Report archive · 18.08.2026
Case No. 20260818 · 18.08.2026
Case № 20260818 · 18.08.2026 08:18 UTC · BTC $64,149 / ETH $1,896
analysis: Fable 5 by Anthropic on our own archive since February
Futures (Hyperliquid). Whales stand short BTC: $516M against $208M long — 2.5 times more. That isn't news, that's the background.
The news is the pace. Over the week the net shifted $92M into short, over three days another $53M — and over the last 24 hours it went $46M into LONG. The weekly short isn't broken, but for the first time this week it exhaled.
ETH tells the same story: short is 3.7 times the long, and over the day the net shifted $26M into long.
Leverage at the top eased a little: wallets running 8× or more number 56 against 69 yesterday, most of them short — and none is pinned to a margin call (yesterday there was one). The dry fuel under the market hasn't gone anywhere, but the one standing on a thread has left the top.
Options (Deribit). Two separate positions. Quiet accumulation: over the week, without blocks, calls were collected at a strike far above the market with a September core, and puts far below — both sides, calls ahead; over three days the emphasis moved to puts right next to spot. Off-block trades: over two days, 37 structures in BTC and 9 in ETH — the biggest in BTC sells a ceiling just above the corridor for this week, the biggest in ETH plays a corridor into tomorrow's expiry. No single direction; types, expiries and sizes are in the full version.
On Bybit separately: we see two different things there and don't merge them. Retail accounts on perps are the crowd — they're in the witness block below. The Bybit options book is a separate cut, and we don't know who stands behind it, so we hang no labels on it.
Money (the Hyperliquid border). Over the week $50.7M came in and $63.1M went out — net minus $12.4M. What leaves is mostly USDC cash through the bridge, and almost all of it lands on an exchange — in even tranches, the latest overnight. Coins through Unit move the other way: more ETH and SOL came in than went out, while BTC was carried out to a market maker. Who exactly deposited the cash, and whether a whale from our registry is among them — in the full version.
This is positioning, not a direction forecast.
What the options market says. This is where you see what insurance against a price move costs. On the weekly tenor it sits at the 3rd percentile of our hourly history since February — cheaper than in 97 hours out of 100 since we started measuring (Deribit). Protection against a fall costs more than a position for a rise (BTC skew +4.0; ETH's tilt is smaller), but both are cheap.
The level below which market makers stop damping the move and start amplifying it dropped 198 points over the week and rose 211 over three days — the market makers won back a little ground from below, but spot stands only just above it. This confirms the whales: the market prices any move cheaply, there's more fuel above — and over the day the perp whales began handing back their short.
What the futures say. When there are more longs, they pay the shorts, and the other way round — that payment is called funding. On Hyperliquid right now shorts pay longs (−0.010% per eight hours), on Binance it's the reverse at +0.003% — both numbers hug zero. This camp is silent: whale positions are one-sided, but the payment for them is near zero on both sides — no pressure.
What the crowd is doing. These are retail traders on Bybit — thousands of small accounts. In BTC 57% of them are long, in ETH 68%; Hyperliquid whales are short at the same time. The gap between how the crowd stands and how the whales stand is, in BTC, average for the month and has barely moved over three days; in ETH it's wider than on 83% of the month's days, but it's standing still too. This contradicts the whales: the crowd is on the opposite side, and in our registry it's the crowd that's wrong more often. But the event is the gap's MOVE, and there's no move.
Any signs of stress. We check whether anyone's already being forced out of positions. Over the day, among the biggest Hyperliquid wallets — not one forced close. On the wider market (OKX swaps) $2.2M closed over the week, mostly longs, and the last 24 hours added only $161k — 7% of the week, flat background. This camp is silent: the fuel is there, nothing's lighting it.
Cross-check in sum: options side with the whales, futures and stress are silent, only the crowd stands against — and it's standing, not moving. This is positioning, not a direction forecast.
The classic index prints 48.7 against 51.6 yesterday — middle of the scale, neither altseason nor bitcoin season (90-day window). Our barometer shows a gap the classic doesn't see.
Alts hold 40.7% of perp open interest ($3.65B), and almost 60% of that volume sits on hot funding — retail is paying to be long alts.
Retail in the top-10 alts is 68.2% long — strongest in XRP (80.5%) and DOGE (77.5%).
