Report archive · 02.09.2026
Case No. 20260902 · 02.09.2026
Case № 20260902 · 02.09.2026 08:19 UTC · BTC $77,354 / ETH $2,415 · analysis: Fable 5.1 by Anthropic
Futures (Hyperliquid). The story here isn't the side, it's the tempo. The group of big accounts stood short in both assets and still does, but the last day did more than the whole week before it: in BTC the daily shift into short was almost ten times the weekly one — the week nudged the position one way in small steps, and yesterday's session added a sharp block on top. ETH shows the same pattern, only softer: a large weekly move into short, a noticeably smaller daily one. Leverage in the group eased slightly at the same time: 60 overleveraged whales under watch against 64 yesterday. Less fuel, but still mostly long fuel.
Options (Deribit). Two positions, and today for the first time this week they point the same way — both sell the top. First — the block trades: ten new structures in a day, four of them sell a ceiling, and not yesterday's ceiling but one a thousand dollars lower. Second — the quiet accumulation: over the week the biggest growth in BTC open interest gathered once again on far September calls above $80,000, meaning small lots are stacking the very thing the blocks are selling. That's not a contradiction: some sell the right to a move up, others buy it — where they agree is that this boundary is the one worth trading. In ETH the quiet trace is unambiguously put-side: all three of the week's biggest increments are puts.
Money (the Hyperliquid border). The border tape last updated on 01.09, so the window here is the seven days to 01.09, not to today. Across the bridge it was dollar cash that moved, and it moved outward: nearly twice as much withdrawn as deposited, and 85% of the whole outflow landed on exchange deposit addresses. Coins through Unit went strictly the other way: four times more BTC came in than left, and ETH, SOL and ZEC all net positive on the way in. The picture repeats for a fourth straight week: cash leaves the venue, coins arrive on it.
This is positioning, not a direction forecast.
What the funds are doing. This camp measures what neither Deribit nor Hyperliquid can see: how much money institutions put into the coin itself through spot ETFs. Funds trade on business days only; the freshest data is for Tuesday 01.09 — there are no figures for today yet.
In BTC the week brought a $253M inflow against $2,083M the week before — an eightfold slower pace.
Tuesday itself already came in as a $236M outflow, and nearly all of it fell on one fund, BlackRock's IBIT; Fidelity added the rest of the outflow, and only Bitwise was buying.
In ETH the pace slipped too but stayed high: $628M for the week against $890M the week before, and that's the 91st percentile of the whole weekly history — even the reduced pace beats almost every week we've seen. Here the inflow was spread across several funds with no clear leader — unlike bitcoin, where one fund decided everything.
This confirms the whales: in bitcoin, institutions moved for the first time this week in the same direction the perp whales added over the day. In ETH the camp still stands on the opposite side — and that's where the gap between money and positions is widest today.
What the options market says. This camp measures not the mood in the news but the price of insurance: how much protection against a move costs. Our fear and greed index stands at 87 out of 100 — exactly where it was yesterday. BTC monthly vol is meanwhile cheaper than in 81% of the hours since February, and the ETH picture is the same. This camp is silent: options price in neither a rise nor a fall — they price in quiet, for a third straight day, even though on the two other arenas everything changed within this one.
What the futures say. This camp measures who pays whom for the right to hold a leveraged position (what funding is). BTC funding in the eight-hour equivalent slipped from +0.0115% yesterday to +0.0107% today — longs still pay shorts, just a little less than a day ago. The basis is positive and thin; in ETH it's effectively zero. This confirms the whales: the big group stands short against a crowd that pays every day for the right to stay long, and that payment didn't rise over the day — there's no pressure for a reversal.
What the crowd is doing. This camp is the retail perp accounts on Bybit: where the mass of small players stands. In BTC it shifted to 55% long against 52% yesterday; in ETH to 66% against 63%. The crowd got more bullish over the day precisely when the big accounts added to short.
