Verdict: FX Trend chip (SMA 50/200 state) (M-029) is rated “NOT VALIDATED — fade claim unreproduced; pre-registered DXY re-test NULL” in its validation review dated n/a.
Key result: Disclosed fade claim (2020-26, 'liquid pairs'): −0.39%/20d, t = −6.9 — not reproducible from any saved script.
Key result: Pre-registered re-test on DXY (2004-26, 21 sessions): NULL: +0.15pp vs all days, 95% CI −0.15 to +0.47pp, q = 0.50.
How it was tested: training data not stated, test data 2004-01→2026-09; out-of-sample design (how it was checked on data it never saw while being built): pre-registered test.
Horizon: disclosed reading: next 20 sessions (re-test: 21 sessions). Coverage: 78 FX board rows: 26 currencies vs USD, vs SGD and vs INR.
Stage: Context-only — registered 2026-10-26 (D-006), not validated. Version v1; the hub record is as of 6 Nov 2026.
The verdict
NOT VALIDATED. Read the chip as a description of where the rate sits, never as a fade or trend-following signal: the t = −6.9 claim is unreproduced and the DXY re-test is NULL.
Development summary: A fixed SMA 50/200 trend rule that was disclosed as a fade signal (t = −6.9). No saved script reproduces that number, and the one pre-registered re-test (DXY, 2004-26) is NULL with the opposite sign.
What the review allows
As a description of where the rate sits against its 50- and 200-session averages, as the tab copy says.
What the review rules out
Fade an 'up' chip (or follow it) on the strength of the −0.39% / t = −6.9 figure — no saved script reproduces it.
Quote the DXY re-test as evidence for the board's pairs; it tests DXY only.
What it predicts
Item
Detail
Question it answers
Is each FX pair in an up, down or mixed trend — and does an 'up' chip carry the disclosed fade reading (strengthened currencies give back over the next 20 sessions), or is that claim not supported?
Method
SMA 50/200 trend state
How it works
up = rate > SMA50 > SMA200, down = the mirror, else flat, per pair on the persisted daily series; a disclosed fade reading (t = −6.9) that no saved script reproduces
Inputs
Latest cross rate, 50-session average, 200-session average
Output
Trend chip (up / flat / down) on every FX board row and the 'All trends' filter
Horizon
disclosed reading: next 20 sessions (re-test: 21 sessions)
Coverage
78 FX board rows: 26 currencies vs USD, vs SGD and vs INR
Markets
Global
Model type
Model · Rule
Stage
Context-only — registered 2026-10-26 (D-006), not validated
Version
v1, first recorded 26 Oct 2026
Validation evidence
Training data: not stated. Test data: 2004-01→2026-09. Out-of-sample design: pre-registered test.
Headline results from the model's development record, as the hub publishes them.
Measure
Result
Compared with
Note
Disclosed fade claim (2020-26 liquid pairs)
−0.39% over 20 sessions, t = −6.9
0% (no effect)
hit 43.5%, per-currency demeaned, date-clustered bootstrap — as registered on T-011 2026-09-03; not reproducible from any saved script
Pre-registered re-test: DXY 'up' → next 21 sessions
+0.15pp
all-days average (+0.08%)
95% CI −0.15 to +0.47pp, p = 0.31, q(BH) = 0.50, n = 1,718 days (115 non-overlapping), 2004-01 → 2026-09: NULL, and the wrong sign for a fade
Split halves of the re-test
+0.36pp / −0.01pp
a stable negative (fade) effect
neither half shows a fade
Chips on the board today
32 up · 28 flat · 18 down
78 board rows
fx as_of 2026-10-23
The hub's standard metrics for this model class (only those the record measured).
Metric
Sample / window
Compared with
Value
Range (CI)
Edge vs control (headline)
2004-01→2026-09
all-days DXY 21-session mean (+0.08%)
+0.15 pp
[-0.15, 0.47] pp
Cells tested
2004-01→2026-09
pre-registered DXY test
1
n/a
Cells surviving
2004-01→2026-09
1 tested (BH q = 0.50)
0
n/a
Limitations
The t = −6.9 fade claim cannot be reproduced from any saved script ( 'unreproduced').
The only reproducible re-test is on DXY, not on the board's pairs.
Pegged or managed currencies (VND, HKD) and wide EM spreads make the chip uninformative for those rows.
Questions
Is the FX Trend chip (SMA 50/200 state) validated?
Its validation review dated n/a rated it NOT VALIDATED. Read the chip as a description of where the rate sits, never as a fade or trend-following signal: the t = −6.9 claim is unreproduced and the DXY re-test is NULL.
What does the FX Trend chip (SMA 50/200 state) predict?
It answers: Is each FX pair in an up, down or mixed trend — and does an 'up' chip carry the disclosed fade reading (strengthened currencies give back over the next 20 sessions), or is that claim not supported? Output: Trend chip (up / flat / down) on every FX board row and the 'All trends' filter. Horizon: disclosed reading: next 20 sessions (re-test: 21 sessions). Coverage: 78 FX board rows: 26 currencies vs USD, vs SGD and vs INR.
How was the FX Trend chip (SMA 50/200 state) tested?
Training data not stated, test data 2004-01→2026-09, out-of-sample design: pre-registered test. Disclosed fade claim (2020-26, 'liquid pairs'): −0.39%/20d, t = −6.9 — not reproducible from any saved script; Pre-registered re-test on DXY (2004-26, 21 sessions): NULL: +0.15pp vs all days, 95% CI −0.15 to +0.47pp, q = 0.50; Status: registered 2026-10-26 (D-006); not validated.