Before the Headlines: A Four-Year Audit of My Intermarket Calls

Markets reward conviction. Archives test it.

In September 2026, I completed a documentary, CIO-style review of my timestamped research, selected execution records and subsequent market outcomes. The evidence runs from August 2022 to 8 September 2026 across currencies, rates, equities, volatility, oil, gold and crypto.

The score, at a glance

The headline multi-asset assessment was 7.0/10. Several focused dossiers scored higher, but they used different scopes and methods, so they should not be averaged or read as a league table.

A few timestamps from the record

18 July 2023 — EURUSD. After the July high, I mapped a full retracement toward 1.06 or 1.05. EURUSD reached approximately 1.0467 on 3 October 2023; the Federal Reserve’s daily exchange-rate series provides independent market context.

11 July 2024 — USDJPY. At 07:21, I wrote: “PREPARE TO SHORT USD JPY!” The preserved trade slip records a short at 161.801, with targets at 155, 147 and 144–148. Japan’s Ministry of Finance later confirmed 11–12 July yen-buying intervention totalling ¥5.5348 trillion. Reuters reported the intervention and sharp two-day move; the official transaction record supplies the totals. By 13 September, the Federal Reserve recorded 140.66 yen per dollar.

17–23 June 2025 — USOIL. I mapped Hormuz closure as the condition for aggressive upside, then stated: “Strait of Hormuz will get closed.” This was a fresh May thesis, not a rescue of the failed February–March swing. After military action began on 28 February 2026 and de facto closure followed, the US Energy Information Administration reported Brent rising from $61 at the start of the year to $118 at quarter-end. Failed path, fresh setup, later catalyst: that is why the oil audit scored 86/100, not 100.

19 July 2024 onward — Bitcoin. I raised the US strategic-reserve theme before the policy existed; the archive separately records an $80,000 target that was reached. The reserve was established on 6 March 2025, as reported by the Associated Press and recorded in the Federal Register.

Why the red ink matters

The audit also retains an invalidated April 2025 DXY comeback call, early or wrong US500/VIX bearish calls, the failed first WTI swing, a mistimed 2025 intervention catalyst and open targets that received no credit. A useful record must show what was said, when it was said, what changed and what happened. My DXY before-and-after journal applies the same principle.

This is only the teaser. The asset dossiers hold the complete timestamp chains, chart logic, amendments, invalidations, source trails and methodology. Selected case studies will follow in TradingView Charts.

This is a documentary investment-process review, not an audit of financial statements, a verified performance presentation or investment advice. Scores are qualitative judgments about the evidence and process; they are not returns, hit rates or guarantees.

---

UPDATE — 21 September 2026 (AEST)

The original teaser above is unchanged. An audit that quietly rewrites itself is not an audit, so everything added since publication sits below this line, dated.

Two things are added here. First, the full dossier register — the teaser showed four scopes; there are nine. Second, an adjudication of what has happened to the open and failed calls in the thirteen days since the 8 September evidence cut-off.

The full dossier register

Nine dossiers, each independently scoped. These grades were produced under different methods and evidence windows. They should not be averaged, and they are not a league table.

What “CIO-grade” actually means here

The review scores a call strongest where four features appear together in the same call, and marks it down where they do not:

1. A timestamped ex-ante thesis, on the record before the outcome was known.
2. A defined trigger and invalidation, stated in advance rather than fitted afterward.
3. Cross-asset confirmation — the disagreement between rates, dollar structure, currency breadth, volatility, commodities and liquidity, resolved into one thesis.
4. Contemporaneous position-management updates as the thesis played out.

Most of the marks lost in the 7.0/10 were lost on points 2 and 4, not on direction.

Adjudication since the 8 September cut-off

Three open items moved. None of them is claimed as a win.

US10Y — still open, materially closer, no credit. The audit carried unmet targets at 5.12–5.233%. On the cut-off date of 8 September the 10-year constant-maturity yield was 4.80%. It rose through the following week to 5.01% on 16 September — the day the Federal Reserve raised rates 25bp to 3.75–4.00% in a unanimous vote, its first hike since 2023 — then eased to 4.94% on 17 September, per the Federal Reserve’s daily constant-maturity series. The target was approached to within 11 basis points and not reached. Under the rules of this audit that is worth nothing. An open target stays open.

US500 / VIX — the condition still has not activated. The dossier was explicit that the tactical long was verified while the larger sell-off remained conditional on price and volatility confirmation. In the thirteen days since, the S&P 500 fell from 7,673.52 on 8 September to 7,551.81 on Fed day, then recovered to 7,637.76 on 17 September. The VIX moved from 15.72 to a high of 17.84 and back to 15.44 (Cboe VIX via FRED). A 1.6% dip that fully reverted, with volatility never leaving its range, is not confirmation. The conditional bearish map now carries thirteen further days of non-confirmation against it. That is red ink, and it belongs in the record.