This confirms the whales: the gap between the mass and the big accounts in BTC widened to 32.5 points — wider than on 97% of the days in the last month. It's a monthly high, and it's been widening for three straight days.
Any signs of stress. This camp measures whether positions are already being broken by force. Among the top Hyperliquid whales none was force-closed over the day. On OKX swaps the day closed $547k, 97% of it longs; the week's volume is $5.2M, and there the long share is noticeably lower.
The last day gave 11% of the weekly volume — an even background with no acceleration. This camp is silent: there's plenty of fuel beneath the market, no ignition. The tail of cascades has historically been a reversal zone, but on our sample that's unproven, so it stays an observation rather than a signal.
This is positioning, not a direction forecast.
💎DEEP
A wallet self-signed "VBVIT". SHORT BTC $64.4M at 10× leverage, entry $78,172, uPnL +$475k, liquidation $98,400 — 27% above spot, a wide margin of safety. Account $19.2M. The position already existed a week ago and has been built up by 107 BTC since: not a fresh entry, a top-up of an old position.
A wallet with no public label. SHORT BTC $31.0M at 20×, entry $77,796, uPnL +$78k, liquidation $92,831 — 20% above spot. Account $8.9M. A week ago there was no position at all: this is an entirely new short, opened inside the week.
A wallet with a public label (name below, in the ANALYST section). LONG BTC $38.4M at 40× and LONG ETH $99.5M at 25×. In bitcoin the entry is $77,374, uPnL +$112k, liquidation $70,593 — 9% below spot. In ETH the entry is $2,443, uPnL −$857k, liquidation $2,337 — only 4% below spot, the tightest liquidation among the venue's big positions. Account $5.9M across both. This is the most squeezed big long on the market.
A wallet with no public label. LONG ETH $73.3M at 20×, entry $2,134, uPnL +$8.7M, liquidation $1,751 — 28% below spot. Account $6.5M. The position already existed a week ago and has grown since.
The biggest gain among the open positions. LONG ETH $48.4M with entry $1,936 and just 4× leverage, uPnL +$9.7M, held for at least a week. The contrast with the previous portrait speaks for itself: the same side of the market, the only difference is when they entered and with how much leverage.
Case №1025 · double purchase at one strike, BTC — bought 810 puts at $75,000 with term 30.10.2026 and bought 810 calls at $75,000 with term 27.11.2026. Both legs bought, so the whale collected no premium — he paid it: ≈ −110.5 BTC (≈ −$8.54M). The day's biggest structure by money and the most expensive of the whole week.
Case №1028 · bear call spread BTC — sold 150 calls at $79,000, bought 150 calls at $85,000, term 25.09.2026. Net premium in the whale's favour ≈ +2.52 BTC (≈ +$195k).
Case №1029 · bear call spread BTC — sold 150 calls at $80,000, bought 150 calls at $84,000, term 25.09.2026. Net premium ≈ +1.71 BTC (≈ +$132k).
Case №1030 · call ratio spread BTC — sold 200 calls at $81,000 and bought only 100 calls at $85,000, term 25.09.2026. Net premium ≈ +2.81 BTC (≈ +$217k). Half as many calls bought as sold, so the upper boundary here isn't fully closed.
Case №1023 · put spread ETH — bought 2,000 puts at $2,200, sold 2,000 puts at $2,000, term 11.09.2026. The whale paid ≈ −10.6 ETH (≈ −$26k) — for that money he gets protection, or a position, on the $2,000–$2,200 range over the next nine days.
Together the four structures that sell the top (№1028, №1029, №1030 and one more on 150 BTC) brought the whales ≈ +9.2 BTC (≈ +$715k) of premium up front. Three of them sell the $79,000–$81,000 boundary on the September expiry — a thousand dollars below the one sold yesterday. Standing apart are two diagonals on 750 and 400 BTC, where whales bought near calls for Thursday and sold far ones for October: that's a position on speed, not on direction.
