Track Record
Every forward-looking call this newsletter has made, graded in public against what actually happened. Misses included, with post-mortems. Published issues are never edited and grades are never retroactively changed.
Where the calls land, by type
Macro data is where we're sharpest. Kinetic geo and diplomatic calls are where we've been wrong most. We publish this so you don't have to guess which kinds of calls to trust.
| Call type | Resolved | Right | Hit rate |
|---|---|---|---|
| regime | 1 | 1 | 100% |
| FREIGHT | 12 | 10 | 83% |
| COMMODITY | 13 | 10 | 77% |
| MACRO DATA | 21 | 12 | 57% |
| KINETIC GEO | 21 | 9 | 43% |
| DIPLOMATIC | 8 | 2 | 25% |
| SUPPLY POLICY | 2 | 0 | 0% |
| FED / DOLLAR | 6 | 0 | 0% |
| CRYPTO | 1 | 0 | 0% |
Calibration
From W25 onward every call carries its own probability, set in 10% steps at publish time and shown on the call below. Once 50 probability-tagged calls have resolved (48 of 50 so far — earlier calls predate the probability field, which is why the total above is larger), this section will publish the Brier score and a calibration curve — how often our 70% calls actually land. Until then there is no number here, because a calibration score computed on a handful of calls would be noise dressed up as rigor. The machinery is built and gated; it switches itself on the week the fiftieth call resolves.
The record, week by week
Each column is one graded week: green right, amber partial, red wrong. Runs of red are left visible on purpose — the misses are the reason you can trust the greens.
The full log
Graded weeks first, newest on top; weeks still awaiting verdicts sit at the end. Grades follow each issue's own published scorecard, cross-checked against daily price history. NO-FIRE means the call's condition never triggered.
No verified physical halt of Hormuz transits (tracker-confirmed at-or-near-zero, confirmed mining, or verified full stop) on any day through Fri Aug 28 -> disrupted holds
No tracker reported a zero-transit day Aug 25-28. Counts ran 2 (Aug 24), a same-outlet 1-versus-5 conflict for Aug 25 (printed unresolved — neither figure is zero), 17 (Aug 26, Kpler via Reuters), 7 (Aug 27, Windward: 3 of 6 dark). The closest thing to a zero — the disputed Tuesday 'zero outbound in 24h' claim — is contradicted by the same publication family's own 5-vessel count and fails the verified bar. Week seven for the flagship; graded Wednesday against sources fixed at the window.
The ULSD-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Fri Aug 28 -> distillate leadership persists
Settles (front-month ULSD x42 minus Brent): 89.66 / 91.08 / 90.00 / 90.36 — never within $19 of the bar. The Aug 24 collapse to ~83 recovered within two sessions and its driver was never publicly attributed; the desk's tape shows the recovery, and we print the unexplained dip rather than pretending it had a story. Five graded weeks of distillate leadership.
IF no verified Hormuz halt occurs through Fri Aug 28, THEN Brent settles inside $85-100 every session -> the premium oscillates but neither breaks; conditional, VOID if the halt occurs
Condition fired (no verified halt); every settle inside: 88.58 / 87.84 / 89.70 / 88.10 against $85-100. Second consecutive conditional band to survive intact under post-settle anchoring.
Gold settles above $4,500 every session through Fri Aug 28 -> the rates regime holds its ground through the Warsh debut
THE GRADE WITH TWO FOOTNOTES, both printed. (1) BASIS: the registered source is COMEX settles on the series the floor was computed from. As captured at registration and at window close (verified twice, Fri and Sun), the anchor was 4,697.10 and the window printed 4,638.10 / 4,598.20 / 4,609.70 / 4,529.90 — every session above $4,500, RIGHT by $29.90 at the tightest, absorbed on Warsh's hawkish Friday. Yahoo later REVISED two of those days (~$52 down: anchor to 4,640.80, Friday to 4,478.10) on the same contract label. On that revised basis the same formula yields a floor of 4,450 — and Friday's 4,478.10 clears it by $28.10. RIGHT on either internally consistent basis; it fails only by mixing the old floor with the revised tape, which is not a grading method. (2) THE CALENDAR: spot gold broke $4,500 on Friday itself (-$146 to ~4,454), and the December contract broke below on both bases in the two sessions AFTER the window (4,431.10, then 4,348.00). The ratchet was priced at one window-sigma and the window ended before the regime turn finished repricing. A calendar-saved win is still a win — and we say the calendar saved it.
