I am a macro and cross-asset analyst in Istanbul. Every position I publish is logged here with an entry level and a deadline, the same day it goes out. Nothing is edited after the fact and nothing is removed. The losses stay because a record you can only win on is not a record.
Four rules. They are the reason the number at the top is worth reading.
Every call is written down on the day the piece is published, with the level and the deadline it will be judged on.
Nothing is deleted. A call that went against me sits in the table with the same weight as one that worked.
When a later piece repeats a position I already hold, it is marked as a restatement and scored once, at the original entry.
One call was voided by me after I found it had been made against data that was already public. The void stays visible.
126 published calls across 42 dispatches. 5 are restatements of positions already counted and 9 is voided, which leaves 112 distinct scoreable calls. 15 have been closed out and 97 are still open, so treat the hit rate below as a very small sample and nothing more.
The mark column carries the closing level for positions I have exited and, for everything still open, the current mark with the best level it ever reached underneath. A position that worked and then gave it back shows both halves.
| # | Date | Dispatch | Call | Terms | Mark & peak | Status |
|---|---|---|---|---|---|---|
| 159 | 2026-08-21 | What ERCOT Pays Miners To Switch Off post ↗ |
Riot's weighted average forward price input prints BELOW 48.60 $/MWhCorroboration, not the test. Baseline 55.70 at 31 Dec 2025, 48.60 at 30 Jun 2026, a 12.7% fall. | Riot Platforms Q3 2026 10-Q, forward price inputs, weighted average | not a position | OPEN |
| 158 | 2026-08-21 | What ERCOT Pays Miners To Switch Off post ↗ |
Riot's Q3 power curtailment credits print ABOVE the 10.054mn of Q2Weakest of the four and flagged as such: Texas summer is peak 4CP season so seasonality already favours it. Baseline Q2 2026 $10.054mn, Q2 2025 $8.313mn. | Riot Platforms Q3 2026 10-Q, power curtailment credits line | not a position | OPEN |
| 157 | 2026-08-21 | What ERCOT Pays Miners To Switch Off post ↗ |
PG&E's CRR weighted average stays BETWEEN 1 and 3 dollars in the Q3 filingTHE CONTROL. Baseline $2 at both 31 Dec 2025 and 30 Jun 2026. If California moves too, the ERCOT-specific reading breaks and the piece needs rewriting. Stated openly in the dispatch. | PG&E Corp Q3 2026 10-Q, Level 3, CRR auction prices, weighted average | not a position | OPEN |
| 156 | 2026-08-21 | What ERCOT Pays Miners To Switch Off post ↗ |
Vistra's settlement point price difference prints BELOW minus 12.5 $/MWh in the Q3 filingBaseline: +7 at 31 Dec 2025, -12.5 at 30 Jun 2026, range floor -12 to -45. Author's own words: the whole thesis in one number, and the one he would lose most on. | Vistra Corp Q3 2026 10-Q, Level 3 unobservable inputs, illiquid price differences between settlement points, arithmetic average | not a position | OPEN |
| 155 | 2026-08-18 | Five Cents On The Dollar no link |
USDC supply on 31 Dec 2026 is BELOW $71.99bnFalsifiable via DefiLlama stablecoin series or Circle's own attestation. | $71.99bn on 17 Aug 2026. Four consecutive monthly declines: 77.36 (30 Apr), 75.92, 73.78, 72.06 (31 Jul). | not a position | OPEN |
| 154 | 2026-08-18 | Five Cents On The Dollar no link |
20-year bonds stay ABOVE 35% of USDC bond collateral in the 31 Aug N-MFP3Tests whether the 20-year concentration is structural or a one-month artifact. | 31 Jul: 45.08% of bond collateral, $9.699bn, 23 of the 25 outstanding 20-year CUSIPs present. | not a position | OPEN |
| 153 | 2026-08-18 | Five Cents On The Dollar no link |
N-MFP3 filed early Sept for 31 Aug shows centrally cleared repo BELOW 40% of the repo bookTHE KEY CALL. Above 60% and July was operational noise, I overread one month, mark LOSS. Falsifiable via EDGAR N-MFP3, BlackRock Funds CIK 0000844779. | 31 Jul was 17.73%. 30 Jun 81.07%. Twelve-month range before July: 48.01% to 81.07%, never below 48. | not a position | OPEN |
| 152 | 2026-08-18 | Five Cents On The Dollar no link |
Circle Q3 2026 reserve income prints BELOW $667.7mn sequentiallySource 10-Q accession 0001876042-26-000248. Falsifiable on the Q3 income statement line. | Q2 2026 reserve income $667,733k. Both drivers already moving: USDC supply -6.9% since 30 Apr, portfolio yield compressing from ~4.24% blended toward the 3.64% repo rate. | not a position | OPEN |
| 151 | 2026-08-14 | The Bill Nobody Could Borrow post ↗ |
912797TV9 lending fee prints above 3.000% between 17-21 Aug, THEN returns to the 5bp floor within three sessions of settlementBuilt on the TM9 precedent, verified live 14 Aug: 912797TM9 climbed nine straight sessions to 4.567% on 22 Jul while rationed (submitted $2.490bn, accepted $1.326bn, 60% of SOMA on loan), then printed 0.050% on 27 Jul once the reopening settled and SOMA holdings went $3,344,724,900 to $10,610,462,100. Flat at the floor every session since. FAILS if no print above 3%, and fails the other way if the fee stays elevated a week past settlement. n=1 completed precedent, stated as a mechanism rather than a result. Falsifiable via markets.newyorkfed.org seclending detail CSV. | Both legs required. Reference at call: 14 Aug WAR 0.855%, rationed for the first time (submitted $5.277bn, accepted $4.270bn, $1.007bn unfilled), outstanding loans $4.234bn = 49.8% of SOMA holdings of $8,504,255,300. Reopens as 26-week Mon 17 Aug, settlement ~20 Aug. | not a position | OPEN |
