only 4×, uPnL +$9.3M, liquidation at $1,017 (−58%). Held for at least a week, untouched over the day.
◦ The sharpest risk — an ETH long of $97.3M at 25× with liquidation at $2,347, 2% from spot. Over the last day this account added 4,400 ETH — it topped up precisely when liquidation was already close.
🔬Who they are
Both of the biggest shorts belong to Abraxas Capital Mgmt (Heka Funds) — two wallets of one structure hold about $502M of BTC and ETH short combined and sit on a paper loss of $53.9M. That's the fund that hasn't cut for a second week. The third-largest BTC short, $37.7M, is Fasanara Capital, and it added over the day. The sharpest ETH long at 25× is Machi Big Brother; the same account also holds a BTC long of $32.6M at 40×.
Case №1033 · BTC bull call spread, expiry 04.09.2026. The whale bought $82,000C and sold $85,000C, 100 contracts in each leg. Both legs went through at the same price of 0.0002 BTC, so the net premium is close to zero: a position on tomorrow's break above $82,000 cost practically nothing. That's +5.3% from spot, while the market prices a daily move of ±1.9%.
Профіль виплат на експірацію 2026-09-04. Кит заплатив $0 премії. Беззбитковості в показаному діапазоні немає.
Case №948 · ETH condor, expiry 25.09.2026, ninth day in play. Legs: buy $1,700P / sell $1,900P / sell $2,800C / buy $3,000C on 11,500 contracts, entered at spot $2,462. Net premium received, about 36.8ETH — roughly $88k. The whale is paid for ETH staying between $1,900 and $2,800 through late September; today's $2,403 sits near the middle of that corridor.
Профіль виплат на експірацію 2026-09-25. Кит отримав $91k премії. Беззбитковість: $1 867 і $2 830.
Two cases closed over the day, both in profit: №975 and №976, ETH bear call spreads. The thesis was "down", price did −0.2% over the term, and the whale collected $182 and $371.
🔒THE WEEK'S DYNAMICS 🔬
(our registry of top Hyperliquid accounts × the Deribit book, 7-day window)
The whales' net BTC position tilts $258M short (long $237M against short $495M). Over seven days a $90M net shift toward long, while the last day alone gave $129M: the tempo is accelerating, the side is consistent. In ETH the tilt is $227M short; the week shifted $67M toward short, but the last day gave $10M toward long — a turn inside the window, still weak.
The BTC gamma flip rose $1,834 over the week to $66,770 (what it is), and the put wall behind it moved $17,000 higher. When both the flip and the walls crawl after price, the market makers are ceding the level and room for a move grows; when they stand, the level is real.
The price of fear: the volatility premium is −1.1 (IV36.8 against actual HV37.9) and melting by 3.1 per week. Options are cheaper than the market really moves, and getting cheaper — a rare regime.
Changing of the guard: over the week the whales built 192 confirmed structures — 50 bullish, 70 bearish, 21 on volatility, 51 on range. The biggest is a risk reversal on 5,000 ETH. The group has no single position; the lean is mixed.
🔒CROSS-CHECK IN DETAIL 💠
On the nearest expiry, 04.09, BTCmax pain stands at $72,000 (Deribit $71,000 · Bybit $76,000), the put wall at $60,000 holds 1,087 contracts, the call wall at $82,000 — 6,482. On the September 25.09 expiry max pain drops to $70,000, and the densest call wall sits there too, $70,000 on 10,969 contracts. An exact match of max pain with the closing price happens in roughly a third of the 306 expiries we checked — a weak magnet, not an iron one (what max pain is).
The market makers' break-even corridor in BTC is $64,000–$78,500, and spot at $77,886 sits just under the upper edge. In ETH the corridor is $2,150–$2,460, and spot at $2,403 is in the upper third as well.
Put/Call in BTC moved like this: month 0.53, week 0.59, now 0.56 — downside protection was stacked a week ago and is being let go now. In ETH 0.53, then 0.56 and now 0.56: the build-up runs level.
Gamma separates the assets most sharply. The ETH flip is $2,363: at −2% from spot ($2,359) the regime turns negative, and dealers start amplifying a fall instead of damping it. The BTC flip is $66,770, 14% below spot; that cushion is large, and it holds today's whole difference between the two assets.
The fuel map confirms it. Beneath ETH lie $101.7M of longs in the $2,284–2,356 band; above ETH only $8.5M of shorts in $2,452–2,524 — a twelvefold asymmetry, and the lower band overlaps the gamma flip. Beneath BTC$27.5M in $73,981–76,318, above BTC$43.5M in $79,432–81,769: here the advantage runs the other way, upward.
