Evidence pack

Backtest claims with source references, limits and blocked claims.

This page shows the proof layer behind the public Trade Intel results. It separates ready evidence from first-pass research, staged-only workflows and claims that are not allowed yet.

Download manifest Risk disclosure

Manifest

evidence_manifest_20260516_101702.json

Built

2026-05-16 10:17 UTC

Combined windows

5/7 positive

Average simulated return

+6.83%

Best / worst window

+18.97% / -3.74%

Operating boundary

Historical simulation; no live performance claim

7/8 rolling windows profitable

Ready

Crypto

Robustness gates passed

Ready

Futures

6/7 futures windows profitable

First pass

CFDs

Paper/staged support only

Data gated

Prediction markets

Public informed-flow + replay ready

Ready

Options

Visible but performance-disabled

Data gated

Claim Readiness

Each public statement is classified. Ready claims have source artefacts. Gated claims are deliberately shown as unsupported or unavailable so technical buyers can see where the evidence stops.

Claim

Status

Support

Disclosure

Multi-asset simulated backtest produced a positive combined return.

Ready

7 windows, 5/7 positive, +6.83% average return, -3.74% worst window.

Historical simulation only; not investment advice or live performance.

Shares sleeve has passed rolling-window robustness checks.

Ready

8 rolling windows, 7 profitable.

Historical shares simulation only.

Crypto sleeve can be benchmarked against BTC and top-10 crypto baskets.

Ready

6 benchmark windows; robustness passed.

Crypto results include high volatility and simulated execution costs.

Futures sleeve has passed first-pass benchmark checks.

Ready

7 windows, 6 profitable, +8.13% average return, 1.96 average Sharpe.

First-pass futures simulation; needs longer walk-forward coverage before headline use.

CFD support is available for customer-authored paper/staged workflows.

Paper/staged

Execution controls can represent CFD paper tickets, but live routing is disabled.

No CFD live execution, personal advice or performance claims are enabled.

Prediction-market module monitors public informed-flow style signals.

Ready

7 public informed-flow candidates; 8 finance/macro signals.

Public order flow, price, liquidity and CLV only; not private information.

Prediction-market settled replay shows positive simulated P&L.

Ready

47 replay trades, +1.125 one-contract P&L, robustness passed.

Simulated settled replay, not live trading performance.

Prediction-market module detects cross-venue divergence opportunities.

Needs evidence

0 eligible comparable pairs under strict gates.

Only equivalent resolution terms should be used.

Options sleeve is ready for historical P&L backtesting.

Not allowed yet

Current chains 4/5; historical chains 0/5.

Historical options claims require point-in-time chains, bid/ask, IV, greeks and assignment modelling.

Method Controls

The evidence is stronger when the system shows why a result might fail. These are the controls used around the current public evidence pack.

Rolling-window validation across multiple market periods.

Visible weak windows instead of a best-period-only presentation.

Transaction-cost and fill-cost assumptions included where relevant.

Benchmark comparisons against BTC, equal-weight crypto and futures baskets.

Promotion gates that reject lesson changes when validation performance does not improve.

Claim statuses that mark unsupported or gated areas rather than overstating readiness.

Important limits

These are historical simulations, benchmark checks and settled replays. They are not live customer performance, investment advice, a forecast, or a reliable indicator of future results. Some markets, including CFDs and options, are shown only as readiness, research or paper/staged workflows unless product scope and legal controls support more.

Source Files

Evidence manifest

Website-safe claim manifest with source-file references and caveats.

Shares result note

Public risk and service terms that govern how historical results may be used.

Risk disclosure

Required risk language for historical simulations and staged workflows.