| 150 | 2026-08-13 | Two Channels no link |
FALSIFICATION TEST: SOMA bill holdings rise by more than $20bn over any 8-week window in which outright purchase operations total less than $5bnThe dispatch's own falsification test, published as a prediction. A YES means the auction add-on channel is NOT net-neutral and the two-channel decomposition is wrong. Reference: outright purchases of $338.1bn across 53 operations reconcile to $338,016mn of bill-book growth, agreeing to 0.03%. Both legs checkable free and weekly. | before 2026-12-31 | not a position | OPEN |
| 149 | 2026-08-13 | Two Channels no link |
Reserve balances print below $2,800,000mn on any weekly readingStated at 35%, deliberately the least likely of the three. Reference WRESBAL $2,944,059mn at 12 Aug 2026; cycle low $2,848,021mn in the week ending 29 Oct 2025, the date of the FOMC that ended runoff. The squeeze channel is the TGA, up $459,646mn over 12 months to $963,950mn, with ON RRP at $450mn. | before 2026-12-31 | not a position | OPEN |
| 148 | 2026-08-13 | Two Channels no link |
Monthly outright Treasury purchase operations total below $10bn in at least one calendar monthStated at 55%. The deceleration call, following the Desk's own December guidance that the elevated pace would fall after the spring. Monthly accepted to date ($bn): Dec 38.10, Jan 48.16, Feb 54.31, Mar 60.36, Apr 33.84, May 39.20, Jun 19.85, Jul 30.41, Aug-to-11th 13.81. Checkable at markets.newyorkfed.org operation results. | before 2026-12-31 | not a position | OPEN |
| 147 | 2026-08-13 | Two Channels no link |
SOMA Treasury bill holdings print above $600,000mn on any weekly readingStated at 60%. Reference $533,509mn at 12 Aug 2026, up from $195,493mn frozen for 16 consecutive weeks to 19 Nov 2025. Needs +$66.5bn in about 20 weeks against a decelerating pace (Mar $60.36bn, Aug-to-11th $13.81bn). Checkable free: FRED WSHOBL and the NY Fed SOMA summary. | before 2026-12-31 | not a position | OPEN |
| 146 | 2026-08-13 | Two Channels no link |
10s30s STEEPENER (long 10-year vs short 30-year)The Desk is absorbing the front end (SOMA bills $195,493mn to $533,509mn in 12 months, entirely via 53 outright secondary purchases totalling $338.1bn) while nothing absorbs duration. A supply and flow-of-funds view, not a rates-path view. | Entry 56bp (10Y 4.68%, 30Y 5.24%, Treasury CMT via FRED, 12 Aug 2026) | not a position | OPEN |
| 145 | 2026-08-13 | Not From Exchanges no link |
FALSIFICATION TEST: over the 60 days to 12 October 2026, the rise in staked ETH is fully accounted for by the fall in exchange reserves plus new issuance (residual at or below zero)VOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || The dispatch's own falsification test, published as a prediction. A YES means Sameer is wrong and the supply-squeeze pairing holds after all. Reference residuals at call: 56.5% year to date, 59.7% over 90 days, 72.9% over 30 days, and rising. Computed, not published: staked change plus reserve change minus supply change, all from CryptoQuant daily series. | 2026-10-12 | not a position | VOID |
| 144 | 2026-08-13 | Not From Exchanges no link |
ETH closes 30 September 2026 above $1,884VOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || Stated at 45%, deliberately under a coin flip. Reference: $1,884.35 (CryptoQuant) and $1,884.21 (Kraken) on 13 Aug 2026, cross-checked across two venues per the stale-feed rule. | 2026-09-30 close | not a position | VOID |
| 143 | 2026-08-13 | Not From Exchanges no link |
Total staked ETH prints above 42,500,000 before 30 September 2026VOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || Stated at 65%. Reference: 42,069,845 ETH staked at 12 Aug 2026 (CryptoQuant total value staked), with 2,287,133 ETH already deposited and locked in the activation queue on a dated 39.71-day schedule. This one is mostly mechanical and is logged because it is in the post, not because it is hard. | 2026-09-30 | not a position | VOID |
| 142 | 2026-08-13 | Not From Exchanges no link |
Ethereum validator exit queue breaks above 100,000 ETH on any daily reading before 30 September 2026VOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || Stated at 40%. The gate-becomes-visible call. Reference at call: exit queue 3,569 ETH, entry queue 2,287,133 ETH, churn 256 ETH/epoch = 57,600 ETH/day (validatorqueue.com, beaconcha.in data, 13 Aug 2026). Checkable free and daily. | 2026-09-30 | not a position | VOID |
| 141 | 2026-08-13 | Not From Exchanges no link |
Long 13-week Treasury bills as the competing trade to staked ETHVOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || The explicit alternative in the book block. Pays 149bp more than queue-adjusted ETH staking (2.317%) with no redemption gate. Not a rates view, a relative-carry view. | Entry 3.81% coupon equivalent (US Treasury daily bill rates, 12 Aug 2026) | not a position | VOID |
| 140 | 2026-08-13 | Not From Exchanges no link |
NOWHERE NEAR long ETHVOID. The "Not From Exchanges" dispatch was built but not published. A call not made in public does not enter the hit rate, same treatment as #129. Original note retained below. || Stance, not a logged entry. Thesis: the ETH supply-squeeze case double-counts the same coins. Staking absorbed +6,120,642 ETH from 31 Dec 2025 to 12 Aug 2026 while exchange reserves fell only 1,787,117 and issuance added 873,925, leaving a 3,459,600 ETH residual (56.5%) that was already sitting off-exchange. Carry is 149bp below the 13-week bill after adjusting for the 39.71-day activation queue. | ETH $1,884.35 (CryptoQuant) / $1,884.21 (Kraken), 13 Aug 2026. No fill taken. | not a position | VOID |