Whale flow on perps. BTC: a $9M shift into short over 7 days, of which $82M came in the last day. The numbers don't contradict each other: the week ran toward long, and the final day wiped out that whole move and added on top. ETH: $93M into short over 7 days, another $21M in the last day — here the week's direction and the day's coincide, and the tempo is accelerating too. Aggregate tilt: BTC $376M short, ETH $247M short.
Gamma flip shift (what it is). The BTC flip sits at $66,824 and rose 2,253 over 7 days; the put wall pulled up by a full 17,000 over the same stretch. The ETH flip is at $2,302, its put wall up 320 over the week. When support pulls up that sharply while the flip crawls, it means the market makers are moving their defence higher behind price rather than holding the old level.
Fear premium trend. The premium in this window is −0.9 (IV 37.2 against actual HV 38.1) and it's melting at 5.3 per 7 days. Options are still cheaper than the market's real movement: the regime of underpriced protection holds for a second week, but the gap is closing.
Open-interest growth by window. BTC over 7 days: $85,000 +5,837 calls (80% on 25SEP26), $70,000 +5,219 puts, $82,000 +4,014 calls (74% on 4SEP26). ETH over 7 days: $2,200 +19,384 puts, $2,400 +13,293 puts, $2,000 +10,911 puts. The contrast between the assets is total: in BTC the upper calls were stacked quietly; in ETH it was puts and nothing else at all three biggest strikes.
Changing of the guard. Over 7 days the registry logged 228 confirmed structures: 63 bullish, 94 bearish, 20 on volatility, 51 on range. The biggest by size — a call diagonal on 6,000 ETH. The week's overall lean is bearish, and today's ten fit into it: four sales of the top against two diagonals and one large purchase of volatility.
Max pain drift. On the nearest expiry the point moved +9,000 over 7 days — the flow is dragging the consensus up, and doing it faster than spot itself is moving.
Levels by expiry (Deribit and Bybit books combined into one; our study across 356 expiries showed price matching max pain exactly only about a third of the time, so it's a reference point, not a magnet — what max pain is)
· BTC 03.09 — max pain $77,750 (Deribit $78,000 · Bybit $77,500) · put wall $76,000 (352 contracts) · call wall $81,500 (149 contracts)
· BTC 25.09 — max pain $70,000 (Deribit $70,000 · Bybit $75,000) · put wall $70,000 (8,651 contracts) · call wall $70,000 (10,972 contracts)
· ETH 03.09 — max pain $2,430 (Deribit $2,440 · Bybit $2,430) · put wall $2,200 (1,156 contracts) · call wall $2,500 (1,003 contracts)
· ETH 25.09 — max pain $2,150 (Deribit $2,150 · Bybit $2,200) · put wall $2,100 (38,081 contracts) · call wall $3,000 (40,910 contracts)
The market makers' break-even corridor: BTC $76,500–$79,500, ETH $2,360–$2,500. Spot sits inside both. The key point: the $79,000 strike sold today stands inside the bitcoin corridor, not on its upper edge — yesterday's $80,000 stood on the edge. The whales are no longer selling the corridor's ceiling; they're selling the space beneath it.
Put/Call through time. BTC: month 0.53, week 0.60, now 0.56. ETH: 0.52 / 0.55 / 0.54. In BTC the weekly value sits above both the monthly and the current one — downside protection was being stacked more actively a few days ago than it is now; in ETH the ratio is practically motionless.
Vol by tenor in percentiles (against our own hourly history since February)
· BTC: 2-day 25th percentile, weekly 13th, monthly 19th
· ETH: 2-day 25th, weekly 20th, monthly 24th
In BTC the cheapest term is the weekly one: that's where the market prices in the least movement relative to its own norm. It's also the term that carries this week's nearest large expiry.
Gamma levels. The regime is positive on both assets: BTC flip $66,824, ETH flip $2,302 — spot is well above both, so dealers work to damp moves. The hollows with no hedge: BTC $72,000–$73,000, ETH $2,300–$2,350. The ±2% stress test shows both assets still inside the damping zone: at −2% BTC tests the densest protection, at +2% it runs into the ceiling.
