FOLIO 05 — ABOUT
QuantTrail Capital is a private algorithmic trading firm — one founder, a small bench of engineers and researchers, and a short list of things we refuse to do. This folio records who we are, what we hold to, and how the firm came to be.
Dubai, MMXXVI
To our prospective partners —
QuantTrail began with a decision about what not to trust. Not intuition — intuition proposes, it does not decide. Not conviction — conviction is abundant in this industry and priced accordingly. We chose to trust evidence: gathered slowly, tested adversarially, discarded without ceremony when it fails. Most of what we build never touches capital. We consider that the system working.
The firms that fail rarely fail for want of ideas. They fail because conviction outruns containment. So we built the containment first. Every strategy we run carries its risk budget with it — drawdown limits, exposure caps, a kill switch it cannot argue with. When a model approaches its boundary, the system reduces it. No one is consulted, because no one needs to be.
There are things this firm will not do. We do not override the machines by hand; a discretionary override is an untested strategy with a sample size of one. We do not deploy anything we cannot explain to you in plain language, on paper, in an afternoon. And we do not publish a figure we cannot trace back to fills.
We are small, founder-led, and intend to remain difficult to impress. If that temperament suits you, the memorandum is available on request.
S. H. Sagar,
Founder & Chief Investment Officer
The validation gates the letter refers to are documented on Folio 02 — Strategy.
Simulated record: Sharpe 1.42 · max drawdown −8.4%.† Full tables on Folio 04 — Performance.
No. 1
No strategy trades because someone believes in it. It trades because twelve walk-forward windows, a thousand Monte Carlo resamplings and a parameter-stability sweep failed to break it.† Belief is the cheapest input in this industry; we price it accordingly. The order of operations never reverses: evidence first, capital second.
No. 2
Each strategy receives an explicit allocation of loss it is permitted to spend — by day, by month, by position. The budget is enforced in the execution path, not reviewed in a meeting. When it is spent, the strategy stands down. No committee convenes, and no exception has been granted to date.
No. 3
Allocators receive the same ledger we read ourselves. NAV, drawdown, fees and attribution reconcile to individual fills, continuously, across every venue. A figure that cannot be traced is not published. We treat reporting as an engineering product, with its own tests, its own releases and its own standard of proof.
The first hypotheses on FX market structure are written down as falsifiable claims — before a single parameter is fit, and before the firm has a name.
The harness is built before the strategies it will judge: rolling walk-forward evaluation, Monte Carlo resampling of trade histories, parameter-stability sweeps. The gates come first.
One strategy clears every gate and is promoted to a modest live book of proprietary capital. The promotion carries its risk budget, its exposure caps and its rollback policy with it.
Execution adapters reach six venues† — MT5 and cTrader for FX; Binance, Bybit, OKX and Hyperliquid for digital assets — with exposure reconciled continuously across all of them.
Operations settle into three centres — UAE · SG · UK. Research and operations in the Emirates, digital-asset execution oversight in Singapore, FX counterparties in the United Kingdom.
The 2024 promotion set the precedent: authority, risk budget and rollback policy travel together.
United Arab Emirates
DIFC · Gate District
Dubai
Operating through a DIFC-registered entity; not a regulated fund vehicle.†
Singapore
Raffles Quay
Singapore
Digital-asset execution conducted through venue entities; no local licence is held or implied.†
United Kingdom
City of London
London EC2
FX execution routed through regulated brokerage counterparties; the firm is not itself authorised.†
Notes & methodology