Henry Hub natgas settles below $3.20 every session through Fri Aug 28 -> the gas-oil decoupling survives week eight
Settles 2.770 / 2.842 / 2.907 / 2.888 — never within $0.29. The decoupling survives week eight, the board's quietest and most reliable axis.
No verified physical halt of Hormuz transits (tracker-confirmed at-or-near-zero, confirmed mining, or verified full stop) on any day through Fri Aug 14 -> disrupted holds
No tracker reported a zero-transit day, verified mining, or verified stop on any day Aug 11-14. Counts ran single-digit-to-low-teens (Kpler-attributed 6-14/day; Lloyd's List 73 transits for the week of Aug 10-16) and the press citations conflict on exact dailies — but every reading from every source is nonzero, which is what the call requires. The window's near-zero readings (PortWatch 1-vessel days) fall on Aug 9 and Aug 16, outside the window. Graded during the Aug 21 catch-up; the count ecosystem's incoherence is printed, not hidden.
The ULSD-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Fri Aug 14 -> distillate leadership persists
Settles (front-month ULSD x42 minus Brent): 89.7 / 91.8 / 91.5 / 91.4 — never within $19 of the bar. Fourth straight graded week of distillate leadership.
IF no verified Hormuz halt occurs through Aug 14, THEN Brent settles inside $79-98 every session -> the premium oscillates on diplomacy but neither breaks; conditional, VOID if the halt occurs
Condition fired (no verified halt); every settle inside: 88.91 / 88.98 / 87.07 / 88.52 against $79-98. First conditional band to survive intact since the architecture was built — anchored post-settle, bounds one sigma out, exactly as designed.
Gold settles above $4,150 every session through Fri Aug 14 -> the rates-driven regime is real, not a two-print spike
The inverted call: gold settled 4,383.00 / 4,408.90 / 4,363.60 / 4,380.40 — never within $200 of the 4,150 floor. The rates regime is real: confirmed by the soft July CPI (Aug 12) and a further leg to ~4,650 the following week on September-cut positioning ahead of Warsh's Jackson Hole debut.
Henry Hub natgas settles below $3.20 every session through Fri Aug 14 -> the gas-oil decoupling survives week seven
Settles 2.767 / 2.804 / 2.727 / 2.733 — never within $0.40. The decoupling survives week seven.
No verified physical halt of Hormuz transits (tracker-confirmed at-or-near-zero, confirmed mining, or verified full stop) on any day through Fri Aug 7 -> disrupted holds
No tracker-confirmed zero day, verified mining, or verified stop on any day Aug 4-7. Counts (all Kpler-attributed): ~5 on Aug 4-5, 8 on Aug 6, 8 on Aug 7 — thin, dark-heavy (41% AIS-dark), and nonzero every day. The count-publication darkness of early August ended mid-week. Grading caveat printed: the tracker ecosystem's five sources use non-comparable windows and do not arithmetically reconcile; every reading is nonetheless solidly nonzero.
The ULSD-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Fri Aug 7 -> distillate leadership persists
Settles (front-month ULSD x42 minus Brent): 79.00 / 79.99 / 80.55 / 80.35 — never within $9 of the bar. Distillate leadership held a third graded week.
IF no verified Hormuz halt occurs through Aug 7, THEN Brent settles inside $74-94 every session -> the premium oscillates but neither breaks; conditional, VOID if the halt occurs
REGISTRATION FAILURE, disposed per the desk's own standard. The band ($74-94) was computed at the Aug 3 settle but a publish failure kept it off the public page through the entire window. Every settle (79.36 / 79.45 / 82.49 / 83.55) landed INSIDE the band — it would have graded RIGHT — and it is VOIDED anyway: a call the readers could not see before resolution does not count, in either direction. Full disclosure published on the W32 page Aug 10.
Gold settles below $4,150 every session through Fri Aug 7 -> the no-war-bid regime holds a sixth week
The bar broke on Wednesday Aug 5's settle (4,245.80) and every settle after (4,242.00 / 4,340.70); gold printed ~4,397 the following Monday. Graded WRONG on its public terms (text and bar were registered; the exact odds, computed at 50%, were caught in the registration failure — excluded from calibration statistics). The honest attribution cuts deeper than the grade: the break was driven by the Aug 5 ADP miss and Aug 7 NFP print of -23,000 collapsing Fed hike odds and the dollar — a rates repricing, not a war bid. The no-war-bid THESIS outlived its own call; the call's falsifiable bar broke, so WRONG it is.
Henry Hub natgas settles below $3.20 every session through Fri Aug 7 -> the gas-oil decoupling survives week six
Settles 2.682 / 2.688 / 2.640 / 2.662 — never within $0.51 of the bar. The decoupling survives week six; gas ignored both the war and the rates repricing.