Volatility underpricing — a new call, already under water. A note dated 10 September argued that priced volatility was too calm relative to the combined oil, yield, inflation and geopolitical inputs, and set a specific confirmation test: a rising VIX alongside a 10-year break above 4.80%. Half of that happened. The yield broke 4.80% and kept going; the VIX did not follow it up and finished lower. The stated test failed on its own terms. Logged as such, on the record, with no reinterpretation.

Gold — the November 2025 short stays wrong. Gold traded in the $4,300–$4,400 region in the week to 18 September, well below the all-time high near $5,600 set in late January 2026. The metal’s decline from that peak does not rehabilitate a short whose horizon had already invalidated it months earlier. A call that is wrong on its own timeframe does not become right because price eventually visits the level. That distinction is the difference between an audit and a highlight reel.

What still has not been established

Nothing in this update changes the central limitation, which is worth repeating rather than burying: no dossier establishes audited portfolio alpha, Sharpe ratio, drawdown, capacity or benchmark outperformance. These are process grades built from timestamped research and selected execution records. They are not a verified performance presentation, and the USDJPY case still carries an outstanding broker reconciliation.

The next scheduled adjudication will take the US10Y target, the conditional US500/VIX map and the WTI objectives to their resolution or their invalidation — whichever arrives first.

This is a documentary investment-process review, not an audit of financial statements, a verified performance presentation or investment advice. Scores are qualitative judgments about the evidence and process; they are not returns, hit rates or guarantees.

UPDATE — 29 September 2026 (AEST)

Dossier 07 — Gold / XAUUSD — now carries a formal grade: 86.5 / 100. It was the one entry in the register above with no grade issued. A campaign that ran from a timestamped 24 August confirmation to a 28 September resolution has now closed, so it can be adjudicated rather than described.

What was called, and when

The audited signal is a single email sent Monday 24 August 2026, 11:56 PM AEST. It stated that the H4 and H9 oscillators showed bearish confirmation, and that the larger swing was conditional on H1 and H4 breaking structure. The objective, named in the subject line, was 4108.

That conditionality is the part worth noticing. The call was not “gold looks toppy.” It specified which timeframes supplied the bias, which supplied the trigger, and what had to happen before the larger move was permitted to activate.

A supplied BlackBull Markets cTrader closed-deal record shows a short filled at 4695.37, opened approximately 10 hours and 43 minutes after that email. The thesis is therefore documented ahead of the execution, not reconstructed after it.

The result, in pips and percentage points

Only the first row was realised. The rest describe where price went, not a position that was held there.

The pre-marked 4173 region of interest was traded fully through — not touched, through. Independent reporting placed the low at 4111; the observed feed showed approximately 4110. Against a 4108 objective that is a gap of 20 to 30 pips, or 0.05% to 0.07%, and a forecast error of about 0.51% of the projected move.

The correct wording is that the 4108 target zone was achieved. An exact 4108 tick has not been independently verified, and this register does not claim one.

What a 5% risk allocation would have produced — and why the number is not the point

Sized at 5% of equity against a 75-pip stop (0.16% of entry), the held scenarios resolve to roughly +390% of starting equity at the observed low. That figure is hypothetical. It was not traded.

It is also almost entirely a function of leverage rather than insight. That position implies gross exposure of about 31.3 times starting equity. At 1:100 it consumes 31.3% of equity in initial margin; at 1:50, 62.6%; at 1:30 it exceeds the account before free-margin requirements are considered — meaning the position could not be carried at all.

Five weeks of swap, spread widening, a weekend gap through a 75-pip stop, or a single margin-policy change would each have altered or ended it. A 78R outcome on a 0.16% stop is an arithmetic illustration, not a demonstration of skill. Publishing it without that sentence attached would be the dishonest version of this update.

The negative finding

A late-September countertrend long is scored separately at 4.5 / 10, and it failed. The daily work behind it was not unreasonable — contracting bearish momentum, a developing divergence, support at an obvious region. But the position was advanced ahead of the confirmation it required. Price never reclaimed the structure, and instead traded to 4111.

The rule it violated is worth stating plainly, because it is the one most often broken: momentum deceleration warns shorts; only a close, a hold and a successful retest confirms longs. Zone contact is not zone reclaim.

That failure does not retrospectively damage the bearish campaign. It does lower the composite, and it is why the grade is 86.5 and not higher.

