Kwiz Quants
Institutional-grade systematic trading strategies, rigorously validated through a multi-layer pipeline and automatically executed on your MT5 account. You keep full control of your capital. We provide the intelligence.
Live Dashboard Coming Soon
The Kwiz Quants monitoring dashboard will be embedded here once deployed.
How It Works
From strategy selection to automated execution — fully transparent, fully under your control.
Select from our validated strategy library across multiple currency pairs and timeframes — or bring your own strategy for our validation pipeline.
Securely connect your MetaTrader 5 brokerage account through our kwizmt5 R package — a dual-protocol bridge (TCP and HTTP) that talks to your MT5 terminal. Your funds stay in your account — always.
Strategies execute automatically with full transparency. Monitor performance in real time through your dedicated dashboard.
Strategy Library
Each strategy family targets distinct market inefficiencies. No proprietary signal logic is revealed — we show the approach, you see the results.
Cross-pair lead-lag 28 forex pairs
Exploiting short-term predictive relationships between correlated currency pairs. Signals derived from lagged cross-pair price dynamics across multiple timeframes.
Time-varying hedge ratio Majors & crosses
Dynamic state-space pairs trading. A Kalman filter estimates a time-varying hedge ratio between two log-prices; signals trigger on standardised innovations of the resulting spread.
Rolling correlation Euro, yen & commodity crosses
Mean-reversion strategies built on time-varying correlation between structurally related pairs. Entry/exit triggered by correlation breakdowns and subsequent restoration.
Cointegration / spread mean-reversion Majors & crosses
Statistical arbitrage on cointegrated forex pairs. Spread modelled as a mean-reverting process; positions initiated when the spread deviates beyond statistically significant thresholds.
Trend, mean-reversion, momentum, S/R, breakout 105 instruments · 9 asset classes
Per-instrument strategy families spanning trend, channel breakout, support/resistance, candlestick & chart patterns, volatility, volume, momentum, mean reversion, regime, and composite overlays — each scanned across 105 instruments and multiple timeframes.
Our Differentiator
Our strategies are held to institutional-grade validation standards before they ever touch real capital. This disciplined process is what separates Kwiz Quants from signal-selling retail operators.
Each of the 52,248 strategy-pair-timeframe combinations is run through 10-fold walk-forward cross-validation with parameter grid search (70% train / 30% test per fold) on MT5 broker historical OHLC data — pulled via the kwizmt5 bridge and persisted as per-symbol-timeframe parquet caches. Multiple pipeline runners claim work atomically from the shared queue (Neon Postgres in cloud mode, SQLite WAL locally) so the search parallelises across machines. Fold results are aggregated and fed into a 6-dimension Composite Robustness Score (CRS): Deflated Sharpe Ratio, Probability of Backtest Overfitting, trade sufficiency, fold hit-rate, return CV, and exit resilience — reported as a robustness diagnostic, separate from the promotion gate.
Strategies that survive Stage 1 are re-run inside the native MT5 Strategy Tester using broker tick data — real spreads, real execution costs. Our KwizStrategyTester EA runs the same strategy logic in tester mode as it does in live trading, eliminating implementation drift. MT5 launches natively on Windows or via a Wine64 bridge on macOS. This stage catches gaps that OHLC backtests miss.
Only strategies that survive both stages and post a positive MT5 Strategy Tester monthly return above the configured threshold (default 1% / month) are promoted to the live pool. The Composite Robustness Score is reported alongside but no longer gates promotion — robustness diagnostics and the production-fitness decision are intentionally decoupled. Strategies that fail the monthly-return bar but still score well on robustness drop to the monitor tier; the rest are quarantined. This design ensures every promoted strategy has produced realised, broker-side returns under tick-level execution.
# Kwiz Quants — 2-Stage Validation Pipeline
# 52,248 strategy × pair × timeframe combinations
# Stage 1: walk-forward grid search (broker OHLC)
cv <- run_unified_cv(
strategy, n_folds = 10,
train_pct = 0.70, grid_search = TRUE
)
crs <- score_robustness(cv,
dims = c(“dsr”, “pbo”, “trade_sufficiency”,
“fold_hit_rate”, “return_cv”, “exit_resilience”)
) # reported, not a gate
# Stage 2: MT5 tester — KwizStrategyTester (broker tick data)
mt5 <- run_mt5_tester(strategy, ea = “KwizStrategyTester”)
# Promotion gate: realised MT5 monthly return
if (mt5$monthly_return_pct >= 1.0) {
promote_to_live(strategy)
}
Infrastructure
Kwiz Quants runs on production-grade infrastructure designed for reliability, transparency, and performance.
Signal generation, 2-stage validation pipeline, and portfolio management — all in R. Dashboard, Payments API, and Strategy API are containerised in Docker with renv-locked packages.
Our own MT5 EA. A single EA that runs identically in MT5’s Strategy Tester and in live trading — no mode switching required. Handles 105 instruments across 9 asset classes with split take-profit and position recovery.
13-module Rhino/Shiny app: strategy selection, live trading orchestration, real-time P&L, backtesting UI, and client management — backed by Plumber APIs.
Cloud-native data layer: Neon Postgres holds the work queue and strategy index; backtest results land as columnar Parquet in Google Cloud Storage. The same cloud_store abstraction falls back to a local SQLite WAL queue and on-disk Parquet for offline runs, so research and production share one storage contract.
Pricing
All plans include the multi-layer validation pipeline and real-time monitoring. Currently in early access — contact us for pricing details.
From the Quants Blog
Research and engineering notes from the Kwiz Quants team — strategy validation, position sizing, R infrastructure, and the realities of systematic trading from East Africa.
Request early access to Kwiz Quants or schedule a demo to see the platform in action.