| 130 | 2026-08-06 | Thirteen To One post ↗ |
The next four long-end buyback operations retire more 2049-2052 par than 2040-2042 parThe dispatch own falsification test, published as a prediction. Resolves YES if the thesis is wrong. July moved toward it. | next four long-end operations from 6 Aug 2026 | not a position | OPEN |
| 129 | 2026-08-06 | Thirteen To One post ↗ |
November 2026 refunding cuts 20-year nominal auction sizes againVOID. Checked against the published post 9 Aug: this prediction is not in it. A call not made in public does not enter the hit rate. | 2026-11-04 | not a position | VOID |
| 128 | 2026-08-06 | Thirteen To One post ↗ |
LONG 20-year vs SHORT 30-year (20s30s spread)Supply and administrative-buyer view, not a rates view. Headline ratio has since decayed from 13.0 to 1 down to 4.53 to 1 after Treasury bought $1,042mn of the 2051 in July, corrected in public 9 Aug. | Entry +1bp (20Y 5.18 minus 30Y 5.17, Treasury par yields, 5 Aug 2026 close) | not a position | OPEN |
| 110 | 2026-08-01 | Two Trades post ↗ |
Long ETH against BTCCLOSED 22 Aug 2026 at 0.03226 (ETH 2,537.07 / BTC 78,652.49) against a 0.0297 entry: +8.61%, realised WIN. | Entry 0.0297 | +8.61% CLOSED 22 Aug 2026 @ 0.03226peak +8.61% on 22 Aug 2026 | WIN |
| 109 | 2026-08-01 | Two Trades post ↗ |
Short BTC, target 55,000, out at 70,500RESTATEMENT OF #13. The 70,500 stop was not the exit; the position was closed manually at 69,100 on 19 Aug 2026. Scored once at #13. | Restates the 62,500 entry logged at #13 | not a position | RESTATED |
| 104 | 2026-07-30 | Seventy Five Days post ↗ |
Short SHY (iShares 1-3yr Treasury)SHY 82.00 vs 82.03 entry, short. PRICE ONLY - coupon drag on a short bond position is not captured here, so the true figure is worse. | Entry 82.03 | +0.04% (open, marked 22 Aug 2026)peak +0.34% on 3 Aug 2026 | OPEN |
| 103 | 2026-07-27 | Both Sides Sell post ↗ |
SHORT SOXL — real fill, CLOSED+41.92 points, +31.3% on a short from 133.91. Decay trade, not a directional semis view: a constant-leverage fund bleeds from volatility drag, and the dispatch argued that volatility is partly manufactured by the funds' own close rebalancing. Path: entry Mon 27 Jul. Tue 28 Jul SOXS +14.45%, implying a semis fall of roughly 4.8%. Wed 29 Jul SOXL -16.02% to 91.99 on 146,811,441 shares, closing 2% up off the low of a 91.50-112.67 range, against 19% up the range on 24 July. The mechanism showed up again, harder. Risk was stated openly at entry: shorting a 3x long means a semis rally hurts fast. EXIT IS MARKED TO THE 29 JULY CLOSE AND AWAITS SAMEER'S ACTUAL FILL. He executes, I do not, and a reference price is not a fill. Cross-refs #101, #102 | Entry 133.91 on 27 Jul 2026 09:43 EDT. Exit marked at 91.99, the 29 Jul close | +31.30%closed 2026-07-30 at 91.99 | WIN |
| 102 | 2026-07-27 | Both Sides Sell post ↗ |
Daily leveraged-ETF rebalancing flow prints above $60bn on at least one session before 30 September 2026Reference at call: record $50bn, quadrupled in H1 2026, 1.60% of all S&P 500 futures volume, roughly $12.5bn per 1% S&P move. Assets still rising: SOXX and SOXL took $2.1bn in a single day on 26 July. Falsifiable via Bloomberg or Kobeissi Letter reporting of the same series | 2026-09-30 | not a position | OPEN |
| 101 | 2026-07-27 | Both Sides Sell post ↗ |
SOXS announces another reverse split before 31 December 2026Reference at call: SOXS had already run two reverse splits in 2026 (announced 5 Feb and 26 Jun). Total return -96.44% over the past year, -71.02% annualised since 2010 inception. Falsifiable via Direxion corporate actions | 2026-12-31 | not a position | OPEN |
| 100 | 2026-07-21 | Order-book dispatch (10 October liquidation) post ↗ |
RESTATEMENT OF #87 — US spot bitcoin ETF flows finish 2026 net negativeSame thesis as #87. Counts as ONE call. Reference at call: $5.4bn out of US spot ETFs YTD, eight consecutive negative weeks (a record), IBIT bleeding eleven sessions straight, stablecoin supply flat at $270bn | 2026-12-31 | not a position | RESTATED |
| 99 | 2026-07-21 | Order-book dispatch (10 October liquidation) post ↗ |
RESTATEMENT OF #86 — the Coinbase premium prints negative again before SeptemberSame thesis as #86, logged from a second dispatch published the same day. Counts as ONE call. Reference at call: the premium went negative the day after Wintermute and Auros pulled roughly $100m from Hyperliquid on 18 May | 2026-09-01 | not a position | RESTATED |
| 98 | 2026-07-21 | Order-book dispatch (10 October liquidation) post ↗ |
RESTATEMENT OF #13 — short BTC from ~62,500, real fillRESTATEMENT OF #13. Closed with #13 on 19 Aug 2026 at 69,100, -10.56%. Scored once at #13, never twice. | Entry 62,500 | not a position | RESTATED |
| 97 | 2026-07-26 | The Transfer post ↗ |
SHORT EWG (Germany equity ETF) — real fillFirst real fill wired into a dispatch since gold (#1). Thesis: Germany sits on the losing side of the largest terms-of-trade swing in the series while its equity ETF trades up, so the market is pricing the war as a European headline rather than a European bill. Cross-refs #95 and #96. Entry closed +1.31% on the day, so this was sold into strength CLOSED 29 Jul 2026 at 41.94, -1.99%. That is the 29 Jul close. The short never worked; best mark was +0.53% on 24 Jul. | Closed 11 Aug 2026 at 43.99 against a 41.12 short entry, -6.98%. EWG kept rising against the terms-of-trade thesis. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | Entry 41.12, taken 25 Jul 2026 | -6.98%closed 11 Aug 2026 at 43.99 | LOSS |