Fuel beneath the market and above it (Hyperliquid liquidation heat map, snapshot 08:16 UTC, index $77,438 and $2,418)
· BTC: $33.4M of longs in the $73,566–$75,889 band, $45.3M of shorts in $78,987–$81,310 — more above the market than beneath it for the first time this week.
· ETH: $109.7M of longs in $2,297–$2,370 against $121k of shorts in $2,466–$2,539.
The day's main change is right here. Yesterday ETH showed $20.3M against $10.0M — a twofold asymmetry to the downside. Today the long fuel is five times bigger and the short fuel has all but vanished. And that fuel lies exactly in the $2,300–$2,350 hollow where dealers hold no hedge: two independent measurements point at the same band.
This is a relief map of risk, not a forecast: liquidations lag price, they don't lead it.
Fragility index. The same group of overleveraged whales holds $729.8M of positions at 8× and above, 69% of that money stands long, and none of them is anywhere near a margin call yet. Meanwhile DVOL is 37 — vol is cheap, so the market isn't pricing this cascade risk at all. It's "fuel, no ignition": not a timing, a state of the structure.
🆓TAIL
signal trades win last total since Dust Strategy V2 37 5% −$1,125 +$92,969 21.03 Dust Strategy V1 31 10% −$1,125 +$37,031 20.03 Volatility Convergence V3 22 45% −$507 +$14,608 26.05 Skew 2.0 V2 36 36% −$631 +$11,839 26.05 Fear Flash V2 21 33% −$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). They share one trait: these are signals that chase continuation of a move, and this market hasn't offered continuations for a second month — in a damping regime, where dealers hold price inside a corridor, that logic systematically pays for entry and never gets the move. The day was nearly flat: +$7 over 24 hours, and that illustrates the regime better than any single signal. The most profitable signal, Dust Strategy V2, wins only 5% of its 37 trades — and still leads the table in money. Win rate isn't money.
Yesterday the system judged the market "risk tilted down". Over the day BTC moved −1.1% — price went that way, but softly, below the threshold at which we count a confirmation. A week ago the conclusion was "no clear tilt", and BTC then moved −2.1% over the day — neutrality didn't hold, there was a move after all. One conclusion over the week partially held, the other didn't: a normal frequency for an analysis of positioning, a poor one for a forecast. We don't do the second.
The scenarios opened on 30.08 were closed by the code today. "The corridor holds under max pain" (weighed at 45%) confirmed: price never left the range for the whole term and finished at $77,921. "Burn-up through the short fuel" (33%) not confirmed — the high of $79,110 fell short of the $79,800 level. "Slide into the long fuel" (22%) not confirmed either — the low of $77,478 stayed above the $76,600 level. Separately, yesterday's "Burn-up through the short fuel" (22%) was cancelled at the boundary within a day: price touched $76,899 against a $77,000 boundary. Two scenarios from 31.08 and two from 01.09 close over the coming days.
Score by forecast type (the journal's whole history, computed by code)
· "Burn-up through the short fuel" — 4 of 12 confirmed, 2 cancelled at the boundary
· "The corridor holds" — 3 of 10, 2 more partial, 3 cancelled at the boundary
· "Squeeze up into the short fuel" — 4 of 10, 4 cancelled at the boundary
· "Slide under the gamma flip" — 4 of 10, 4 cancelled at the boundary
· "Slide into the long fuel" — 1 of 3, 1 cancelled at the boundary
The weakest type is "Burn-up through the short fuel": 4 confirmations of 12 issued — a third, worse than a coin toss. We neither hide it nor drop it, because in two cases of twelve it wasn't wrong but was cancelled at the boundary before its term — but today's 32% has to be weighed with exactly that score in hand.
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.
Radar is assembled automatically from live Deribit / Hyperliquid / OKX data · time-stamped snapshot. A journal of the system's decisions, not investment advice. © 2026 INDICIA DESK.