Hyperliquid whales hold $651M in alts and only 27% of it long; the biggest position is HYPE, and that's mostly short. Whales are structurally short — the signal will be in the CHANGE of the gap, not its size. Full table: indiciadesk.com/en/altseason
Continuation ⟳. The upward tilt is alive: fuel above is growing, insurance is close to the cheapest since February, and the perp whales' weekly short exhaled for the first time over the day. The squeeze we gave 30% on 16.08 was confirmed by code. But the ceiling above the corridor stands, and until it's broken on volume this isn't a trend — it's a corridor with a tilt.
Forecasts this week: 12 — 5 confirmed · 1 partly · 6 not.
What this does NOT mean: it isn't an invitation to go long and it isn't a promise of another squeeze. Retail is already 57–68% long — the crowd is on that side. The call wall above the corridor and the market's biggest wall in September (Deribit; levels and contracts in the detailed cross-check) damp the move up, and half of the week's scenarios didn't come true — that's written above.
🔒 In today's full RADAR: the fund that started cutting both its BTC shorts over the day · the structure of the day — a sold ceiling above the BTC corridor: expiry, legs, size.
ANALYST access → indiciadesk.com/en/agent · 3 days free
💎DEEP
Portraits of the day (Hyperliquid, snapshot 13:16 UTC).
The biggest BTC short. $115.4M at 40× leverage, entered at $63,991 — uPnL now −$262k. Liquidation at $64,854, 1% from price, on a $2.8M account: he's held it a week, and every half percent up is a question of survival for him.
Two 5× shorts from the same hand. The first — $72.7M, entered at $63,237, uPnL −$1,020k, liquidation $102,389 (+60%): CUT by 170.2 BTC over the day. The second — $61.5M, entered at $62,955, uPnL −$1,132k, liquidation $105,843 (+65%): a week ago the position didn't exist, and over the day it also shrank by 70.5 BTC. Low leverage, a big loss, cutting — not adding. Who this is — below, in the Analyst section.
Two more 40× on a thread. A $64.1M short from $63,973 with liquidation at $65,096 and a $32.8M short from $63,582 with liquidation at $64,662 — both 1% from price, uPnL −$164k and −$284k. The second cut its position by 1,280.6 BTC over the week — so it's already trimming. Together with the biggest short, that's three positions that burn in the $64.7–65.1k zone — right under the $64,500 call wall and just before the short fuel begins.
The market's deepest loss. An ETH short of $94.8M at 3× from $1,700 — uPnL −$9,827k, held for at least a week, liquidation $2,190 (+15%). Neither cutting nor adding.
The only long in ETH's top-6. $28.5M at 25× from $1,905, uPnL −$121k, liquidation $1,582 (−17%). One against five shorts.
The biggest gain. An ETH short of $8.5M from $2,122, 20× leverage, uPnL +$1,014k — held for at least a week.
BTC $72.7M −$1,020k cut 170.2 BTC over the day BTC $61.5M −$1,132k new this week · cut 70.5 BTC ETH $65.7M −$541k was there a week ago ETH $41.8M −$295k liquidation $3,847Fasanara Capital — an ETH short of $63.8M, 15× leverage, entered at $1,885, uPnL −$397k, liquidation $2,300 (+21%), account $16.7M. The same address appears below in the money — it was topping up cash at the border. An ETH short of $36.0M at 3× from $1,789 with uPnL −$2,039k — nick Bennett; a BTC short of $41.2M at 15× from $63,581 (uPnL −$357k, liquidation $66,976) — nick pablogv from OpenSea. Both self-labels.
Structures with a breakdown (Deribit, two days).
Over two days the detector recognised 37 off-block structures in BTC and 9 in ETH. The biggest, with expiry and legs (net premium isn't in the snapshot):
• bear call spread on 2,000 BTC to 21.08 — sold the $66,000C, bought the $67,000C. The position earns if BTC stays under $66,000 by Friday; the maximum loss is capped by the bought call. The sold ceiling stands just above the upper edge of the market makers' corridor.
• call backspread on 600 BTC to 25.12 — two prints of 300: sold the $68,000C, bought twice as many $78,000C. Positioned for a STRONG rise into December — it loses on a soft climb into $68–78k and wins on a shot above.
• condor on 2,000 ETH to 19.08 — bought the $1,840P, sold the $1,880P and $1,920C, bought the $1,960C. A position on the $1,880–1,920 corridor into tomorrow's expiry — "selling the quiet" for a single day.