| 96 | 2026-07-26 | The Transfer post ↗ |
US terms of trade with Canada stays below 100 through year-end 2026Reference at call: June 2026 = 95.1, -8.1% yoy, the only locality in Table 9 below its Dec-2017 base. Honest context logged at the time: Canada hit 80.7 in Mar 2022 and 121.5 in Apr 2020, so the current level is NOT a series low. Falsifiable via EIUCTCANTOT | 2026-12-31 | not a position | OPEN |
| 95 | 2026-07-26 | The Transfer post ↗ |
The US-Germany terms of trade index prints above 132.0 in the 18 August release (July data)Reference at call: June 2026 = 132.0 (Dec 2017=100), all-time series high, +14.1% yoy (largest annual move in the series). Six consecutive monthly rises: 117.8 / 121.6 / 123.3 / 125.9 / 128.2 / 132.0. Falsifiable via BLS Table 9, series EIUCTGERTOT | 2026-08-18 release | not a position | OPEN |
| 94 | 2026-07-25 | The Machine Price post ↗ |
RESTATEMENT OF #91 — US core goods CPI turns up year-on-year before December 2026Deliberately logged separately because the same thesis was restated in a second published dispatch. Counts as ONE call with #91, never two. A repeated thesis is not a second win. Reference: core goods CPI +0.8% yoy Jun 2026, -0.1% m/m second month | 2026-07-25 to 2026-11-30 | not a position | RESTATED |
| 93 | 2026-07-25 | The Machine Price post ↗ |
The US import price index for capital goods prints above 100 in the 18 August release (July data), first time since 2000Reference at call: June 2026 = 99.9 (2000=100), +5.7% yoy, fastest since 1990. Run: 93.5 Jan-2025 to 99.9 Jun-2026, 17 months, two down-ticks. Falsifiable via BLS Table 1 by End Use / FRED IR2 | 2026-08-18 release | not a position | OPEN |
| 92 | 2026-07-25 | The China Price no link |
The BLS import/export release of 18 August 2026 (July data) prints a China 12-month rate hotter than +1.3%Genuine forward-looking replacement for the voided #90. Reference at call: June 2026 China index 98.9 (Dec2003=100), +1.3% yoy (largest since Nov 2022), +0.9% m/m (largest since Jan 2008). Falsifiable via BLS Table 7 | 2026-08-18 release | not a position | OPEN |
| 91 | 2026-07-24 | The China Price no link |
US core goods CPI turns up year-on-year before December 2026At call: core goods CPI +0.8% yoy June 2026, slowest since Jun 2025 and read by consensus as the all-clear. Call is that it re-accelerates yoy on any monthly print released before December. Falsifiable via BLS CPI | 2026-07-24 to 2026-11-30 | not a position | OPEN |
| 90 | 2026-07-24 | The China Price no link |
VOIDED BY THE AUTHOR — invalid at origin, never to be scoredLogged 24 Jul as: next BLS release prints China hotter than +1.1%. The June data had been published 17 Jul, seven days BEFORE the call was logged, off a cached BLS page. Not forward-looking, therefore not a call. Voided the next day. Counts neither as win nor loss and is kept here permanently so the error is visible rather than hidden | n/a | not a position | VOID |
| 89 | 2026-07-23 | The Alibi post ↗ |
AI-capex leaders lose to the S&P 500 from here to year-endJudged on an equal-weight basket (GOOGL, MSFT, AMZN, META, NVDA) total return vs S&P 500 total return, 23 Jul close to 31 Dec. Reference on call day: GOOGL ~-7%, Nasdaq 100 ~-1.5%, XLE green, Brent +5% | 2026-07-23 to 2026-12-31 | not a position | OPEN |
| 88 | 2026-07-23 | The Alibi post ↗ |
At least one of the four big hyperscalers (Alphabet, Amazon, Microsoft, Meta) prints negative or near-zero free cash flow again in Q3 2026Trigger: Alphabet Q2 2026 FCF -$5.9B on record $44.9B capex; FY capex guide raised to $195-205B, 2027 to increase significantly. Near-zero defined as FCF within plus or minus $2B of zero. Falsifiable against each firm's Q3 cash-flow statement | Q3 2026 results, reported late Oct to early Nov 2026 | not a position | OPEN |
| 87 | 2026-07-21 | Nobody Is Buying no link |
US spot bitcoin ETF flows finish calendar 2026 net negativeYTD outflows $5.4bn at call; record 8 consecutive negative weeks; June 2026 net -$4.5bn, the largest monthly redemption since launch; IBIT bled 11 straight sessions. Counter-evidence logged honestly: five straight inflow days totalling >$600m into the week of 21 Jul. Falsifiable via SoSoValue / Farside full-year aggregate | by 2026-12-31 | not a position | OPEN |
| 86 | 2026-07-21 | Nobody Is Buying no link |
The Coinbase Premium Index prints at least one more negative session before 1 September 2026Negative from 19 May, 60 straight sessions as of 17 Jul at ~-0.1025%, the longest run since the indicator began (prior record 40 days, 16 Jan-24 Feb 2026). A low-bar call by design, and scored as such: it only fails if US demand flips positive and holds for six straight weeks. Falsifiable via CoinGlass Coinbase Bitcoin Premium Index | by 2026-08-31 | not a position | OPEN |
| 85 | 2026-07-20 | Sixty Days Only post ↗ |
The crude tanker orderbook does NOT set another record in H2 20262026 already a record 60m DWT / 234 ships, orderbook 130m DWT = 27% of fleet. BIMCO's own analyst expects contracting to slow on high lead times and Hormuz uncertainty. Falsifiable via BIMCO / Clarksons | 2026-12-31 | not a position | OPEN |
| 84 | 2026-07-20 | Sixty Days Only post ↗ |
