F01Public signal accountabilityWalk-forward · SPX and ES · as of Jul 15, 2026

Falsification Ledger.

Define the claim before reading the outcome. Keep only independent observations. Publish the failure. Withhold the statistic when the evidence floor is not met.

Publication rule

Failures carry the same weight as successes.

Performance is visible only when at least n=50 independent walk-forward events survive a 21-session separation. A refusal is a result. A descriptive indicator is not converted into a trade after the fact.

Definitions audited
7
Performance published
1
Performance withheld
5
Not a signal
1
Publication floor
n≥50
Publishable result · 1 of 7

The COT claim did not survive cleanly.

The upper and lower ES commercial-positioning deciles produced n=55 independent events. Across both benchmarks, the average direction-adjusted return is negative at every published horizon. Raw market drift is shown separately so it cannot be mistaken for signal edge.

Exact definition

ES commercial COT extremes

Entry when the trailing 52-week ES commercial-net percentile crosses into the upper decile at 90 or the lower decile at 10. Repeated carried-forward daily values do not create new events.

Directional claim
Upper-decile entry bullish, lower-decile entry bearish.
Candidate entries
n=85
Independent entries
n=55
SPX

S&P 500 cash index

FAILED DIRECTIONAL TEST · n=55
HorizonHit rate95% intervalAverage rawAverage alignedIndependent n
+1 session38.2%26.5% to 51.4%-0.01%-0.09%n=55
+5 sessions45.5%33.0% to 58.5%+0.02%-0.23%n=55
+21 sessions50.9%38.1% to 63.6%+1.11%-0.20%n=55

Maximum drawdown of the non-overlapping 21-session signal sequence: -17.72% · n=55

Where the signal fails

A regime is marked failed when its n is at least 10 and either its 21-session hit rate is below 50% or its average direction-adjusted return is not positive.

RegimeVerdictHit rateAverage alignedIndependent n
risk-onFAILED40.5%-0.83%n=37
neutralHELD68.8%+0.82%n=16
risk-offINSUFFICIENT100.0%+3.46%n=2
ES

E-mini S&P 500 continuous futures

FAILED DIRECTIONAL TEST · n=55
HorizonHit rate95% intervalAverage rawAverage alignedIndependent n
+1 session34.5%23.4% to 47.7%+0.00%-0.08%n=55
+5 sessions45.5%33.0% to 58.5%+0.02%-0.18%n=55
+21 sessions49.1%36.4% to 61.9%+1.08%-0.06%n=55

Maximum drawdown of the non-overlapping 21-session signal sequence: -16.34% · n=55

Where the signal fails

A regime is marked failed when its n is at least 10 and either its 21-session hit rate is below 50% or its average direction-adjusted return is not positive.