No verified physical halt of Hormuz transits (tracker-confirmed at-or-near-zero, confirmed mining, or verified full stop) on any day through Fri Jul 31 -> disrupted holds, by however thin a margin
Graded on tracker counts: Kpler 12 crossings Jul 28 and 14 on Jul 29 — the week's highest, through the heaviest US strikes of the war and two IRGC closure declarations. Jul 30-31 numeric counts were never published by any priority tracker (Kpler characterized Jul 30 only as 'a sharp fall' after the rebound); no tracker reported a zero-transit day, verified mining, or a verified stop for any session in the window. The bar was a VERIFIED halt; none exists. Coverage gap printed, not hidden: two of five window days carry no published number.
The ULSD-diesel-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Fri Jul 31 -> distillate leadership persists
Settles (front-month ULSD x42 minus Brent): 90.25 / 92.80 / 87.76 / 82.98 on the post-roll displayed series (captured live Jul 28 pre-roll: 84.32). Never within $12 of the $70 bar on any candidate series. Distillate leadership held through both the crash and the spike. The Jul 31 contract roll re-displays history, so the series is named rather than assumed.
IF no verified Hormuz halt occurs through Jul 31, THEN Brent settles inside $88-108 every session -> the two-front premium is real but still supply-capped; conditional, VOID if the halt occurs
Condition fired (no verified halt); consequence failed on session one. Jul 28 settled below the $88 floor (84.09 post-roll display; 85.98 captured at the time — both under 88) after Monday's 8.7% collapse parked spot on the floor before the window opened. The next three settles — 90.74 / 89.03 / 90.12 — sat back inside the band. A band broken by 24 hours of diplomacy, re-entered the day the pause died. The conditional architecture again contained the miss to its own leg.
Gold settles below $4,150 every session through Fri Jul 31 -> the no-war-bid regime holds a fifth week
Settles 4,036.30 / 4,034.70 / 4,100.10 / 4,049.10 — every session under the 4,150 bar (displayed Dec-26 series post-roll; nearer months settled at or below in contango). The bar WAS pierced intraday: ~4,162 printed Thursday about 30 minutes after the settlement window, and faded. The call resolved on settles per its registered text and survived on exactly that rule. The no-war-bid regime holds a fifth week — bent, not broken.
Henry Hub natgas settles below $3.20 every session through Fri Jul 31 -> the gas-oil decoupling survives week five
Settles 2.662 / 2.725 / 2.758 / 2.747 — never within $0.44 of the bar. Gas ignored the sharpest escalation of the war entirely; summer fundamentals own this price. The decoupling survives week five.
Iran's declared closure does not become a verified physical halt (transit counts at or near zero, confirmed by maritime trackers) on any day through Jul 27 -> disrupted, not closed, holds even under the Jul 19 re-declaration
Survived by ONE tanker. Kpler counted 1 crossing on Jul 23 (VLCC New Giant, 2M bbl Basrah — the lowest since May 7) and 3/day Jul 22-24, against ~90 pre-war; gCaptain's Jul 24 data shows no zero day. An AI-aggregated tracker claimed a zero-transit day Jul 25 but self-discloses auto-composition and no Tier-1 source corroborates it — excluded per the verified bar. Iran's Jul 26 mine claim (Tasnim) is uncorroborated by CENTCOM/UKMTO, same pattern as Jul 19. Reduced is not halted, but barely. (gCaptain/Kpler Jul 24, CNBC Jul 21/23.)
The ULSD-diesel-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Jul 24 -> distillate leadership persists under escalation
Held every settle with room; Friday's crack sat near $79 (CME ULSD x42 minus ICE Brent). Distillate leadership intact through a second crude repricing week.
IF no verified physical halt occurs through Jul 27, THEN Brent settles inside $80-95 every session -> the repriced-but-capped regime; conditional, graded condition and consequence separately, VOID if the halt occurs
Condition FIRED (no halt verified), consequence FAILED: settles 89.22 / 91.01 / 94.07 / 100.69 / 96.78 — Jul 23 and 24 both above $95. The FOURTH consecutive upside price surprise (78, 88, 96.78, 100.69) — but the first fully contained by the new architecture: conditional, non-flagship, graded on its own leg. The driver was a genuinely new fact no band anticipated: Houthi strikes on Saudi tankers in the RED SEA (Jul 22-23), threatening the Yanbu/Petroline bypass — a second chokepoint front. (CNBC/Bloomberg Jul 23.)
