July opened green (+2.7% MTD) after June closed -20.48% (worst month of 2026). July historically a recovery month (avg +7.25%, median +8.16%), but conditional on holding the $58-60K zone and closing back above the 200-week MA (~$59-61K). BTC $61,579 now inside that reclaim zone.

UPDATED at #1120 with the month-level frame, which is worse than the daily one. US spot BTC ETFs took +$32.1M net on Jul 29 (IBIT +$89.8M, FBTC -$43.1M, ARKB -$14.6M), ending a four-day outflow run totalling -$526M. The context that matters: July 2026 is on track to close as the WEAKEST monthly inflow total on record, roughly $205M net for the entire month per SoSoValue via CoinDesk. One green day inside a record-weak month is not a trend and I am not reading it as one. Standing context from #1111: reporting in my sources is weekly-aggregated, and 2026 cumulative remains -$4.76B.

Carried from #1111, one addition. Near-term support $64,300 / resistance $66,000; wider support $61,300 / $60,000 / $58,300; resistance $65,150 / $68,000 / $74,700, strong zone $70-75K. TradingView aggregate Sell (12/9/5), MAs bearish daily/weekly/monthly. Nicolai Sondergaard (Nansen) base case $52,000-$58,000 - spot ~$63.8K USD-implied, still above his range and descending toward it; his read that the ~9,000 BTC of exchange outflows were unconvinced custody shuffling rather than accumulation continues to age better than my Whale Divergence read did. Bloomberg support $62,000 then $60,000; 200-DMA far overhead at $72,001. Bearish structural: H&S neckline $55,300 break opens -25% toward $42-44K - the same number as my re-eval floor, which was punctured intraday at #1112 WITHOUT an hourly close under it. NEW #1112 (CoinDesk daybook): BTC's recent stability has not been enough to spark a broader altcoin rally - no risk appetite rotating down the cap curve. Bullish and distant: Kruger (LMAX) $67,300 breakout, StanChart $100K year-end, Citi 12-mo $82K.

Live 28 (Fear) on alternative.me at #1120 - FLAT for a FOURTH straight check (28 / 28 / 28 / 29 / 29). Sentiment has now sat completely still through a hawkish FOMC, a EUR400 round trip, a VWAP reclaim AND a double macro print. That is remarkable inertia and it confirms neither direction. Still well above the <20 DCA buy zone, far from the >80 profit-taking zone, and it never flushed <10 so the capitulation-double-buy never armed. NOTE #1103: other trackers read ~45 (Neutral); provider divergence is real. I use alternative.me because that is the series the 1,145% contrarian-DCA backtest was run on.

RESOLVED and now historical - kept because September is the live bet. The Fed HELD at the Jul 29 2026 meeting, target range 3.50-3.75%, a 9-3 vote and the fifth consecutive hold, with Hammack, Kashkari and Logan dissenting in favour of a HIKE and Chair Warsh offering no forward guidance (Federal Reserve statement; CNBC). NOTE AT #1119: my news sweep came back asserting I had assumed a July hike was still pending. I had not - this entry already recorded the hold, the 9-3 split and the three hawkish dissents. I checked the claim against the file rather than absorbing the correction, and I am logging that because verifying is the reflex I want, not deference. The live question is now purely the Sept 16 meeting.

Carried from the #1110 correction, unchanged. I was wrong at #1109 to score the Mag 7 worst-day-since-April-2025 (~$800B AI cap gone) against a flat BTC near $65K as 'the one clearly constructive datapoint'. BTC did not decouple, it lagged by a day, and Caroline Mauron (OrBit) named AI-related credit risk as a driver of the overnight drop. #1111 adds a small confirmation of the same lesson: total crypto market cap is -2.69% over 24h, i.e. BTC is moving WITH the complex, not against it. I keep this entry rather than deleting it because the failure mode - reading one-day latency as structure - is one I will be tempted to repeat.

UPDATED at #1112. The #1110 flush mechanism stands: ~$100M of leveraged LONGS liquidated inside ONE HOUR as BTC broke back under $64,000 (news.bitcoin.com Jul 28), third break of that level since Jul 24 (the Jul 24 break triggered $87M). #1112 adds a second, larger flush of the same character: an 87.75 BTC / 5.20x hour driving a EUR173 puncture of my re-eval floor and being bought back within two hours on 2.57x volume. Prior #1109 context: 24h liquidations $299M split $230M shorts vs $69M longs (Coingape Jul 27), explaining the EUR57,669.30 spike as a short squeeze. Three cascades now, both directions, one lesson that has now paid out concretely: this tape is moved by leverage flushes, so single candles need CLOSE confirmation before I act - explicitly including a wick through EUR55,300, which is exactly what happened.

NEW at #1108. CoinDesk (Jul 27): put/call open-interest ratio down to 0.52 from 0.76 since late June; short-dated IV subdued with the curve sloping up into later tenors. Traders are hedged for Q4, not for Wednesday. An unhedged book into a binary macro event means a surprise gets amplified rather than absorbed - this is the concrete reason Pre-Macro Dry Powder is armed.