RegimeVerdictHit rateAverage alignedIndependent n
risk-onFAILED37.8%-0.74%n=37
neutralHELD68.8%+1.09%n=16
risk-offINSUFFICIENT100.0%+3.29%n=2
Audit all 55 independent events
DateDirectionRegimeFoldSPX +21ES +21
Mar 13, 2012Lower-decile entryrisk-on1-0.60% raw-0.73% raw
Jun 05, 2012Lower-decile entryrisk-on1+6.39% raw+5.95% raw
Jul 17, 2012Lower-decile entryrisk-on1+3.07% raw+3.31% raw
Aug 28, 2012Lower-decile entryrisk-on1+2.69% raw+2.36% raw
Oct 16, 2012Lower-decile entryrisk-on1-6.53% raw-6.18% raw
Dec 04, 2012Lower-decile entryrisk-on1+4.22% raw+3.72% raw
Mar 05, 2013Lower-decile entryrisk-on2+1.31% raw+1.14% raw
Apr 16, 2013Lower-decile entryrisk-on2+5.35% raw+5.45% raw
Sep 17, 2013Lower-decile entryrisk-on2+0.98% raw+0.48% raw
Oct 22, 2013Upper-decile entryrisk-on2+1.52% raw+1.73% raw
Jan 21, 2014Upper-decile entryrisk-on3-0.22% raw-0.12% raw
Mar 04, 2014Lower-decile entryrisk-on3+0.91% raw+0.61% raw
May 20, 2014Upper-decile entryrisk-on3+4.63% raw+4.84% raw
Jun 24, 2014Upper-decile entryrisk-on3+1.95% raw+1.93% raw
Sep 16, 2014Lower-decile entryrisk-on3-6.83% raw-7.64% raw
Feb 10, 2015Lower-decile entryrisk-on4-0.13% raw+0.08% raw
Jun 23, 2015Upper-decile entryrisk-on4-1.04% raw-0.85% raw
Aug 11, 2015Upper-decile entryrisk-on4-6.32% raw-6.26% raw
Sep 22, 2015Upper-decile entryrisk-on4+3.92% raw+3.96% raw
Mar 22, 2016Upper-decile entryrisk-on5+2.03% raw+1.97% raw
Jun 14, 2016Lower-decile entryrisk-on5+4.26% raw+3.99% raw
Jul 19, 2016Lower-decile entryrisk-on5+0.85% raw+0.97% raw
Mar 07, 2017Lower-decile entryrisk-on6-0.65% raw-0.85% raw
Jun 20, 2017Upper-decile entryrisk-on6+1.49% raw+1.38% raw
Aug 08, 2017Upper-decile entryrisk-on6-0.40% raw-0.25% raw
Sep 12, 2017Lower-decile entryrisk-on6+2.35% raw+2.28% raw
Oct 31, 2017Lower-decile entryrisk-on6+2.81% raw+2.92% raw
Dec 05, 2017Lower-decile entryrisk-on6+4.32% raw+4.35% raw
Jan 30, 2018Lower-decile entryrisk-on7-5.13% raw-5.18% raw
Mar 06, 2018Lower-decile entryrisk-on7-2.39% raw-2.29% raw
Jan 15, 2019Upper-decile entryneutral8+5.19% raw+5.30% raw
Apr 02, 2019Upper-decile entryneutral8+1.75% raw+1.76% raw
Dec 17, 2019Lower-decile entryneutral8+4.29% raw+4.17% raw
Feb 04, 2020Lower-decile entryneutral9-8.30% raw-8.61% raw
Jun 23, 2020Upper-decile entryneutral9+3.33% raw+3.50% raw
Oct 27, 2020Lower-decile entryneutral9+7.05% raw+7.22% raw
Apr 27, 2021Lower-decile entryrisk-on10+0.22% raw+0.34% raw
Aug 03, 2021Lower-decile entryneutral10+2.28% raw+2.41% raw
Oct 12, 2021Lower-decile entryrisk-on10+6.80% raw+6.94% raw
Jan 18, 2022Lower-decile entryrisk-on11-2.23% raw-2.21% raw
Mar 22, 2022Upper-decile entryneutral11-2.61% raw-2.54% raw
May 03, 2022Upper-decile entryneutral11+0.03% raw+0.14% raw
Jun 21, 2022Upper-decile entryrisk-off11+6.22% raw+6.20% raw
Apr 11, 2023Upper-decile entryrisk-off12+0.70% raw+0.37% raw
Aug 08, 2023Lower-decile entryneutral12-1.07% raw-1.38% raw
Sep 26, 2023Lower-decile entryneutral12-2.03% raw-2.43% raw
Jan 09, 2024Lower-decile entryrisk-on12+5.08% raw+4.69% raw
Apr 02, 2024Lower-decile entryneutral13-3.60% raw-4.07% raw
Jul 02, 2024Lower-decile entryneutral13-1.13% raw-1.59% raw
Oct 15, 2024Lower-decile entryrisk-on13+2.93% raw+2.61% raw
Mar 11, 2025Lower-decile entryrisk-on14-2.07% raw-1.54% raw
Jun 10, 2025Upper-decile entryneutral14+3.66% raw+4.62% raw
Sep 16, 2025Upper-decile entryneutral14+0.97% raw+1.58% raw
Mar 31, 2026Lower-decile entryrisk-on15+10.42% raw+10.24% raw
Jun 02, 2026Upper-decile entryneutral15-1.66% raw-1.25% raw
No performance number published

Six claims refused or withheld.

