Risk-scoring infrastructure · FinTech

We predict which accounts breach their drawdown limits — before they breach.

Trade System is a risk-scoring layer for prop firms and brokers. We score every trading account from its raw trade stream and flag hidden risk days before a limit breach. We do not manage client funds.

A single account, 40 trading days before a breach Balance (what leaderboards show) Equity (actual exposure)
↑ divergence flagged here — days before the balance curve turns

The problem

The industry ranks traders by returns. Returns hide risk.

The account that looks best on a leaderboard is usually the one absorbing the most unrealised risk — right up until the session it loses everything.

Prop firms

Their entire economics rest on one prediction: which funded accounts will breach drawdown limits. Most have no systematic way to make it, so breaches are booked as a fixed cost of doing business.

Brokers

Copy-trading clients follow the top of the leaderboard, blow up, and churn. The acquisition cost is paid; the lifetime value never arrives.

Platforms

Leaderboards rank by realised P&L — the exact metric a martingale or grid strategy optimises for, right until it fails.

How it works

A risk-scoring API, not a fund.

Institutions send us trade streams. We return a per-account risk score. We hold no funds and execute no trades.

1

Ingest

Raw trade-level streams from MT4 / MT5 and copy-trading venues. No custody, no execution.

2

Extract

Behavioural features: balance–equity divergence, sizing escalation, exposure concentration, holding-time tails.

3

Score

Per-account probability of a drawdown-limit breach over a forward window, refreshed continuously.

4

Act

Delivered via API and alerting. The firm decides: reduce allocation, cap exposure, or intervene.

Our core signal is the divergence between balance drawdown and equity drawdown — the structural signature of martingale and grid exposure. It is independent of profitability, which is precisely what a returns-based leaderboard cannot see.

Why us

Everyone ranks. Nobody scores risk.

The incumbent layer sorts by outcome. We model the process that produced it.

 
Ranks by return
Models hidden risk
Sold to institutions
MQL5 / Signal Start
Yes
No
No
ZuluTrade / Myfxbook
Yes
No
No
Prop firm dashboards
Yes
No
In-house
Trade System
No
Core product
Yes

Team

We are the people whose money was on the line.

Not outsiders building a fintech tool from a market map. Five years operating trading desks, and a quantitative research practice that kills its own strategies when they fail out-of-sample.

Daryn Doldashev
CEO

Commercial and institutional sales. Built and ran a trading operation with 1,000+ trained traders. Owns the ADGM regulatory workstream.

Yesset Zhussupov
CTO

Data infrastructure and MLOps. Broker-side ingestion, model retraining, API delivery. Built the production pipeline that runs the scoring engine.

Adilkhan Zhubanaliyev
Head of Quantitative Research

Owns the scoring engine: feature engineering, model architecture and validation — walk-forward, purged cross-validation, deflated Sharpe, permutation testing.

We do not manage client capital. Trade System holds no funds, takes no discretionary mandate and executes no trades. We sell software to regulated institutions — a scope that sits cleanly outside fund-management licensing.

The leaderboard shows you who won.We show you who is about to lose.

info@trade-system.net