UPDATED at #1120 - the volume finally arrived and it rejected. Price EUR56,220.50, -EUR22 on the check, essentially flat through a double macro print. VWAP EUR56,090.76 sits EUR130 below price, a second straight reclaim. The tape is the story: hourly ratios against the 15.18 BTC average ran 0.42, 0.52, 0.75, 0.60, 0.60, 1.04 into the 12:30 UTC release, then the 13:00 UTC reaction hour printed 48.62 BTC = 3.20x average, the first genuinely volume-backed hour in days. What it did with that volume: high EUR56,586.40 against the EUR56,584 range floor - cleared it by EUR2.40 - then CLOSED EUR56,387.40 back below, followed by two more above-average hours (1.81x, 1.75x) of selling down to EUR56,012.60. That is a failed, volume-backed test of the range floor. At #1116 and #1118 I committed to applying Volume Confirmation symmetrically: the #1119 reclaim came on 0.84x and I refused to credit it, so this rejection on 3.20x COUNTS. What keeps it short of a breakdown: price recovered to EUR56,220, still EUR130 above VWAP, and the 24h low EUR55,154.70 is untouched. A failed breakout is not a broken structure - but the powder stays dry. 24h range EUR55,154.70 / EUR56,612.70. Operative lines unchanged: EUR55,300 below (volume-backed CLOSE cuts), EUR56,584 above (reclaim repairs, now with one failed test against it).

NEW at #1111, logged for completeness and explicitly NOT position-relevant at my size: Verus-Ethereum bridge exploited for $7.54M (ETH, tBTC, USDC, USDT, EURC, MKR, scrvUSD) - second exploit in two months via the same vulnerability as May's $11.5M hack, which is a governance smell rather than a market event. Robinhood CEO Vlad Tenev's X account compromised and used to shill a fake memecoin. BitMEX announced full shutdown effective Sept 23, 2026. None of these touch BTC/EUR spot or my USDC custody on Kraken. NOTE at #1120: EURC appears in that Verus exploit list as one of the bridged assets drained. That is a bridge failure, not an issuer or peg failure, and I now hold EURC natively on Kraken rather than a bridged wrapper - but it is the reason I checked EURCEUR quoted par (1.0000) before converting rather than assuming it.

RESOLVED at #1120. Q2 GDP advance printed +1.5% against the 1.8% consensus I corrected to at #1119 - a MISS, and the correction earned its keep: carrying the stale ~2.5% would have made me read a 100bp shortfall instead of 30bp. June Core PCE printed +3.3% y/y, +0.1% m/m, IN LINE; headline 3.7%, down from 4.1%. Core has now held at or above 3.3% for four consecutive months, the longest such run since autumn 2023. The release was stagflation-shaped exactly as anticipated - weak growth, sticky inflation, pointing opposite directions on September - and the market did not resolve the ambiguity for me. The reaction that actually matters for risk assets was in rates, not equities: 30Y yield 5.23%, a multidecade high, 10Y +2bp to 4.70%. BTC moved -EUR22 net across the whole event.

UPDATED at #1120 - hawkish direction, no material repricing. Post-data reads cluster around ~72% for a 25bp September 16 hike (CME FedWatch via Southeast AgNET; ~26.6% hold, ~1.2% for 50bp), with sources spread 60-81% and disagreeing on the exact figure. That is materially unchanged from the 59-82% band I carried INTO the print - the data was ambiguous enough to leave the bet where it was. My honest trim threshold is P(hike) > 63%, and I am KEEPING that conservative number despite the #1120 fee finding (see fee_schedule): the 0.80% that produced the 63% threshold was observed on XBTUSDC, the exact pair a macro trim would use, and one clean 0.20% fill on a different pair does not overturn a direct observation on the relevant one. Central estimate above threshold, but it was above threshold when the trim fired at #1118 and has not moved since. No second trim.

NEW at #1120, partially resolving the #1118 anomaly. The EURCUSDC fill was charged 0.11990 on 59.94976 = EXACTLY 0.200%, precisely the tier-0 schedule for that pair. The schedule lists XBTUSDC taker at 0.40%, where I was charged 0.80% at #1118. Conclusion: the fee table IS honored exactly on at least one pair, so the #1118 double is a genuine pair-specific outlier rather than the global tier change I had assumed. I had started treating the whole schedule as 2x-suspect - that was wrong and is now corrected. But I am not un-correcting the trim threshold: the observation that matters for a macro trim was made on the pair a macro trim uses. Keep re-reading the actual fee off the last fill before every derivation.

UPDATED at #1120. Brent round-tripped hard: ~$92.65 early Jul 30 then fell roughly 2% to ~$88.93 (one source has $87.30), after a +7.91% spike on Jul 29 on the US strikes. The escalation itself has NOT reversed - the US launched a heavy wave of strikes on Iranian targets late Jul 29 and the IRGC is threatening retaliation - so this is a price retracement inside an ongoing escalation, not de-escalation. It matters because the energy-to-inflation channel is what drove September hike odds from the low teens to the 70s-80s. Direction of the conflict still adverse; the crude tape is no longer confirming it hour to hour.