Gold settles below $4,150 every session through Jul 27 -> the no-war-bid regime holds; gold trades rates and the dollar, not the strait
Settles 4,010.30 / 4,071.10 / 4,146.90 / 4,046.60 / 4,067.60 — the Jul 22 print came within $3.10 of the bar and backed off. Even a $100-Brent, two-chokepoint week produced no war bid in gold; the no-war-bid regime is now confirmed at the war's most escalated prices. (COMEX settles; graded on data through publish per desk precedent — W31 windows move to Friday ends to remove the Monday ambiguity.)
Henry Hub natgas settles below $3.20 every session through Jul 27 -> the gas-oil decoupling survives its hardest test yet
Settles 2.86-2.92 all week, never within $0.28 of the bar; gas closed DOWN 1.4% in a week crude rose 9.9%. The decoupling held through its hardest test. (NYMEX settles; same Friday-window note as gold.)
Iran's declared closure does not become a verified physical halt (transit counts at or near zero, confirmed by maritime trackers) on any day through Jul 20 -> disrupted, not closed; the tell is traffic, not statements
No tracker-confirmed halt on any day. Transits fell to two-month lows — Kpler counted 14 on Jul 12 and 7 on Jul 17, versus ~88-94 pre-war — and RFE/RL called it a 'near halt', but every named tracker showed nonzero flow all week; Bloomberg (Jul 16) documented dark-AIS ships and Oman ship-to-ship transfers continuing. The IRGC re-declared the strait closed Jul 19 and claimed two tankers destroyed; CENTCOM disputed it and no tracker confirmed a zero-transit day. Reduced is not halted; the verdict stays Disrupted. (Kpler/CNBC Jul 13, Bloomberg Jul 16, CNBC Jul 17, RFE/RL Jul 18.)
Brent front-month settles inside a $70-85 band every session through Jul 20 -> the oversupply-caps-war-premium equilibrium holds
Settles Jul 13-17: 83.30 / 84.73 / 84.95 / 84.23 / 88.10 — the Jul 17 close broke the $85 ceiling three days before the window ended. Third consecutive same-direction price miss (W28 called <75, settled 78; W29 called <=85, settled 88.10): the model's short-war-premium bias, named and now retired — the flagship moves off raw price bands from W30.
The ULSD-diesel-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $65 on every settle through Jul 17 -> distillate stays the barrel's tightest link
The crack held with room on every settle: $77.29 / 83.87 / 80.88 / 85.06 / 82.61 Jul 13-17 (CME ULSD x42 minus ICE Brent front-month, the pre-registered sources) — never within $12 of the $65 bar. Distillate stayed the barrel's tightest link through the escalation, exactly the transmission the call tested.
No dated, both-sides-confirmed US-Iran negotiating round is scheduled through Jul 20 -> diplomacy stays frozen under the escalation
No round convened or was scheduled with a date both sides confirmed. The floated 'Jul 18 resumption' traced to a single unnamed-source report (Kurdistan24, Jul 2) and never materialized: Iran's deputy FM called the Islamabad MoU 'suspended' on Jul 18 after US strikes killed ~50, and by Jul 19 Washington described only 'mixed signals' amid a ninth night of strikes. (Al Jazeera Jul 18, CNN Jul 19.)
The Crude Signal crisis score prints at or above 72 on every daily close through Jul 17 -> the crisis regime persists
Printed 78 on every daily close Jul 13-17 (public archive, /data). The crisis regime persisted through the declared closure's first full week.
No new US-Iran kinetic strike AND no Iranian attack on or physical blocking of transiting vessels through Jul 13 -> the truce survives the funeral's end
Both legs failed. Strikes resumed Jul 7-8 (three commercial tankers struck in and near the strait; the US hit Iranian air-defense and anti-ship targets), the US declared the June 28 stand-down over Jul 8, and on Jul 12 Tehran formally declared the strait closed to shipping. New US-Iran kinetic strikes AND Iranian attacks on transiting vessels both occurred (Reuters/AP/Bloomberg, Jul 7-12). This was the 60 pct tail-risk call that the truce would survive the funeral's end; it broke decisively.
A new US-Iran negotiating round convenes or is formally scheduled (dated, both sides confirmed) by Jul 13
No round convened, and none was scheduled with a specific date confirmed by both sides. Escalation replaced diplomacy: the US struck Iran Jul 7-8 and again over the weekend, Tehran denied reports that talks were continuing (Washington Times, Jul 10), and only unnamed regional officials floated a possible Jul 18 resumption. No dated, both-sides-confirmed round existed by Jul 13 (Al Jazeera/CNN/CNBC, Jul 10-13).