• bear call spread on 1,500 ETH to 25.12 — sold the $2,800C, bought the $3,200C: someone's selling ETH's rise above $2,800 to year-end, with the risk capped at $3,200.
• a three-leg structure on 255 ETH to 25.09 — sold the $2,100C, bought the $1,800P, sold the $1,600P: insurance against a fall to $1,600, paid for by selling the rise above $2,100.
Cases in play (our registry).
Case №546 · strangle on 7,500 ETH to 25.06.2027 (Deribit). Put $1,000 and call $4,000, entered at spot $1,806. A position on a big move either way over ten months — for now spot stands in the middle.
BTC 7d $70,000 calls +4,017 (64% on 25.09) $67,000 calls +2,255 (90% on 21.08–04.09) $60,000 puts +3,335 (11 expiries) BTC 3d $63,500 puts +1,025 (4 expiries) $60,000 puts +918 (87% on 21–28.08) $64,000 puts +812 (10 expiries) ETH 7d $2,000 calls +13,275 (9 expiries) $1,700 puts +12,832 (60% on 21.08) $2,050 calls +11,687 (86% on 28.08) ETH 3d $2,000 calls +6,811 · $1,950 +6,065 · $2,050 +5,340 (96% on 28.08)
A whale from the registry was topping up cash. Fourteen USDC tranches of $0.5–2.0M, $12.85M together from 12 to 15.08 — from address 0x7fdafd…, the same one that's signed Fasanara Capital in our whale registry and holds the $63.8M ETH short at 15×. That's margin under the short. The last tranche was 15.08, none new over three days.
Wintermute (market maker). BTC — three tranches of 44–45 BTC to its address on 13, 14 and 17.08, $8.51M together; ETH — $4.8M withdrawn to it, SOL — $1.61M. Coins came in mostly from Binance: ETH $12.1M, SOL $4.54M. Over the last 24 hours — $3.23M USDC to Coinbase and $2.79M in BTC to Wintermute.
The weekly shift of BTC whales on Hyperliquid into short (the numbers are in THE GIST) was ACCELERATING until yesterday: more than half of the week's shift landed in the last three days — and over the day it went the other way. It's a reversal of pace, not of position: the net short of $308M stands. In ETH the weekly shift is tiny ($6M into short), and the daily one went into long: whales are handing back there too.
BTC's gamma flip stands at $63,139 — its move over the week and over three days is described in the cross-check above; it now sits a little below spot. ETH — $1,840, down 17 over the week.
The fear premium (VRP) is growing: plus 0.7 over the week, plus 0.3 over three days — with the cheapest vol since February the market still overpays for protection relative to how it actually moves. Structures over the week: 108 — 27 bullish, 22 bearish, 18 on vol, 41 on range — no tilt, the biggest a call backspread on 3,000 ETH. Over three days 29 (8/7/6/8), the biggest a bear call spread on 2,000 BTC.
Changing of the guard. Yesterday's "whale on a thread" — a $98.7M short at 40× with the liquidation right at price — is no longer in BTC's top-6: there's no position of that size there any more, and what became of it the data doesn't show. In its place a fresh $61.5M short at 5× from Abraxas entered the top — and it's already cutting.
Max pain and the market makers' corridor (Deribit). For 19.08 BTC max pain is $64,000 (Deribit $64,000, Bybit $63,750) — spot stands just above; the market makers' corridor is $63,000–65,000, spot in the middle. By our registry of 306 expiries, an exact match with max pain happens about a third of the time — the magnet is weak. The week's walls: put wall $63,500 (566 contracts, 457 of them on 19.08) — support; call wall $64,500 (406, 187 of them on 19.08) — resistance. On 25.09 max pain is $70,000 and the call wall sits at the same level with 11,350 contracts — the market's biggest wall; the 25.09 put wall is $60,000 with 4,926.
ETH: corridor $1,840–1,920, max pain $1,880 — and both the week's put wall and call wall stand at that same $1,880 (1,455 and 1,032 contracts): the market is pinned to a single point. On 25.09 max pain $2,000, call wall there too — 29,312 contracts.
OI structure. BTC — 338k contracts, a third on 25.09 with a core at $70,000, another 30% on 25.12 with a core at $80,000. The biggest strikes: $70,000 (22k, 87% calls) and $60,000 (21k, 82% puts) — top and bottom are defined. ETH — 1,546k, split evenly between 25.12 ($3,200) and 25.09 ($2,000); the biggest strike $2,500 — 98k, 95% calls.
