PMTS 2024: Public Dashboard, MAM Launch, and the Evolution of Our AI Engine

2024 was the year PMTS stopped being a private trading system and became a public platform. Everything we promised in previous updates — transparency, verified results, investor access — is now live.

This is the longest update we have written, because more happened in 2024 than in any previous year.

The Public Dashboard — Radical Transparency

In Q1 2024, we launched the live trading dashboard at pmts.elysiumdubai.net. Every metric is pulled from real broker accounts via our DataSync Expert Advisor. Nothing is simulated. Nothing is delayed.

What the dashboard shows:

• Real-time equity curves — Balance and equity tracked every 60 seconds across all active accounts
• Monthly performance breakdown — Win rate, profit factor, Sharpe ratio, Sortino ratio, Calmar ratio, maximum drawdown
• Individual trade history — Every entry, exit, stop-loss hit, and take-profit captured with timestamps
• Profitability calendar — Month-by-month returns with color-coded heatmap
• Risk metrics panel — Live drawdown tracking, consecutive wins/losses, average holding time

The dashboard automatically syncs with MetaTrader 5 through our PMTS_DataSync.mq5 Expert Advisor. When a position opens or closes on the broker side, the data appears on the dashboard within seconds. The system handles account snapshots every 60 seconds for balance and equity tracking, and streams deals, orders, and positions in real time via POST requests to our REST API.

We believe this level of transparency is unprecedented in the algorithmic trading space. Most competitors show screenshots or PDFs. We show live data from the broker, verifiable in real time.

MAM — Multi-Account Management for Investors

The second major launch of 2024 was the MAM (Multi-Account Manager) system. This allows external investors to allocate capital to the PMTS strategy through their own regulated broker accounts.

How MAM Works

The architecture is straightforward but required months of engineering:

1. Master account — PMTS trades on a single master account at MultiBank Group (regulated, 11+ licenses)
2. Proportional allocation — When the master account executes a trade, the MAM system distributes the P&L proportionally to all connected investor accounts based on their allocation percentage
3. Independent accounts — Each investor maintains their own broker account. Their capital never leaves their account. We never touch, manage, or have access to investor funds
4. Fee structure — 35% performance fee on monthly profits only. No management fee. No upfront cost. If the system does not make money, we do not get paid

The MAM Distributor (MAMDistributor.php) runs as a server-side process that calculates each user’s share of every trade. It implements a High Water Mark (HWM) system — if a user has a losing month, the losses must be recovered before any performance fees are charged. This ensures we are fully aligned with investor interests.

User Panel

Alongside MAM, we built a complete user panel where investors can:

• View their personal portfolio performance and equity curves
• Track deposits, withdrawals, and fee history
• Configure notification preferences (trade alerts, weekly reports)
• Manage security settings (2FA via Google Authenticator)
• Select their preferred display currency (EUR, USD, GBP)

AI Engine Evolution — From GPT-4 to GPT-4o

In May 2024, OpenAI released GPT-4o (“omni”), a faster and more capable model. We migrated our AI Master Analyzer immediately. The improvements were significant:

• Response latency — GPT-4o processes our analysis prompts 2-3x faster than GPT-4, reducing the total analysis generation time from ~90 seconds to ~35 seconds
• Reasoning quality — More nuanced understanding of contradictory signals. When technical indicators say BUY but sentiment is bearish, GPT-4o produces better-calibrated probability assessments
• Multilingual output — Native support for Spanish analysis alongside English, enabling our i18n system to serve AI-generated content in both languages without translation artifacts

We also implemented a smart refresh system that triggers analysis regeneration based on three conditions:

1. Economic calendar events — Automatic refresh 30 minutes before and 15 minutes after high-impact events (FOMC, NFP, CPI)
2. Price volatility — If gold moves more than 0.5% since the last analysis, a new analysis is generated immediately
3. Scheduled refresh — Full analysis regeneration every 2 hours during market hours

Reversal Detection 2.0

One of the most requested improvements from our quant team was a better reversal detection system. The original version used simple RSI divergence and MACD crossover signals. Version 2.0 implements a 6-factor confluence analysis:

Each factor is evaluated independently and the confluence score determines the reversal probability tier: LOW (25-49%), MODERATE (50-74%), or HIGH (75-100%). The system only acts on reversals with HIGH probability AND confirmation from the AI Master.

7-Language Internationalization

We expanded from 2 languages (English, Spanish) to 7 languages: English, Spanish, Arabic, German, Portuguese, Hindi, and Chinese. This was not just a UI translation — every public page is server-side rendered in the requested language for SEO, with proper hreflang tags so Google indexes each language variant independently.

The i18n system uses JSON translation files with over 1,500 keys per language, covering every UI element, dashboard metric label, and legal disclaimer. Arabic support includes full RTL (right-to-left) layout rendering.

Security & Infrastructure Hardening

With real investor capital on the platform, security became paramount:

• Two-factor authentication — TOTP-based 2FA via Google Authenticator with backup codes
• JWT session management — Token-based auth with automatic expiry and refresh
• Rate limiting — API-level rate limiting on all authentication endpoints
• HSTS and CSP headers — Strict transport security and content security policies
• Encrypted credentials — All API keys, database credentials, and Stripe keys stored in environment variables, never in code
• Audit logging — Every administrative action logged with timestamp, user, and details

Multi-Tenant White-Label System

Looking toward B2B expansion, we built a multi-tenant white-label system. Partners can deploy their own branded version of PMTS with custom logos, colors, domain, and language settings — all powered by the same backend infrastructure and trading engine.

The Tenant system is managed through the admin panel and supports per-tenant configuration of: branding, enabled pages, language subsets, and feature flags.

Payment Processing

We integrated Stripe for automated fee collection:

• Monthly performance fees calculated automatically from MAM distribution data
• High Water Mark ensures no fees on recovery months
• Support for balance deduction (from trading account) or card payment
• Automated billing on the 1st of each month
• Invoice generation with UAE-compliant formatting

Looking Back, Looking Forward

From a research project in 2015 to a live, regulated, AI-powered trading platform in 2024 — the journey has been longer and harder than we imagined. But the core thesis has held: a system that reads institutional order flow, validated by AI synthesis, with institutional-grade risk management, can produce consistent results in the gold market.

2025 will focus on scaling — more accounts, more instruments, more languages, and most importantly, more verified performance data. We will let the results speak.

Past performance does not guarantee future results. Trading involves substantial risk of loss. Only invest capital you can afford to lose.

— Lorenzo Ballanti, Founder & Lead Developer, Elysium Media FZCO, Dubai Silicon Oasis