Brent front-month closes below $75 every session through Jul 13
Brent breached the $75 line on settles after the Jul 7-8 re-escalation. Closes Jul 8-10: 78.02 / 76.30 / 76.54 (Crude Signal daily series); the Jul 8 settle of 78.02 alone fails 'every session.' Only Jul 7 (74.16) closed below the line.
Mojtaba Khamenei makes a verifiable public appearance (in person or new video/audio; written statements do not count) by Jul 13
No verifiable appearance surfaced in the window. Ali Khamenei was buried Jul 10 at Mashhad; Mojtaba was barred from and absent at the funeral over assassination-risk fears, with no image, video, or audio released. Coverage through Jul 13 was still speculating whether he would appear at a later memorial (Times of Israel/SCMP/Business Today, Jul 9-12). Written statements are excluded by the call; none of the qualifying forms (in-person, video, audio) were verified.
Brent front-month closes below $80 every session through Jul 6, including across the Jul 5 OPEC+ meeting -> deflation regime holds
Closes Jun 29-Jul 3: 73.15/72.92/71.57/71.80/72.13; Monday Jul 6 71.88 after the seven-member OPEC+ +188k bpd August decision (Rigzone/Interfax Jul 5-6). Never approached 80.
Hormuz transits stay below ~80/day sustained through Jul 6 (AIS + any CENTCOM count) -> reopening stays impaired by renewed hostilities
AIS (PortWatch) 27-52/day across the window; press aggregation of Windward/CENTCOM 25-35/day; Jul 4 count 25 with 6 diversions on IRGC route warnings. Both source classes far below the 80/day bar.
DXY closes above 100 every session through Jul 6 AND CME year-end hike odds stay above 60% -> Warsh's hawkish regime holds
Dollar leg clean: closes 101.11/101.19/101.39/100.86/100.86, never below 100. CME year-end hike-odds leg UNVERIFIABLE at resolution: primary FedWatch tool unreachable, secondary trackers quote non-comparable metrics. Graded on the verifiable leg, labeled.
The Jun 28 US-Iran stand-down holds through Jul 6: no new US-Iran kinetic strike and no Iranian physical (vessel-blocking) closure -> ceasefire survives its first full week
No new US-Iran kinetic strike and no Iranian attack on or blocking of transiting vessels Jun 29-Jul 6 (Reuters/AP/CNN wire sweep); the pause was formalized as an explicit one-week funeral truce with talks resuming after Jul 9.
DXY closes above 100 AND the 2y stays above 4.10% every session through Jun 29 -> Warsh's hawkish reset sticks, dollar bid holds
DXY held >100 all week and hit a 13-mo high ~101.6 (Jun 25); but the 2y slipped to 4.09% on Jun 25 (in-line PCE trimmed hike premium), nicking the >4.10%-every-session threshold by 1bp. Thesis right (hawkish reset stuck, dollar bid at 4-mo oil lows); never neared the 4.0% invalidation. Sources: Investing.com DXY, Trading Economics 2y.
Brent front-month closes below $85 every session through Jun 29 -> deflation regime holds
Clean. Brent closed below ~$78 every session and fell toward $72 by Jun 26 (WTI sub-$70 for the first time since Feb), never near $85, holding its lows even through the weekend US-Iran strikes. Source: Yahoo Finance front-month closes.
Hormuz transits stay below ~70/day sustained (clearly under ~94 pre-war norm) through Jun 29, on both AIS + CENTCOM counts -> reopening stays partial
Held with the dual-source/looser bar. Transits briefly touched ~70/day mid-week before the Jun 25 Ever Lovely strike reversed momentum, then ~40 (Jun 27), ~28 (Jun 28), ~22 (Jun 29), far under ~94 norm and below 70. Sources: Windward, IMF PortWatch.
US-Iran MOU holds through Jun 29: no US-Iran kinetic resumption and no Iranian physical (vessel-blocking) closure -> $110 closure trade stays dead
Missed. The MOU broke into a kinetic exchange: US strikes Jun 26, drones hit Ever Lovely (Jun 25) and tanker Kiku (Jun 27), US counter-strikes, IRGC missiles at US bases in Kuwait and Bahrain (Jun 28), then a stand-down. Lowest-conviction call (60%) on our weakest type. Sources: CNBC, Al Jazeera, CNN.
FOMC holds 3.50-3.75% and the SEP dot plot erases the last 2026 cut (hawkish hold, no hike) -> DXY holds 99-100, 2y ~4.0-4.1%, risk grinds higher
Condition right: the Fed held 12-0 and the median 2026 dot moved past zero cuts to imply a hike. But we under-priced the consequence. Warsh out-hawked the call: DXY broke 100 to ~100.8 (not the 99-100 band), the 2y cleared 4.19% (above ~4.0-4.1%), and equities sold off on FOMC day rather than grinding higher. Right thesis (hawkish hold, cut erased), too timid on magnitude.
