Quantitative research & trading systems
Tickflow Capital is a Nairobi-based quantitative research and trading technology firm. We build systematic trading systems for XAUUSD and publish the full walk-forward record - including the Windows that failed.
Validation record
The strategy is evaluated with rolling walk-forward analysis on XAUUSD. Each iteration trains on 24 months of prior data and validates on the following 12 months, stepped forward one year at a time and repeated four times across 2020–2023. A window passes if profit factor is at or above 1.0 and trade count is at least 10. Signals use only data available at the time of decision.
Full methodology, gate definitions and reproducibility requirements are published on the validation methodology page.
| Window | Validation period | Trades | Profit factor | Result |
|---|---|---|---|---|
| 1 | 2020 | 18 | 3.424 | PASS |
| 2 | 2021 | 6 | 0.796 | FAIL |
| 3 | 2022 | 10 | 1.664 | PASS |
| 4 | 2023 | 21 | 1.731 | PASS |
Window 2 (2021) is retained in the record as a failure. It contained six trades, five of them in January before the regime filter's moving averages crossed. The failure is descriptive, not explanatory - we do not claim to know why the window lost; we publish it and let the record stand. The strategy is long-only on XAUUSD in the current phase; other configurations are gated behind their own validation.
Profit factor is gross profit divided by gross loss. Three of four windows clear the 1.0 breakeven threshold. Window 2 (0.796) does not, and is retained in the record.
Out-of-sample holdout (simulated)
This period was not used in training or parameter selection. It is a blind holdout backtest, not a live trading record. Returns are gross of commission and swap.
Methodology notes. The Monte Carlo figure reports the fraction of 10,000 simulations with positive terminal equity, where simulations permute trade order with replacement. Trade-order permutation does not test for edge; it tests sensitivity to sequence. Any positive-edge claim rests on the walk-forward record, not on the Monte Carlo result. Buy-and-hold XAUUSD over the same holdout period is a required comparison and is being computed. Trial count for parameter selection across the current strategy family will be disclosed alongside the benchmark.
Systems
Two production systems sit behind the research: an internal trading and execution stack, and an open-source machine-learning platform released under the MIT licence.
Trading Intelligence & Execution System. An event-driven backtesting and validation engine, a sequential pre-trade risk framework, an execution bridge with preflight validation and fill reconciliation, and an append-only audit log with trace-ID correlation. Built to run our own capital.
An AI stock intelligence platform: LSTM neural networks for directional price prediction, FinBERT for financial-news sentiment analysis, and a hybrid model combining both with technical indicators. Full REST API with authentication, rate limiting and Swagger documentation.
Products
A point-in-time research environment for macro state, regime and XAU/XAG analysis, released as independently validated modules rather than one opaque signal. Each module opens only when it clears its own validation gates.
macro_regime_gmm · 0.1.0 · Research
Classifies the current macro regime from a point-in-time composite of four canonical dimensions - growth, inflation, liquidity and risk - and will publish posterior probabilities across a set of empirically derived macro states once the model is fit.
The module does not output a position, size or buy/sell signal. It publishes regime state and probability. Pricing is not yet set; joining the waitlist involves no payment or commitment. Full status
Join the regime waitlistOpen · No signup
The validation record behind the platform, published in full before any module is offered for access.
About
Tickflow Capital Limited is a quantitative research and trading technology firm incorporated in Kenya on 13 January 2026. We trade our own capital with the tools we build. We do not manage client money or provide investment advice. Selected tools and research are offered to systematic traders, proprietary firms and research teams.
Founder & Head of Research
Quantitative researcher and software developer. Builds the research stack, the validation engine and the execution infrastructure end to end, and operates the internal trading account.
Get in touch
Regime waitlist, research partnerships, licensing and consulting enquiries go directly to the founder. Expect a reply within one to two business days.
Include your organisation type, use case and the module or service you are asking about, so the first reply can be substantive.