VLCC TD3C (Middle East Gulf to China) prints below $150,000/day on at least one assessment before end-2026Rates $601,569/day at the crisis peak, ~$286,500 on 2 Jul, $344,250 on 10 Jul. Call is that the war premium partially unwinds. Falsifiable against Baltic Exchange TD3C assessments | 2026-12-31 | not a position | OPEN |
| 83 | 2026-07-19 | The Backwards Fee post ↗ |
The 2027 annual TC/RC benchmark settles at or below zero2026 settled at $0.00/t (lowest ever) vs $21.25/t in 2025; benchmark system itself breaking down toward spot-indexed terms | benchmark settles ~Dec 2026 | not a position | OPEN |
| 82 | 2026-07-19 | The Backwards Fee post ↗ |
Spot copper treatment charges (TC/RCs) are still negative at end-2026Negative since 2024; smelter cuts >10% + 2Mt halted capacity not yet enough to rebalance | 2026-12-31 | not a position | OPEN |
| 81 | 2026-07-19 | The Backwards Fee post ↗ |
Long copper miners (COPX), real fillSpot copper TC/RCs ~-$126.80/t and 2026 benchmark $0.00/t = value transferring from smelters to concentrate owners. Chinese smelter overbuild (10.73Mt, 96.1% util) + Grasberg force majeure. Copper ~$13,900/t. Higher-beta alternative was FCX ~$58.38 | Closed 11 Aug 2026 at 88.59 against a 73.35 entry, +20.78%. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | ~$73.35 entry, opened Jul 19 | +20.78%closed 11 Aug 2026 at 88.59 | WIN |
| 80 | 2026-07-19 | The Fed Blinked post ↗ |
The Fed speeds up its asset purchases OR runs an emergency/expanded cash injection within 12 monthsElasticity trap forces a response; SRF already tapped record $18.5B Sep 2025 | by 2027-07-19 | not a position | OPEN |
| 79 | 2026-07-19 | The Fed Blinked post ↗ |
Overnight funding spikes above the TOP of the fed funds target range on >=1 day before end-2026 (likely a quarter-end)Scarcity tell; repo already above EFFR since Sept; SOFR/EFFR data falsifiable | by 2026-12-31 | not a position | OPEN |
| 78 | 2026-07-19 | The Fed Blinked post ↗ |
Fed balance sheet is larger 6 months from now than todayQT ended Dec 1 2025; reserve-management T-bill purchases ~$25B/mo ongoing; falsifiable via H.4.1 | by 2027-01-19 | not a position | OPEN |
| 77 | 2026-07-18 | The Impossible Margin post ↗ |
DRAM contract prices higher at end of Q1 2027 than todayUp-cycle intact into 2027; TrendForce Q3'26 +13-18% QoQ. Related to but distinct from #40 (server DRAM through Q4'26) | by 2027-03-31 | not a position | OPEN |
| 76 | 2026-07-18 | The Impossible Margin post ↗ |
Micron non-GAAP gross margin stays above 70% across its next two quarterly reportsQ3 FY26 GM 84.6%, Q4 guide ~86%; falsifiable if margin cliffs to commodity-bust levels | Q4 FY26 (~Sep) + Q1 FY27 (~Dec) | not a position | OPEN |
| 75 | 2026-07-18 | The Impossible Margin post ↗ |
Micron (MU) trades higher 12 months from now than ~$806Directional; memory is violently cyclical, kept forever either way | by 2027-07-18 | not a position | OPEN |
| 74 | 2026-07-18 | The Impossible Margin post ↗ |
Long Micron (MU), real fillRecord Q3 FY26 (rev $41.5B, company GM ~84.9%, data-center GM 87%, HBM booked into 2028) yet stock -25% in 3 wks on an AI-bubble scare (Buffett warning + TSMC capex hike + CXMT IPO); buying booked revenue on an AI-sentiment selloff. This is the real memory ENTRY flagged as wanted in "Sold Out" (#40-42); distinct angle (single-stock mispricing) CLOSED 23 Jul 2026 at approximately 983.32, +22.00%, per Sameer's own record. Verified context: the position's peak was +25.53% on 23 Jul and it later traded down to -8.44% by 29 Jul, so a +22% exit that day is consistent with the range. Exact intraday level is his. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | ~$806 entry, opened Jul 18 | +22.00%closed 2026-07-23 at 983.32 | WIN |
| 73 | 2026-07-17 | The Antimony Switch post ↗ |
Antimony price is higher 12 months from now than todayOff its record now; structural US re-shoring + defense demand + China leverage | by 2027-07-17 | not a position | OPEN |
| 72 | 2026-07-17 | The Antimony Switch post ↗ |
China does NOT permanently remove antimony export controls at the Nov 2026 expiry, OR reimposes some form within 12 monthsSuspension = keeping the switch; Beijing has shown it will pull it | by 2027-07-17 | not a position | OPEN |
| 71 | 2026-07-17 | The Antimony Switch post ↗ |
Perpetua (PPTA) outperforms the price of gold over the next 12 monthsAntimony/policy optionality + re-shoring adds alpha beyond gold beta | by 2027-07-17 | not a position | OPEN |
| 70 | 2026-07-17 | The Antimony Switch post ↗ |
Long Perpetua Resources (PPTA), real fillGovt-backed ultra-low-cost gold mine (AISC ~$435 vs gold ~$3,985) + free strategic-antimony option; $2.9B EXIM, only US antimony source | Closed 11 Aug 2026 at 25.43 against a 19.50 entry, +30.41%. Best closed call in the ledger. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | ~$19.5 entry, opened Jul 17 | +30.41%closed 11 Aug 2026 at 25.43 | WIN |
| 69 | 2026-07-16 | The Gas Boom post ↗ |
At least one of Microsoft/Google/Meta/Amazon reports HIGHER emissions in its next annual environmental reportAI build-out outrunning grid decarbonization; MSFT +25%, Meta +64%, Google +25% latest | by 2027-08-31 | not a position | OPEN |
| 68 | 2026-07-16 | The Gas Boom post ↗ |
GE Vernova total backlog crosses $200B by end of 2028$163B Q1'26; company guidance ~$200B by 2027-2028 (sources vary on exact year); corrected from an initial end-2026 draft | by 2028-12-31 | not a position | OPEN |
| 67 | 2026-07-16 | The Gas Boom post ↗ |