Brent front-month closes below $90 every session through Jun 22 (peace-premium unwind holds as the dominant regime)
Clean. Brent traded ~$78-81 all week and never approached $90; the peace-premium unwind held as the dominant force. Brent ~$78.6 Jun 22.
Hormuz daily transits stay below 47, half the ~94 pre-crisis baseline (reopening stays paper-only on mine-clearing)
Honest split. Our pre-registered source (AIS-based PortWatch/Windward, ~20/day) held under 47, but CENTCOM's more complete military count (dark-vessel inclusive) hit 55 on Jun 20. By the named source the call holds; by the fullest count the threshold was touched. Strait still a trickle vs ~94 norm, but graded down for the ambiguity.
US-Iran ceasefire holds through Jun 22: no US-Iran kinetic resumption, no formal Iranian withdrawal -> $110 closure trigger stays dead
Our lowest-conviction call (65%, weakest type), landed. No US-Iran kinetic exchange since Jun 9-10; Iran stayed at the Switzerland table and signed the MOU; oil fell to $78. The Jun 20-21 re-closure declaration was an MOU violation, not a framework withdrawal.
May CPI (Wed Jun 10) hot: headline +0.4% MoM or core +0.3%+ -> hike tail goes live, DXY breaks 100
Headline fired hot (+0.5% MoM, +4.2% YoY) but core printed soft (+0.2%). Consequences failed: DXY stalled at 99.5 with no break of 100, and gold rose toward $4,340 rather than testing $4,250. The energy-driven headline was undercut by the soft core and the Iran deal collapsing the oil input. Right on the headline, wrong on every market consequence.
May CPI soft -> hike tail fades, relief rally
Core fired soft (+0.2% MoM). DXY unwound to 99.5 and gold reclaimed toward $4,400 as called, but the 'oil handed back to the strait' clause missed when the peace framework sent oil lower instead. Right on the dollar and gold, wrong on oil.
Iran takes closure action -> Brent gaps
Condition never fired. Iran's Jun 11 'total closure' declaration and Jun 12 drone fire (both downed) laid no new mines, seized no vessel, and transit did not newly collapse. A peace framework landed instead and Brent fell to $83. An unfired condition is void, not a miss.
Pause-break fade trade
Condition never fired on its terms: the direct Iran-Israel missile axis held all week, and the disabled Settebello was Palauan-flagged (a US blockade action), not a US-flagged ship hit by Iran. Void.
EIA shows SPR sub-350 -> runway headline escalates
Condition fired exactly: the SPR drew 7.9mb to 349.2mb, under 350 as called. Consequence failed: WTI fell to ~$80 rather than holding the predicted ~$90 floor, as the peace framework made the reserve-runway math irrelevant overnight. Right on the reserve, wrong on the floor.
Iran acts on closure (mines, seizures, transit to zero) -> Brent gaps through $110
Iran never acted; no gap occurred and Brent slipped to ~$94.5 (Jun 8 close $94.24). The necessary-condition read held: no action, no gap.
Macron channel reopens or ceasefire resumes -> Brent unwinds toward $88
Oil softened to the low-$90s but diplomacy stayed broken; the easing came from jobs + OPEC+ barrels. Right direction, wrong driver.
US-Iran strike with US vessel hit, Gulf base casualty, or tanker sinking -> Brent $115+, risk-off
Casualty leg fired (Kuwait airport, 1 dead Jun 2; Haifa, 4 dead Jun 7); Brent FELL (Jun 5 close $93.09). The macro regime overrode the geopolitical tail.
May jobs <100K or unemployment 4.4%+ -> cut path reopens, DXY rolls to 98
Payrolls +172K, unemployment 4.3% - the mirror image; cut path slammed shut, a hike got priced, DXY firmed to ~99.9. The framing correctly named NFP as the pivot.
EIA 8mb+ SPR draw -> runway alarm goes from Q3 risk to certainty
SPR down 7.9mb to 357.1mb, right at threshold; commercial crude -8mb. Runway is now a standing headline.
Iran mine-clears or US suspends blockade -> Brent breaks $90 within 48h
Neither happened; Brent tagged $91.12 Friday on no-deal + soft GDP, not on execution (May 29 close $92.05). Wrong mechanism.