PCR in time. BTC put/call over the month 0.46, over the week 0.57, now 0.56 — downside protection was built up over the month, held over the week. ETH from 0.54 over the month to 0.51 now — disarming against a fall. Two books up to 30 days diverge: on Deribit BTC 0.65, on Bybit 1.05 — relatively more puts on Bybit; who's on the other side there, we don't know.
Skew by tenor (Deribit). BTC 2d P26/C24, 7d P30/C27, 30d P36/C32 — puts pricier on every tenor, and the whole curve sits at the 8th, 3rd and 10th percentile since February. ETH 7d P39/C38 — 2nd percentile. Cheap everywhere; no shock is priced on two days, and monthly vol is only slightly dearer than weekly.
Gamma and the shot zones (our calculation from the Deribit book). BTC's flip at $63,139: at −2% ($62,954) price breaks through it and market makers switch to amplifying the fall; at +2% ($65,523) it's still under the big $70,000 call wall, but the local $64,500 one is already passed. Under spot it's empty at $60,000–61,000 — a break there accelerates. ETH at −2% ($1,863) is still in positive gamma — the dip gets bought; empty at $1,800–1,850.
Fuel (HL cascade map, 06:46 UTC). Above spot, $286.2M of short positions liquidate in the $65,676–67,607 zone; below spot, $24.8M of longs in $61,169–63,100 — the same tilt as in THE GIST, now in zones. ETH is the same in miniature: $15.0M below, $2.9M above.
VRP. The premium is +8.4 (implied 34.6 against realized 26.1 over the week) — the market pays more for protection than it actually moves; on the daily window the premium is wider still.
🆓TAIL
Pendulum · Wheel — standing position: +$251 on the day · +$6,567 in total · counted since 11.05.2026.
signal trades win rate last total since Dust Strategy V2 35 6% −$1,125 +$95,219 21.03 Dust Strategy V1 29 10% −$1,125 +$39,281 20.03 Volatility Convergence V3 15 47% −$195 +$8,248 26.05 Skew 2.0 V2 27 37% −$191 +$3,952 26.05 Flat Wings 4 75% +$801 +$1,744 11.05Deepest in the red: Wheel Trail V8 −$4,962 · Skew 2.0 V3 −$4,630 · Fear Flash V1 −$4,078 — all three hunt a reversal that compressed vol never delivered. The most profitable signal wins only 6% of its 35 trades: win rate isn't money.
The 16.08 scenarios reached their deadline — the code's verdicts:
• BTC "The corridor holds" (from 2026-08-16, we weighed it at 45%): PARTLY — inside the range 62% of the time (close 64,225)
• BTC "Squeeze up into the short fuel" (from 2026-08-16, we weighed it at 30%): CONFIRMED — high 64,519 against the 64,300 level
• BTC "Drift below the gamma flip" (from 2026-08-16, we weighed it at 25%): CANCELLED AT THE BOUNDARY — high 64,519 against the 64,000 level
Read it this way: the scenario with the smaller weight (30%) came true, the heaviest (45%) only partly, because spot was inside the range only 62% of the time. We didn't guess the distribution of weights — and that entry stays in the registry.
Yesterday's verdict said: "continuation — the week's upward tilt is alive, but the $64,500 ceiling holds". Over the day BTC +0.4% — it rose, but weakly, below the threshold we count as a move; a week ago the same conclusion gave +1.0%. The ceiling wasn't broken — that confirms "holds", not "tilt".
The 17.08 scenarios (corridor 62,000–64,500 / squeeze 64,657 / drift 62,765) stay open until the morning of 19.08. Over the week (11–17.08), forecasts: 12 — 5 confirmed · 1 partly · 6 not (cancelled at the boundary). Since 08.08, when the code started counting: 21 scored — 8 confirmed · 1 partly · 12 not.
Confirmation by type since 14.06 (210 checked): "relative" 71% of 210 and "ETH range" 66% of 150 — reliable. The "BTC range" type came true in 36% of 150 cases — worse than a coin toss, the system doesn't use it; the same for "BTC smart-money trace" at 33% of 75.
Not investment advice and not a recommendation to trade. Derivatives trading carries a high risk of loss; INDICIA DESK is a research publisher, not a licensed adviser under 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.