💎SCENARIOS IN DETAIL 💠
⏱ 2 days, to 05.09.
1. The corridor holds — 35%.BTC stays inside $75,600–$79,400 for the whole term. Cancellation level: a touch of $75,000. History of the name: confirmed 3 times out of 10, twice more partially.
2. Burn-up through the short fuel — 42%.BTC touches $79,432 — the lower edge of the short-fuel band. Cancellation level: a touch of $75,600. History of the name: 4 of 14, 3 more cancelled at the boundary.
3. ETH slides into the long fuel — 23%.ETH touches $2,284 — the bottom of the fuel band, not its near edge. Cancellation level: a touch of $2,470. History of the name: 1 of 3.
This is how we weigh the odds — not a promise. The code will compute the verdict from spot history.
🔭WHAT TO WATCH
💠 ETH gamma flip $2,363. While spot is above it, dealers damp moves; below it they amplify, and the fuel band starts right behind it.
🔬 The upper edge of the BTC market-maker corridor, $78,500, and the $82,000 call wall, which holds 17,103 contracts across all Deribit expiries: between them the dealers' hedge is the thinnest on the whole map.
🔬 BTC funding. Within a day it fell from +0.0107% to +0.0001%; a return to clearly positive values would mean the crowd has followed the whales.
🆓TAIL
📊 POLYGON
(our own signals at real prices, not a backtest)
⚙️ Pendulum · Wheel: $0 on the day · +$8,641 in total · counted since 11.05.2026.
signal trades win last total since
Dust Strategy V2 375%−$1,125+$92,969 21.03
Dust Strategy V1 3110%−$1,125+$37,031 20.03
Volatility Convergence V3 2245%−$507+$14,608 26.05
Skew 2.0 V2 3735%−$392+$11,448 26.05
Fear Flash V2 2133%−$255+$3,440 16.05
What lost and why: deepest in the red are Wheel Trail V8 (−$7,088), Wheel Trail V6 (−$6,081) and Against the Current · top V1 (−$5,598). All three belong to the family that sells a move against the trend, and all three suffered from the same thing: the market went further than they assumed. The table's main number isn't in the money column but next to it: the most profitable signal wins 5 trades in 100 across 37 trades. Win rate isn't money, and that's exactly how this table should be read.
📅REVIEW OF PREVIOUS CONCLUSIONS
(verdicts are computed by code from spot history)
What we said on 31.08 and what price did. The stillness scenario ("The corridor holds under max pain", 45%) confirmed: BTC held inside the range for the whole term and finished at 77,513. Neither movement scenario confirmed. The upper one stopped at 79,110 against the required 79,650 — $540 short. The lower one in ETH fell two dollars short: the low was 2,396 against a 2,394 level.
Over the week the code closed 20 scenarios: 8 confirmed, 8 not, 4 cancelled at the boundary before their term.
What we changed today because of it. The name "Burn-up through the short fuel" was cancelled at the boundary two days running: on 02.09 at 76,899 against a 77,000 boundary, on 03.09 at 76,463 against 76,500. Both times price crossed the boundary by a fraction of a percent, while BTC's priced-in daily move is ±1.9%. In other words we placed the boundary inside ordinary daily noise and knocked ourselves out. Today the boundary is widened to $75,600 — about one and a half daily sigmas from spot.
Failing types, as they are. "Burn-up through the short fuel" confirmed 4 times out of 14 across its whole history — that's 29%, and three more times it was cancelled at the boundary. We weighed it at 32–42%, so we've been consistently overrating it. "Squeeze up into the short fuel" — 4 of 10, "Slide under the gamma flip" — 4 of 10: both near a coin toss, we don't treat them as standalone signals and they don't enter the verdict of the day. Only "The corridor holds under max pain" holds up — 3 of 3 — and that's exactly why we dropped that name today: spot stands well above max pain, and the wording would have been untrue.
📏Scale calibrator
• What protection costs. Our fear and greed index — 85 out of 100. We compute it from the price of volatility and the option skew, not from news, which is why it diverges from the well-known index.
• Cheap or expensive for this market itself.BTC weekly tenor — 7th percentile of the hourly history since February; monthly — 12th. Cheaper than almost always.
• Which way protection tilts.BTCskew−0.0, dead centre: no reliable tilt. ETH+1.4 — puts pricier than calls. 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.