All chains verified
POLARIS
Intelligence
Menu
POLARIS Intelligence · Public preview

Verifiable signal intelligence, cryptographically proven.

True North of the Market.

Every signal, paper trade, and AI rationale is appended to a SHA-256 hash chain stamped FB-LTD-17096178. A daily cron is the single authoritative verifier.

Audit clock
105 / 90
Day 105 of 90 · advances each calendar day the cron records signals on chain.
Tamper-evident by construction
1,289 records · 6 chains
Every signal, trade, fill, insight and capital event is hash-linked from genesis; any edit breaks the chain at the exact row.
Reproducible research
Sharpe 1.59 · as traded
Reported with the live risk rails applied, the layer most track records omit. The research figure is higher; this is the one that trades.
Broker-confirmed execution
Alpaca · one-directional rails
Orders route through hard risk rails; broker fills are chained as independent evidence, not self-reported figures.
§1 · Validated strategy performance

As traded, not as researched.

Production strategy: multi-signal ensemble (momentum + quality + low-volatility), sector-residualised, 12% volatility target. 756 out-of-sample trading days across six walk-forward folds, June 2023 to June 2026. The figures below are the strategy with its live risk rails applied: the per-trade stop, drawdown kill switch and position, sector and liquidity caps that govern real capital. Most published track records omit this layer. Ours does not.

1.59
Sharpe ratio (as traded)
−8.8%
Maximum drawdown
2.54
Calmar ratio
22.3%
CAGR (net of costs)
What the rails cost

Without the risk rails the same strategy scores Sharpe 1.86, maximum drawdown −9.05% and CAGR 32.2% over the identical window. That is the number a conventional backtest would publish, and it describes a system that is never stopped out. The rails cost roughly a third of the return; we report the lower figure because it is the one the platform actually trades.

Past performance does not guarantee future results. Figures are walk-forward out-of-sample backtest results, net of 0.1% transaction costs, reproducible from the locked research artefact. The backtest universe uses current index membership and a point-in-time-current sector map, which can introduce survivorship and metadata-timing effects; the 12% target is realised at ~13.3% (long-only, no leverage).

§2 · How the chain is verified

Six steps. One authoritative verifier.

Signal
generated
Canonical encode
tz-stable JSON
SHA-256
row_hash
Chain append
prev → head
Daily cron
22:30 UTC
verify_chain
integrity_runs
Each row's row_hash = SHA-256(canonical_json(row) + prev_hash). The daily cron walks the chain from genesis and recomputes every hash; the dashboard displays the cron's verdict.
Tamper-evident
Any single-character change to any row breaks the chain at the exact tampered row. The next cron run flips status to broken.
Backend-independent
Canonical encoding is timezone-stable and byte-identical across DuckDB (local) and Postgres (cloud).
Authoritative
Python is the single verifier. The dashboard reads integrity_runs and displays the verdict — browsers never re-verify.
§3 · Live chain state

1,289 rows across 6 independent chains.

Refreshed on every request. We expose row counts and verification status only — individual signals, prices, and AI rationale stay behind the demo-access wall.

capital flows
Verified
2
rows on chain
quant insights
Verified
1
rows on chain
broker fills
Verified
131
rows on chain
signals
Verified
385
rows on chain
ai rationale
Verified
385
rows on chain
paper trades
Verified
385
rows on chain
§4 · Request demo access

For institutional buyers, evaluators, and partners.

Demo access is granted on request. Complete the form below; we reply within one business day with credentials and a guided walkthrough of all platform surfaces.

We’ll reply from polaris@fruitfulbough.uk. Your details are used only to send credentials — not shared, not added to any list.

POLARIS Intelligence is a software platform operated by Fruitful Bough Ltd (company no. 17096178, registered in England & Wales). Output is impersonal commentary on a data product, identical for every subscriber. Not personal investment advice. Past performance does not guarantee future results.