Thesis ledger
Dated forecasts and market interpretations with assessment criteria and subsequent outcomes. This records research, not portfolio performance.
Publication range: 2 Jul 2026 – 5 Nov 2026. Source: Hawk Thorne research record. Thesis descriptions retain the content recorded in the ledger.
Forecasts
25Claims about future events, assessed against a stated condition and time horizon.
- Held
- 14
- Broke
- 9
- Unresolved
- 0
- Awaiting assessment
- 2
Interpretations
134Assessments of current conditions. New evidence can sustain, revise or retire an interpretation. These do not contribute to forecast accuracy.
- Sustained
- 12
- Revised
- 26
- Retired
- 14
- Awaiting assessment
- 53
- Historical assessments
- 29
Assessment rules and forecast calibration
Since 27 July 2026, forecasts and interpretations follow separate assessment rules. Earlier interpretations graded against price thresholds remain in the history and are excluded from calibration.
Declared probabilities compared with assessed outcomes. Brier is the mean squared error; a lower score indicates less error within this sample. A small sample does not establish future forecasting accuracy.
Forecast calibration
scored forecasts: 23 · Brier: 0.248
The sample includes only resolved forecasts with a recorded probability. Interpretations and unresolved entries are excluded.
EUR/PLN's rise to a 20-day high is mostly a dollar-strength effect, not evidence corroborating the bond market's rate-cut skepticism; the 10-year POLGB spread over 6-month WIBOR remains the primary signal on NBP's easing room.
What would prove it wrong
This reading is sustained if EUR/PLN continues tracking USD/PLN and EUR/USD moves rather than diverging from them; it is retired if EUR/PLN weakens against the euro even as EUR/USD stabilises or reverses, indicating a genuinely złoty-specific repricing has taken hold.
Read analysis#How it settled
EUR/USD fell further rather than stabilising or reversing (a fresh 20-day low at 1.1334 on 29 September), so the 23 September retirement condition, which required EUR/USD to stabilise or reverse while EUR/PLN still weakened, was not met; EUR/PLN's 0.07% session move and 0.74% five-session move against USD/PLN's much larger 1.91% five-session move keep the original dollar-driver reading intact.
Leveraged funds extended their 2-year Treasury net short by a further 16,524 contracts in the week to 15 September 2026 to 1,321,644 contracts, growing into a yield at the 100th percentile of its trailing year and a Fed funds futures path pricing 70bp of additional tightening, a position the desk's 18 September falsifier has not yet broken but which the 22 September auction will test directly.
What would prove it wrong
The 2-year Treasury yield trades at or below 4.60% in the sessions following the 22 September 2026 auction, showing the leveraged short has outgrown the policy path it is betting against.
Read analysis#How it settled
no DGS2 close below 4.6 through 2026-10-04
The Nasdaq Mini futures net short covered for a second straight report (25,064 contracts in the week to 15 September) as the Nasdaq Composite rallied to a fresh 20-day high, exhibiting the ordinary short-squeeze mechanism while the position remains near its three-year stretch extreme.
What would prove it wrong
If the Nasdaq Mini net short remains at or above the 85th percentile of its three-year range in the next COT report while the Nasdaq Composite closes above 28,200, the forced-seller squeeze thesis fires; if the short keeps covering toward below the 80th percentile, the read holds.
Read analysis#How it settled
no ^IXIC close above 28200 through 2026-10-14
WTI's managed-money net long, built to the top of its three-year range in the week to 15 September 2026, is a fresh exposure absorbing a diplomacy-driven price drop the report itself predates, not evidence yet for or against the desk's 17 September asymmetry call.
What would prove it wrong
The reading is tested when the Commitments of Traders report covering the week after 15 September 2026 publishes: if it shows WTI's managed-money net long extending further into the price decline, the exposure reading is sustained; if it shows the long being trimmed, the reading should be revised toward positioning already adjusting to price.
Read analysis#How it settled
The 22 September report shows WTI's net long rising to 5,879 contracts, but gross longs fell 2,031 contracts while gross shorts fell 5,163, a mixed de-risking the 21 September note's binary extend/trim test did not anticipate.
The record 2.66pp spread between the 10-year POLGB yield and 6-month WIBOR on 20 September signals the bond market sees less room for further NBP cuts than WIG-BANKI's flat-but-still-elevated pricing assumes.
What would prove it wrong
This reading is sustained if the NBP's next projection round keeps or extends the mid-2027 return-to-target timeline from the March round; it is retired if that round instead pulls the CPI path back toward target sooner, restoring room for the cuts WIG-BANKI is pricing.