Israel-Hezbollah escalates or Iran rejects MoU -> Brent reclaims $105
Both legs fired (Beaufort Castle May 31; talks suspended); Brent surged ~7% to $97, not $105. Right trigger, short on magnitude.
New Home Sales / Durables miss -> DXY to 98.50, gold $4,600
Home sales missed but Durables beat +7.9%; DXY held ~99, gold fell to $4,460.
Core PCE +0.3% MoM or headline 2.7%+ -> DXY tests 100, gold breaks $4,500
Core softer at +0.2%, headline hot 3.8%; gold broke $4,500 (to $4,460) but DXY never tested 100. Gold leg landed, dollar leg didn't.
SPR draw <5mb signals deal confidence; 6mb+ signals hedging
SPR drew 9.06mb to 365; the hedging branch fired and the deal collapsed days later. Read diplomacy correctly when headlines didn't.
PF Plus clears convoy with no engagement -> Brent -$5-8
PF Plus did not restart; Brent gave back $15 anyway on deal optimism (May 22 close $103.54). Right direction, wrong mechanism.
IRGC engages convoy or US vessel -> Brent $115+
No US escalation; Brent did the opposite, cracking to $97. Graded against the standing $115 template.
UMich <47 or Philly/Empire miss -> DXY rolls to 98, gold $4,650
UMich crashed to 44.8 and Philly missed huge, but DXY ROSE to 99.32 on inflation expectations. Trigger fired, wrong direction.
Fed speaker reframes services-PPI sticky -> DXY tests 100
Minutes were hawkish; DXY touched 99.32, never cleared 100. Half-fire.
SPR 6mb+ draw to sub-378 -> WTI floor $100 holds
SPR drew 10mb to <375 but WTI broke $100 down to $93 (May 22 close $96.60) on the diplomatic surprise. Right number, wrong outcome.
CPI hot (headline >3.5% or core 3.4%+) -> 2y to 3.95%+, DXY 98.50, gold -$50
Core MoM +0.4% hot; 2y to 4.08%; DXY cleared 99; gold gave back $182. Cleanest single fire, per W21 scorecard.
Project Freedom restarts / US vessel fired on -> Brent $115+ in one session
MSV Haji Ali sank May 13; Brent hit $109.26 (May 15 close), not $115. Half-fired.
Trump accepts modified Iran proposal -> Brent gaps to $90
Trump dismissed it; Brent went to $109+.
Warsh dovish confirmation -> DXY breaks 97
Warsh confirmed 54-45, but hot CPI/PPI overwhelmed; DXY cleared 99 instead.
CLARITY markup advances -> ETH catches BTC/SOL within 5 sessions
Markup advanced 15-9, but crypto sold off on CPI: BTC -6%, ETH -10%, SOL -12%.
SPR drained 5mb+ to sub-388 -> oil floor $100 firms
SPR drew 8.6mb to ~384; WTI held >$100, peaked $105.42 (May 15).
Saudi cuts -> Brent $110+, run at $120
Saudi did not cut; Brent ripped on conflict then cracked to ~$101 (May 8 close $101.29). Mechanism wrong.
Saudi releases capacity -> Brent breaks $105, tests $100
Brent broke $105 (May 8 close $101.29) but Saudi never released capacity. Right outcome, wrong cause.
Project Freedom succeeds -> WCI freight eases 2-3%
Project Freedom paused after 48h; WCI rose +3% to $2,286.
Fed speakers more hawkish -> DXY clears 99
Speakers were hawkish; DXY fell to a 10-week low of 97.84 (May 8). Hawkish-Fed-lifts-dollar mechanism broken.
Trump accepts Iran proposal -> Brent gaps -$15, tests $98
Trump rejected the proposal ('TOTALLY UNACCEPTABLE'); trigger never fired. Graded WRONG in the W20 scorecard.
Powell cites Iran/energy CPI risk -> DXY through 99, gold breaks $4,650
Per W19 scorecard: Powell cited energy; gold cracked $4,650 to $4,545; DXY tagged 98.92.
IRGC seizes US-flagged vessel -> Brent $115+ in one session
No US seizure, but Brent hit $118.03 (Apr 29) anyway on UAE/OPEC exit. Right idea, wrong trigger.
Pakistan restart confirmed -> Brent retests the $90 floor
Pakistan opened corridors but Brent went the opposite way, to $118.
Core PCE >3.0% YoY -> 2y yield +10bps
Core PCE 3.2%; 2y +exactly 10bps to 3.94%. Clean call per W19 scorecard.
CLARITY markup advances -> ETH/SOL re-rate vs BTC
Markup did not advance; BTC outperformed. W19 marks it Open.