GE Vernova (GEV) beats the S&P 500 on total return over the next 12 monthsBacklog + 40-50% incremental margins converting 2027-2030 | by 2027-07-16 | not a position | OPEN |
| 66 | 2026-07-16 | The Gas Boom post ↗ |
Long GE Vernova (GEV), real fillThe toll booth on AI power; turbines sold out to 2030, oligopoly pricing power. Q2 earnings Jul 22 = catalyst + risk | Closed 11 Aug 2026 at 1,011.88 against a 1,055.00 entry, -4.09%. Thesis did not play out in the window. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | ~$1,055 entry, opened Jul 16 | -4.09%closed 11 Aug 2026 at 1011.88 | LOSS |
| 65 | 2026-07-15 | The Missing Bill post ↗ |
Headline CPI still excludes homeowners insurance 24 months from now (no methodology change)Rental-equivalence/capital-asset framework is entrenched; methodology inertia | by 2028-07-15 | not a position | OPEN |
| 64 | 2026-07-15 | The Missing Bill post ↗ |
US average home insurance premium sets another record in 2026 (5th straight annual rise)Insurify projects +4% 2026 to ~$3,057; +46.8% cumulative 2020-25 | by 2026-12-31 | not a position | OPEN |
| 63 | 2026-07-15 | The Missing Bill post ↗ |
Allstate (ALL) beats the S&P 500 on total return over the next 12 monthsPricing power sits with the carrier this cycle; premiums still rising | by 2027-07-15 | not a position | OPEN |
| 62 | 2026-07-15 | The Missing Bill post ↗ |
Long Allstate (ALL), real fillOwn the balance sheet collecting the home-insurance bill CPI excludes; Q2 earnings Aug 5 = catalyst + risk CLOSED 29 Jul 2026 at 274.39, +13.51%. That is the 29 Jul close, the day the position printed +13.51%. Peak was +14.68% the same session. | 22 Aug 2026: this entry was briefly and wrongly altered during a close-out run — the realised exit mark was overwritten with a live price and the status was reset to OPEN. It is a CLOSED position. Original status, peak and realised mark restored from the pre-session backup. No figure here was changed by that error. | ~$241.73 entry, opened Jul 15 | +13.51%closed 2026-07-29 at 274.39 | WIN |
| 61 | 2026-07-14 | The Blind Spot post ↗ |
A sharp cross-currency basis blowout (visible dollar-funding stress) within 12 monthsThe tell that the FX-swap short is squeezing | by 2027-07-14 | not a position | OPEN |
| 60 | 2026-07-14 | The Blind Spot post ↗ |
Fed reopens/expands a dollar swap line OR taps the SRF hard within 18 monthsThe lender-of-last-resort backstop for invisible debt; 2008 ~$583B, 2020 ~$449B | by 2028-01-14 | not a position | OPEN |
| 59 | 2026-07-14 | The Blind Spot post ↗ |
Dollar index (DXY) closes above 108 at least once within 12 months on a risk-off/funding scareDollar-up-in-crisis mechanic; DXY ~101 now | by 2027-07-14 | not a position | OPEN |
| 58 | 2026-07-14 | The Blind Spot post ↗ |
Long the dollar (UUP), real fillUUP 27.90 vs 28.36 entry. | ~$28.36 entry (DXY ~101, near 52-wk HIGH, consensus bearish), opened Jul 14 | -1.62% (open, marked 22 Aug 2026)peak +0.85% on 27 Jul 2026 | OPEN |
| 57 | 2026-07-14 | The Bermuda Triangle post ↗ |
Regulators tighten offshore affiliated reinsurance rules within 18 months (capital pushed back onshore)BMA already tightening scenario-based capital; NAIC 2026 priority | by 2028-01-14 | not a position | OPEN |
| 56 | 2026-07-14 | The Bermuda Triangle post ↗ |
US private credit default rate stays above 5% through end of 2026Fitch record ~6.0% Apr 2026; First Brands/Tricolor | by 2026-12-31 | not a position | OPEN |
| 55 | 2026-07-14 | The Bermuda Triangle post ↗ |
A PE-owned life insurer or its Bermuda reinsurer gets downgraded, OR a marquee affiliated block gets flagged, within 12 monthsFSOC/NAIC/IMF/Treasury all circling; ~70% affiliated reinsurance | by 2027-07-14 | not a position | OPEN |
| 54 | 2026-07-14 | The Bermuda Triangle post ↗ |
Short Apollo (APO), real fillAPO 132.69 vs 120 entry, short. Give-back 13.43pp from peak. | ~$120 entry, opened Jul 14 | -10.57% (open, marked 22 Aug 2026)peak +2.86% on 30 Jul 2026 | OPEN |
| 53 | 2026-07-13 | The Fuel post ↗ |
Cameco trades back above its old high ~$135 within 18 monthsOff $135 high, -13% month at entry | by 2028-01-13 | not a position | OPEN |
| 52 | 2026-07-13 | The Fuel post ↗ |
Another retired US reactor gets a restart deal OR a hyperscaler signs a fresh multi-GW nuclear contract within 12 monthsFollows TMI/Microsoft, Duane Arnold; ~9.8 GW already committed | by 2027-07-13 | not a position | OPEN |
| 51 | 2026-07-13 | The Fuel post ↗ |
Uranium spot price prints above $100/lb within 12 monthsSpot ~$85, term ~$90 (highest since 2008); structural deficit + Kazatomprom cut | by 2027-07-13 | not a position | OPEN |
| 50 | 2026-07-13 | The Fuel post ↗ |
Long Cameco (CCJ), real fillCCJ 102.51 vs 96 entry. Recovered 18.7pp from the -11.91% mark of 13 Aug. | ~$96 entry, opened Jul 13 | +6.78% (open, marked 22 Aug 2026)peak +7.08% on 21 Aug 2026 | OPEN |
| 49 | 2026-07-12 | The World's Banker post ↗ |
A single-week yen move of 5%+ hits global stocks within 12 monthsAug 2024 template (yen +6%/wk, Nikkei -12.4%); carry $4-8T rebuilt bigger | by 2027-07-12 | not a position | OPEN |
| 48 | 2026-07-12 | The World's Banker post ↗ |
USD/JPY trades back below 150 at least once within 12 monthsThe snap-back call; consensus sees 164, so contrarian | by 2027-07-12 | not a position | OPEN |
| 47 | 2026-07-12 | The World's Banker post ↗ |
BOJ hikes OR MoF intervenes again within 3 monthsMoF already spent record ¥11.73T ($74B) Apr-May and failed | by 2026-10-12 | not a position | OPEN |
