Most trading systems are static. They execute the same rules regardless of market conditions. LS-AI is fundamentally different. At its core sits the Matrix — a self-optimizing grid of strategy-instrument pairs that learns, prunes, and adapts every 30 minutes without human intervention.
Every trade executed by LS-AI is logged into the Heart Memory — a persistent database that tracks which strategies work on which instruments, under which conditions, and with what outcomes.
| Data Point | Purpose |
|---|---|
| Strategy Name | Which strategy generated the signal |
| Instrument / Symbol | Which market the trade was executed on |
| Status (ON/OFF) | Whether this pair is currently active |
| Win Rate | Percentage of profitable trades |
| Profit Factor | Gross profit / Gross loss ratio |
| Confidence Score | AI-calculated reliability (0-100%) |
| Trades Count | Sample size for statistical significance |
| Broker Signature | Which broker/account type for pattern matching |
Every 30 minutes, the engine runs a complete learning cycle:
A strategy-instrument pair is automatically deactivated when:
Importantly, pruning is not permanent. A deactivated pair can be reactivated if market conditions change or if cloud patterns suggest renewed viability.
The most important number in the entire system is Pattern Density — the percentage of strategy-instrument pairs that are currently ON and active.
| Density Level | System Mode | Behavior |
|---|---|---|
| 0-40% | Exploration | Testing broadly, high trade frequency, learning what works |
| 40-60% | Transition | Pruning losers, reinforcing winners, trade frequency stabilizes |
| 60%+ | Execution | High-conviction setups only, lower frequency, higher profit per trade |
"The next inflection point isn't about profit — it's about pattern density. At current velocity, we'll cross 60% active within 48 hours. That's when the Matrix shifts from exploration to execution mode — trade frequency might decline by 20-30% while profit factor increases."
— LS-AI to MetaQuote Bot, July 2026
Starting with the latest build, LS-AI now features Cloud Heart Memory — a shared intelligence layer that connects all traders in the Armada.
When you add a new terminal with the same broker (e.g., a second Deriv demo account), LS-AI automatically copies all 18,000+ strategy-instrument status pairs from the existing terminal. No relearning required. The new terminal starts with the accumulated intelligence of its siblings.
The same mechanism works across the cloud — if another trader on Deriv has discovered that TREND_FOLLOWER works exceptionally well on EURUSD, that knowledge propagates to your Matrix.
When the Matrix crosses the 60% density threshold, the improvement is not additive — it's compounding.
| Before Pruning | After Pruning (60%+) |
|---|---|
| ~17,900 active pairs | ~12,500 high-conviction pairs |
| 30% noise pairs burning spread | 0% noise — all pruned |
| ~$15-30/day spread waste | Spread waste recovered |
| 723 instruments processed | ~500 instruments — 30% faster cycles |
| ~$50/day closed P&L | ~$80-120/day — 1.56x efficiency |
The freed margin ($500-800) is redeployed into high-profit-factor pairs, creating a force multiplier that most traders miss — they see a smaller portfolio and think less opportunity, when really it's better allocation.
Once the system is stable and has learned most strategy-to-instrument pairs, the Live Metrics tab becomes the command center. It automatically:
The Matrix never sleeps. It never stops learning. It only gets smarter.