If Tehran returns to Pakistan talks within the week, Brent holds $90-100
Graded by W18 scorecard: Brent held the called range intraweek, then moved to $101 through it.
If Iranian forces retaliate against a US-flagged vessel, the floor breaks down fast; Goldman's $120 reactivates
No US-flagged retaliation that week; Iran hit two non-US vessels. Condition unmet.
If DXY slips below 97 and gold tags $4,900, the structural thesis hardens
Per W18 scorecard: DXY held 98.0-98.8. Neither trigger fired.
If Iran retaliates against the US Navy, expect a rapid move toward $150
Multiple kinetic US-Iran exchanges occurred (W20, W23, W24); Brent peaked at $118.03 (Apr 29) and sat ~$94 by W24. $150 never printed.
If the ceasefire fails Tuesday night, expect oil to surge toward Goldman's $150 scenario
Apr 8 deadline passed without a deal; Brent FELL to $94.75 (Apr 8 close). Oil never approached $150 at any point in the crisis.
Relayed Goldman $115 base case, $150-200 if blockade extends to June
Third-party scenario, not an owned call. Brent in June 2026: ~$94. The $150-200 band never came close.
No verified physical halt of Hormuz transits (tracker-confirmed at-or-near-zero, confirmed mining, or verified full stop) on any day through Fri Sep 4 -> disrupted holds
Resolves by tracker counts (Kpler, gCaptain, Lloyd's List, PortWatch), Sep 4. Flagship, week eight — a SHORT board, registered Wednesday post-settle with a two-session window because the desk missed its Monday and owes a board, not a second gap. 60%: the corridor framework holds on paper while flow stays thin and contested.
The ULSD-over-Brent crack (front-month ULSD $/bbl minus Brent $/bbl) holds above $70 on every settle through Fri Sep 4 -> distillate leadership persists
Resolves by CME ULSD and ICE Brent settles, Sep 4. Anchor 100.70 — back near the record print after the unexplained Aug 24 dip; the widest cushion this call has carried. Sixth graded week, same bar.
IF no verified Hormuz halt occurs through Fri Sep 4, THEN Brent settles inside $92-100 every session -> the premium oscillates but neither breaks; conditional, VOID if the halt occurs
Resolves by ICE Brent settles, Sep 4, conditional on the flagship. Two-session window; anchored to today's 95.38 settle after a +8.7% three-session rip, bounds each >=1 window-sigma at 2.4%/day worst-case vol, MC-priced 0.73.
Gold settles below $4,550 every session through Fri Sep 4 -> the hawk repricing holds; the test flips with the regime
Resolves by COMEX settles (December contract), Sep 4. THE TEST FLIPS: W35 defended a floor and won by $29.90 on the last session before Warsh's hawkish debut broke the level for real (4,431.10 then 4,348.00 the two sessions after the window). A desk that kept betting the floor after the regime turned would be marking its own myth. The ceiling is $4,550 — one window-sigma above today's 4420.90 settle, same construction as ever, opposite direction. One settle above invalidates.
Henry Hub natgas settles below $3.20 every session through Fri Sep 4 -> the gas-oil decoupling survives week nine
Resolves by NYMEX settles (September contract), Sep 4. Anchor 2.957 — the tightest this call has ever run (8.2% of room). If the bar finally breaks, it breaks honestly; we do not retire a call because it got close.
Expect another sharp move upward in oil if the mid-April supply cliff hits
No discrete cliff event was ever declared; no price level or date to test.
A diplomatic breakthrough would likely cause a rapid reversal in oil
Directionally trivial; no threshold or date.
Methodology
- A call counts as falsifiable only if it states a level or event, and a horizon. Early issues (W14–W17) mostly relayed third-party scenarios; the Binary Triggers framework (W18 onward) made nearly every call gradeable. Both eras are shown — including the unfalsifiable statements, because pretending they weren't made would defeat the point.
- RIGHT — threshold met, direction correct, inside the horizon. PARTIAL — right direction or trigger, short on magnitude or wrong on mechanism. WRONG — wrong direction or wrong outcome. UNRESOLVED — window still open or condition never triggered.
- From W25 onward, grading gets stricter: a call whose trigger condition never fired grades VOID instead of WRONG, so the ledger separates "the setup never happened" from "we read it wrong."
- Grades are written by the newsletter itself in the following week's scorecard, then re-verified against daily close data. Published grades are never changed afterward.
Nothing here is investment advice. See About for the full disclaimer.
Get the next call before it's graded.
A handful of specific, falsifiable calls every Monday, then the verdict the week after. Free.
Subscribe Free