These cards carry the same visual weight as the publishable result. They state exactly what is missing: a testable claim, the required point-in-time intraday data, or enough independent events.

Published indicatorNOT A SIGNAL

Futures Key Levels

Plots prior-session, opening-range, initial-balance, VWAP, pivot, ADR, settlement, weekly, and monthly references. The published indicator defines no entry, exit, long, or short event.

Claimed direction
None. It is a reference grid, not a directional signal.
Candidate events
n=0
Independent events
n=0
Publication floor
n≥50

No signal exists to test. Assigning an outcome would invent a claim the indicator explicitly does not make.

Source: Published Pine v6 indicator definition

Published indicatorWITHHELD

Liquidity Levels Pro sweep reclaim

A high-zone sweep requires the bar high to exceed the zone top and close back below it. A low-zone sweep requires the bar low to breach the zone bottom and close back above it, after the Pine zone-state and separation rules are applied.

Claimed direction
High-zone reclaim bearish, low-zone reclaim bullish.
Candidate events
n=0
Independent events
n=0
Publication floor
n≥50

The repository has no point-in-time five-minute archive containing the evolving pivot-zone state. A daily approximation would not reproduce the published signal.

Source: Published Pine v6 indicator definition

Published indicatorWITHHELD

Institutional Orderflow Pro pulse

An edge-triggered long or short pulse after relative volume, delta z-score, close-in-range, cumulative delta, session VWAP direction and slope, prior-day proximity, and divergence gates agree.

Claimed direction
Bull pulse bullish, bear pulse bearish.
Candidate events
n=0
Independent events
n=0
Publication floor
n≥50

The repository has no point-in-time intraday OHLCV archive needed to reconstruct session VWAP, cumulative delta, and the bar-level gates. Daily candles are not a valid substitute.

Source: Published Pine v6 indicator definition

Site signalWITHHELD

Macro regime flips

Risk-on entry when the five-leg composite crosses from below 60 to 60 or higher. Risk-off entry when it crosses from above 40 to 40 or lower.

Claimed direction
Risk-on entry bullish, risk-off entry bearish.
Candidate events
n=79
Independent events
n=41
Publication floor
n≥50

Only n=41 independent walk-forward observations remain after the 21-session separation rule. Publication requires n>=50.

Source: dhawal.org five-leg macro regime composite

Site signalWITHHELD

Net GEX sign changes

Positive entry when net gamma exposure crosses from non-positive to positive. Negative entry when it crosses from non-negative to negative.

Claimed direction
Positive-GEX entry bullish, negative-GEX entry bearish.
Candidate events
n=0
Independent events
n=0
Publication floor
n≥50

Only n=0 independent walk-forward observations remain after the 21-session separation rule. Publication requires n>=50.

Source: Delayed Cboe options-chain snapshot through dhawal.org options analytics

Site signalWITHHELD

2s10s curve inversion crosses

Inversion entry when the 10-year minus 2-year Treasury spread crosses from non-negative to negative. Un-inversion entry when it crosses from non-positive to positive.

Claimed direction
Inversion entry bearish, un-inversion entry bullish.
Candidate events
n=10
Independent events
n=4
Publication floor
n≥50

Only n=4 independent walk-forward observations remain after the 21-session separation rule. Publication requires n>=50.

Source: Federal Reserve FRED T10Y2Y daily series

Walk-forward protocol

Frozen rules, forward-only evaluation.

The first 1,260 sessions form the initial training history. Each following 252-session block is an out-of-sample test fold. Signal thresholds come from the published production definitions and are never optimized against a test fold.

Only one event is retained in any 21-session window, matching the longest forward horizon. Hit rate counts direction-adjusted returns above zero. The Wilson interval is two-sided at 95%. Average raw return shows market drift; average aligned return multiplies bearish events by -1.

Maximum drawdown compounds the non-overlapping 21-session direction-adjusted event sequence in date order. It is a signal-sequence diagnostic, not a live strategy equity curve, and includes no transaction-cost assumption.