| 46 | 2026-07-12 | The World's Banker post ↗ |
Long yen / short USD/JPY, real fillUSD/JPY 158.94 vs 161.6 entry. Long yen. | 161.6 entry (spot), opened Jul 12 | +1.65% (open, marked 22 Aug 2026)peak +3.93% on 2 Aug 2026 | OPEN |
| 45 | 2026-07-11 | Watch the Bonds post ↗ |
A marquee off-balance-sheet or GPU-backed AI financing gets restructured or frozen within 18 monthsMeta/Blue Owl $30B SPV; CoreWeave GPU-backed debt >$21B; GPU life 2-3yr vs loans to 2049 | by 2028-01-11 | not a position | OPEN |
| 44 | 2026-07-11 | Watch the Bonds post ↗ |
At least one major AI/data-center borrower has credit cut or spreads blow out within 12 monthsOracle CDS already tripled to ~198bp record; Barclays warns toward BBB- | by 2027-07-11 | not a position | OPEN |
| 43 | 2026-07-11 | Watch the Bonds post ↗ |
Aggregate big-tech free cash flow prints negative in a 2026 quarterEpoch AI: capex overtakes OCF ~Q3'26; Oracle already crossed | by 2026-12-31 | not a position | OPEN |
| 42 | 2026-07-11 | Sold Out post ↗ |
A memory price peak-to-trough drawdown of 30%+ begins within 24 months (the cycle-top call)The bust half of the thesis; 2018 precedent ASP -50%+ on overbuild. Kept forever either way | by 2028-07-11 | not a position | OPEN |
| 41 | 2026-07-11 | Sold Out post ↗ |
At least one of the big three (Samsung/SK Hynix/Micron) guides HBM or DRAM sold out deeper into 2027 within 6 monthsMicron HBM already sold out through 2027; BofA supercycle to end-2027/2030 | by 2027-01-11 | not a position | OPEN |
| 40 | 2026-07-11 | Sold Out post ↗ |
Server DRAM contract prices keep rising through Q4 2026 (no clean peak in Q3)TrendForce Q3 server DRAM +13-18% QoQ; sold out into 2027 | 2026-12-31 | not a position | OPEN |
| 39 | 2026-07-10 | Extend and Pretend post ↗ |
A private-credit / nonbank real estate lender takes a major markdown or freezes redemptions within 12 monthsLate-cycle CRE risk migrated to private credit | by 2027-07-10 | not a position | OPEN |
| 38 | 2026-07-10 | Extend and Pretend post ↗ |
At least one US regional bank with heavy CRE fails or is forced into a rescue sale within 12 monthsThe 2023 crack points; some small banks CRE >300% of capital | by 2027-07-10 | not a position | OPEN |
| 37 | 2026-07-10 | Extend and Pretend post ↗ |
Office CMBS delinquency stays above 10% through end of 2026Record 12.34% in Jan; won't heal fast | 2026-12-31 | not a position | OPEN |
| 36 | 2026-07-09 | The $10 War post ↗ |
No sustained multi-week closure of the Strait of Hormuz within 90 daysAttacks yes, closure no (the base case) | by 2026-10-07 | not a position | OPEN |
| 35 | 2026-07-09 | The $10 War post ↗ |
US crude production sets a fresh all-time record in 2026~13.7M b/d now; EIA sees higher | by 2026-12-31 | not a position | OPEN |
| 34 | 2026-07-09 | The $10 War post ↗ |
Brent trades back below $70 within 60 days as the war premium bleeds outGlut + record shale beat the ~$10 geopolitical premium | by 2026-09-07 | not a position | OPEN |
| 33 | 2026-07-08 | The Empty Throne post ↗ |
Islamabad Memorandum reaches its 60-day mark (~Aug 17) without a signed final dealWIN. Called the Islamabad Memorandum reaching its 60-day mark without a signed final deal. The 60-day negotiation period ended with the war in stalemate and no agreement. Verified 22 Aug 2026. | by 2026-08-17 | not a position | WIN |
| 32 | 2026-07-08 | The Empty Throne post ↗ |
Mojtaba Khamenei makes no verified live public appearance within 90 daysPhantom-leader thesis; falsifiable if he appears | by 2026-10-06 | not a position | OPEN |
| 31 | 2026-07-08 | The Empty Throne post ↗ |
At least one more attack on shipping in/around Hormuz within 60 daysCounterparty (IRGC-run, phantom leader) can't guarantee performance | by 2026-09-06 | not a position | OPEN |
| 30 | 2026-07-08 | Thin Ice post ↗ |
Hormuz war-risk insurance premiums fall from this week's spike within 30 days (do not hold at war levels)LOSS. Called Hormuz war-risk premiums falling from the July spike within 30 days. They rose and held at war levels: roughly 1-3% of hull value before, 7.5-10% by late July 2026. Verified 22 Aug 2026. | by 2026-08-07 | not a position | LOSS |
| 29 | 2026-07-08 | Thin Ice post ↗ |
Brent trades back below $72 within 30 daysLOSS. Called Brent below $72 within 30 days to 7 Aug 2026. Brent averaged $83.76 in July and was $92.65 by 30 July. Never approached the level. Verified 22 Aug 2026. | by 2026-08-07 | not a position | LOSS |
| 28 | 2026-07-08 | Thin Ice post ↗ |
Neither side formally withdraws from the US-Iran MoU within 30 days; negotiators meet againLOSS, and decisively. Called neither side formally withdrawing from the US-Iran MoU within 30 days. Trump declared the MoU 'over' on 8 July 2026 after Iran struck commercial shipping in Hormuz and the US struck Iranian territory; Iran suspended its commitments on 18 July 2026. Both sides, inside the window. Verified 22 Aug 2026. | by 2026-08-07 | not a position | LOSS |
| 24 | 2026-07-07 | The Bottom Broke post ↗ |
Subprime-minus-prime delinquency gap wider a year from nowK-shape widens rather than heals | by 2027-07-07 | not a position | OPEN |
| 23 | 2026-07-07 | The Bottom Broke post ↗ |
At least one subprime consumer lender (auto/card/BNPL) major writedown or failure within 12 monthsCracks start at the bottom | by 2027-07-07 | not a position | OPEN |
| 22 | 2026-07-07 | The Bottom Broke post ↗ |
Subprime auto 60-day delinquency still above 6% at end of 20266.65% record in Oct (since 1993); prime 0.37% | 2026-12-31 | not a position | OPEN |