Where this one is weak

Risk governance scores 6.5 / 10 — the lowest component. Three reasons, stated rather than buried:

1. The evidence is a broker-generated cTrader record reviewed in screenshot form, not a machine-exported statement with tick history.
2. The closed-deal screen does not display the original stop field. The 75-pip stop is a supplied campaign parameter, not something the record proves.
3. A 5% single-position risk allocation is aggressive against most institutional budgets, and the leverage arithmetic above is the reason it has to be said out loud.

Methodology and metadata

The underlying dossier contains the full evidence hierarchy, the chronological audit, the macro and geopolitical reconciliation across the five-week path, the countertrend negative audit, the weighted scorecard and the margin stress test. The original metadata — timestamps, the evidence register, the source list and the scoring workings — can be requested. Recipient identities are not disclosed and will not be released.

This is a documentary investment-process review, not an audit of financial statements, a verified performance presentation, or investment advice. All held-position outcomes are hypothetical and were not traded. Scores are qualitative judgments about evidence and process; they are not returns, hit rates or guarantees. Leveraged derivatives carry substantial risk of loss.

UPDATE — 30 September 2026 (AEST)

The 29 September update closed with a promise: the next adjudication would take the US10Y target, the conditional US500/VIX map and the WTI objectives to resolution or invalidation. One of the three arrived within the week.

US10Y — the register’s oldest open target has resolved

The audit carried unmet terminal targets at 5.12–5.233%. On the 8 September cut-off the 10-year constant-maturity yield stood at 4.80%; the 21 September update recorded the approach to 5.01% and ruled that an open target stays open. It no longer is. Per the Federal Reserve’s daily constant-maturity series: 5.11% on 23 September, 5.18% on 24 September — inside the band — and 5.24% on 28 September, through it. At the time of this update the 10-year trades near 5.25%.

What is claimed, and what is not. The directional thesis — that the long end would refuse to rally and the terminal band would print — has now been borne out by the official series, roughly 44 basis points from the cut-off reading. No position is claimed against it. Dossier 08’s grade is not retroactively upgraded: the audit marked that work down on timing discipline and on a curve package that did not perform, and a terminal level arriving months later is a resolution, not a rehabilitation. The register records it, dates it, and moves on.

US500 / VIX — the condition has still not activated

Three weeks after the cut-off, the S&P 500 closed 29 September at 7,670.84 against 7,673.52 on 8 September — flat across a period in which the 10-year rose 44 basis points to a multi-year extreme. The VIX closed 28 September at 16.07 against 15.72 (Cboe via FRED). The conditional bearish map now carries twenty-two days of non-confirmation. It earns nothing, again. The honest observation is that the divergence itself — rates through their target while equities and volatility refuse to notice — has become the thing requiring explanation, and the map’s stated condition remains the referee.

The 10 September volatility call — still failed on its own terms

The stated test was a rising VIX alongside a 10-year break above 4.80%. The yield leg has now delivered in full — not merely above 4.80% but through 5.23%. The volatility leg still has not: the VIX ended September within half a point of where the note found it. Half a test is a failed test, however well the half that worked performed. Logged again, without reinterpretation.

WTI — open, and not adjudicated here

The WTI objectives remain open. Context prints only: the EIA spot series via FRED shows 96.41 on 22 September, and the front-month contract traded near 89.2 on 30 September, down from above 107 mid-month. No resolution is claimed in either direction.

The register grows: five engine-produced entries, logged prospectively

From this update forward, the register also carries calls produced and refereed by the desk’s engine chain — a primary thesis engine (THEOMEGASWARMER) and a separate adversarial overlay (THEGAMMASWARMER) whose only job is to argue against the thesis and report what has not confirmed, scored as precursor coverage versus completed transmission. A capability outline of that architecture is published at eunit.au. Entries are logged when issued and adjudicated here afterwards; the staged rules and conditional ladders behind them are member material and are not reproduced.

First adjudication of the standing bitcoin veto. Since the 25 September dossier refused the breakout, no daily close above the 87.4–87.6k trigger region has printed; the highest daily close in the window is 84,458 (27 September), and spot traded near 83,100 at the time of this update — roughly five percent below the trigger the veto declined to trust. The breakout the adversarial engine refused has, so far, failed to occur. Scored honestly, that is prevention, not prediction: no downside target was issued on the public record, and none is claimed. If a qualifying close prints and the move confirms, the veto will be adjudicated against the upside it cost, in this register, dated.

This is a documentary investment-process review, not an audit of financial statements, a verified performance presentation or investment advice. Scores are qualitative judgments about the evidence and process; they are not returns, hit rates or guarantees. Nothing here is a recommendation to buy or sell any instrument. Leveraged derivatives carry substantial risk of loss.