eTIMSAn eTIMS integration
Our own KRA eTIMS (OSCU) integration, in production
- Invoices auto-submitted to KRA and returned signed with a QR code
- Deployed under each client's own KRA PIN — the software is theirs
- Wired into an existing system via API, or delivered as a ready invoicing web app
EcommerceA live store
WhatsApp checkout with M-Pesa & Paystack, on a proven platform
- Customer journey: browse, cart, checkout, order handed to the merchant on WhatsApp
- Merchant journey: products, orders and payments from one admin
- Eleven industry-ready store templates
Kwiz QuantsThe validation methodology
How a strategy earns trust before it earns capital
- Walk-forward out-of-sample validation
- Deflated Sharpe ratio
- Probability of Backtest Overfitting (PBO)
- Multi-stage MT5 execution testing
The validation methodology
Most trading strategies look profitable in a backtest and fail in live markets. The Kwiz Quants pipeline is built to tell those apart before capital is at risk.
- Walk-forward, out-of-sample validation. Every strategy is fitted on one window of history and judged only on the data it never saw, window after window, so performance is measured the way live trading actually unfolds.
- Deflated Sharpe ratio. When many strategies are tried, the best-looking one is often just lucky. The Deflated Sharpe ratio corrects the observed Sharpe for the number of trials and the non-normality of returns, so a result has to be good after that correction, not just before it.
- Probability of Backtest Overfitting (PBO). Combinatorial splits of the data estimate how likely it is that the strategy selected in-sample underperforms out-of-sample — the direct measure of overfitting.
- Multi-stage MT5 execution testing. Strategies that pass the statistics are then run through tick-level MetaTrader 5 execution testing, so slippage, spreads and order handling are part of the verdict.
- Tiered gates. Only strategies that clear every stage move on; the ones that fail are reported honestly rather than quietly dropped.