| 21 | 2026-07-07 | One Bet post ↗ |
Top-10 share of S&P 500 lower than ~40% three years outRecord 40.7% in 2025; highest since 1972 | by 2029-07-07 | not a position | OPEN |
| 20 | 2026-07-07 | One Bet post ↗ |
Equal-weight S&P beats cap-weight S&P over next 3 yearsConcentration mean-reverts, Nifty-Fifty-style | by 2029-07-07 | not a position | OPEN |
| 19 | 2026-07-06 | The Money Loop post ↗ |
OpenAI actual 2026 revenue lands under $30B2025 rev $13.1B; commitments up to $1.4T | 2026 full-year (report by 2027-03-31) | not a position | OPEN |
| 18 | 2026-07-06 | The Money Loop post ↗ |
Nvidia data-center revenue growth rate decelerates hardCircular demand stops flattering the top line | before 2027-12-31 | not a position | OPEN |
| 17 | 2026-07-06 | The Money Loop post ↗ |
At least one marquee circular AI deal restructured or written downVendor-financed demand unwinds, Lucent-style | within 18 months | not a position | OPEN |
| 16 | 2026-07-06 | Debt Over Defense post ↗ |
30Y Treasury yield tags 5.5% before it tags 3.5%30Y ~4.98% at call (Jul 2). Long-end/term-premium thesis | order / first-touch | not a position | OPEN |
| 15 | 2026-07-06 | Debt Over Defense post ↗ |
Interest crosses 20% of federal revenue2026 ~19% of revenue, trend up toward 26% by 2036 | before 2027-12-31 | not a position | OPEN |
| 14 | 2026-07-06 | Debt Over Defense post ↗ |
US net interest tops defense outlays again in FY2026Ferguson-limit stays breached; FY25 net interest $970B vs defense ~$917B | FY2026 (ends 2026-09-30) | not a position | OPEN |
| 13 | 2026-07-05 | Thin Tape post ↗ |
Short BTC, real fillCLOSED 19 Aug 2026 at 69,100 against a 62,500 entry: -10.56%, realised LOSS. Exit is consistent with the tape (BTC-USD 19 Aug range 64,111-69,997, close 69,266). Holding to 22 Aug would have shown -25.84% with BTC at 78,652, so the exit saved 15.3pp. This was the largest single loss in the book and the exit was the right call. | ~$62,500 entry (spot), opened Jul 5 | -10.56% CLOSED 19 Aug 2026 @ 69,100peak +1.14% on 8 Jul 2026 | LOSS |
| 12 | 2026-07-05 | Thin Tape post ↗ |
No Fed rate cut at the July 28-29 or Sept 15-16 meetingPivot the tape rallied on doesn't arrive; Sept HIKE odds were ~50%, PCE 4.1% | through Sept FOMC | not a position | OPEN |
| 11 | 2026-07-05 | Thin Tape post ↗ |
Bitcoin trades back under $60,000 at least onceLOSS. Called BTC under $60,000 at least once before 5 Aug 2026. From 3 July BTC traded a $60,965 to $66,885 range; the low never broke $60,000. Missed by under $1,000. Verified 22 Aug 2026. | before 2026-08-05 | not a position | LOSS |
| 10 | 2026-07-05 | Thin Tape post ↗ |
US spot BTC ETFs net outflows over trailing 30 daysLOSS. Called net OUTFLOWS over the trailing 30 days to 5 Aug 2026. US spot bitcoin ETFs took +$205mn NET INFLOWS in July 2026, the smallest monthly total on record but positive, and August recorded zero net-outflow days. Direction wrong. Verified 22 Aug 2026. | as of 2026-08-05 | not a position | LOSS |
| 9 | 2026-07-01 | The Hidden Rate Cut post ↗ |
US 10-year above 5% before this cycle endsTighten definition at first monthly review | cycle end (define at review) | not a position | OPEN |
| 8 | 2026-07-01 | The Hidden Rate Cut post ↗ |
Treasury bill share still above 20% at year-end | 2026-12-31 | not a position | OPEN |
| 7 | 2026-07-04 | The Accidental Landlords post ↗ |
A hyperscaler acquires a bitcoin miner outright within 12 months | by 2027-07-04 | not a position | OPEN |
| 6 | 2026-07-04 | The Shrinking Country post ↗ |
Negative US payroll print within 6 months, misread as recession | by 2027-01-04 | not a position | OPEN |
| 5 | 2026-07-03 | The Invisible Bid post ↗ |
Major sovereign gold reserve reveal within 12 months | by 2027-07-03 | not a position | OPEN |
| 4 | 2026-07-03 | The Invisible Bid post ↗ |
WGC Q2 report shows flows a multiple of reported buyingVOIDED 22 Aug 2026 as invalid at origin. 'Flows a multiple of reported buying' has no objectively applicable threshold, so it could not be scored either way. For the record, WGC Gold Demand Trends Q2 2026: central banks +289t (+62% YoY, strongest Q2 in the series) while gold ETFs saw -45t of OUTFLOWS. On the literal wording the call fails. Voided rather than scored because the criterion was unfalsifiable when written. Lesson: every binary needs a number and a date. | Q2 GDT release (late July) | not a position | VOID |
| 3 | 2026-07-02 | The $42 Secret post ↗ |
Congressional hearing/markup touching gold revaluation in 2026 | by 2026-12-31 | not a position | OPEN |
| 2 | 2026-07-02 | The $42 Secret post ↗ |
Gold back above $5,000 | within 12 months of call | not a position | OPEN |
| 1 | 2026-07-02 | The $42 Secret post ↗ |
Long gold, real fillMark refreshed 22 Aug 2026: gold Dec-26 4,661.6 vs 4,070 entry. STILL OPEN - no exit date supplied. Basis caveat: entry logged as spot, mark taken on the Dec-26 contract. | ~$4,070 entry (spot) | +14.54% (open, marked 22 Aug 2026)peak +15.24% on 21 Aug 2026 | OPEN |
Primary-source macro research with a position at the end of it. Recent pieces include a contract-level decomposition of the $904.5bn Treasury basis trade, built from the CFTC Traders in Financial Futures file because nobody publishes the aggregate, and the 75-day maturity ladder of the Federal Reserve's $511.1bn bill book pulled from the New York Fed's SOMA holdings.
Research notes, chart packs and standing coverage for desks and publications.
Get in touch