Read analysis#How it settled
The referee is correct: the stated falsifier requires the NBP's next projection round to confirm or extend the mid-2027 timeline, but the data pack still only contains the March 2026 round, so no new projection has occurred to test this; the reading cannot be graded 'sustained' on unchanged WIG-BANKI pricing and an unretested spread alone, since that is not the confirming event specified.
Leveraged funds extended their 2-year Treasury net short by 33,281 contracts in the week to 8 September 2026 even as the 2-year yield sits at the 100th percentile of its trailing year, a position now at the 97.4th percentile of its own three-year range that a soft 22 September US Note 2-Year auction or a stall in the priced Fed tightening path could unwind sharply.
What would prove it wrong
The 2-year Treasury yield trades at or below 4.60% in the days following the 22 September 2026 auction, showing the crowded leveraged-fund short is over-extended against a policy path not firming as fast as fed funds futures currently price.
Read analysis#How it settled
no DGS2 close below 4.6 through 2026-09-30
WIG20's record close on 17 September is concentrated in banks and Allegro; Orlen, one of its largest weights, fell roughly 6% from its recent peak the same session as large futures blocks traded into the 18 September expiry.
What would prove it wrong
Orlen recovers toward its recent peak in the sessions immediately following the 18 September 2026 WIG20 futures expiry, indicating the divergence was positioning-driven rather than a repricing of the refiner's fundamentals.
Read analysis#How it settled
No evidence on Orlen's price behavior after the 18 September expiry (recovery toward peak or continued decline) is presented, so the falsifier cannot be evaluated and the reading cannot be sustained as stated.
WIG-BANKI's continued five-session gain into 17 September still prices NBP easing from 3.75% that the 15 September CPI print (3.4%, rising) and a proposed fuel-tax holiday through March 2027 make harder to justify without a supportive projection revision.
What would prove it wrong
This reading is sustained if the Council's next projection round keeps the mid-2027 return-to-target date from the March round intact; it is retired if that round instead pushes the CPI path higher or later, removing the premise for further cuts.
Read analysis#How it settled
The falsifier explicitly hinges on the Council's next projection round testing the mid-2027 return-to-target date, but no such projection round has landed in this pack, leaving the claim untested rather than confirmed; meanwhile PMI fell to 48.3 and food prices are falling fastest, data not weighed in the original filing.
Brent's supply risk premium is unwinding as Saudi Arabia reroutes crude through Oman, while WTI's already-thin managed-money short leaves little further cushion for either grade if the decline extends.
What would prove it wrong
The reading is overturned if the next Commitments of Traders report, covering the week following 8 September 2026, shows Brent's managed-money net long extended rather than trimmed alongside the reroute story, or if wire reporting shows the Oman reroute reversing rather than displacing the Saudi supply flow.
Read analysis#How it settled
The 17 September note's falsifier asked for a NEW report to show WTI's or Brent's net long extending, or the Oman reroute reversing; the only report available, dated 15 September 2026, is the same one the note was built on, so neither leg has actually been tested by fresh data, and the note is revised into a forward-looking test rather than graded today.
Fed funds futures are pricing a firmer policy path (52bp higher over 12 months) even as leveraged funds keep covering a once-record Euro FX short, and the two signals cannot both be right about where dollar strength is headed.
What would prove it wrong
The Euro FX net short deepens in the next Commitments of Traders report while EUR/USD trades at or below 1.1535, reversing the covering wave and showing the priced tightening path is dominating the currency.
Read analysis#How it settled
EURUSD=X closed below 1.1535 on 2026-09-17 (close 1.14699)
An oil-driven inflation shock is colliding with a Treasury yield already at the 100th percentile of its year, while high-yield credit spreads at the 7.1st percentile still refuse to confirm the regime's stated credit-stress driver.
Read analysis#What would prove it wrong
If the 11 September CPI print shows core or headline inflation accelerating from the prior month's readings while high-yield spreads widen out of their current range, the premise that credit stress is disconnected from real-time inflation signals would be overturned.
WIG-BANKI's five-session rally into 9 September still prices NBP easing from 3.75% that the Council's unchanged, silent rhetoric on 9 September neither confirmed nor contradicted, leaving the trade dependent on a still-unpublished, expected-hawkish projection update.
What would prove it wrong
This reading is sustained if the Council's next projection round keeps the mid-2027 return-to-target date from the March round; it is retired if that round instead pushes the CPI path higher or later, removing the premise for further cuts.
Read analysis#How it settled
No fresh NBP projection round has landed since the 9 September note; the 15 September CPI print at 3.4% and the fuel-tax relief proposal narrow the case for cuts but do not themselves resolve the Council's stance, so the 9 September reading stands untested rather than confirmed or broken.