Home Who We Are What We Do Join Tomtit Capital Contact Privacy Policy
Our Business

Systematic trading
from first principles.

We build and operate quantitative trading strategies across equities, derivatives, and futures — powered by proprietary machine learning and disciplined risk management.

Financial charts and quantitative data
3+
Asset classes traded
ML
Signal generation
24/7
Automated monitoring
0
Discretionary overrides
What we trade

Tomtit Capital operates systematic quantitative strategies across global equities, equity derivatives, index futures, and fixed income instruments. Our strategies are sector-agnostic and regime-adaptive — designed to perform across market environments rather than optimised for a single regime.

Every position taken by Tomtit Capital is the output of a model. There are no discretionary overrides. There is no "gut feel." The system is the strategy, and the strategy has survived rigorous out-of-sample testing before it touches capital.

If you cannot write down the rule, you cannot trade it. If you cannot explain why it works, you should not trust that it will continue to.
Our strategies
01
Statistical Arbitrage
We identify and exploit temporary mispricings between related instruments using co-integration models, mean-reversion signals, and pairs-trading frameworks validated across multiple market regimes.
02
ML-Driven Momentum
Transformer-based price prediction models trained on OHLCV data, options flow, and sentiment signals. Our models predict return distributions, not point estimates — and they know when not to trade.
03
Sentiment Alpha
Fine-tuned language models trained on earnings call transcripts, SEC filings, and financial news. We score sentiment at the document level and combine it with price signals for conviction-weighted positioning.
04
Regime-Adaptive Allocation
Hidden Markov Models over volatility surfaces, yield curves, and market breadth indicators identify the current market regime. Different strategy mixes are deployed depending on the detected state.
05
Options Flow Signals
We monitor unusual options activity — volume anomalies, implied volatility skew, and order flow imbalance — as a leading indicator of institutional positioning before price moves.
06
Risk Execution Engine
Proprietary execution algorithms minimise market impact and slippage. Reinforcement-learning-based order routing continuously optimises against live market microstructure conditions.
Data analysis and machine learning infrastructure
Our infrastructure

We build all of our own infrastructure. Our data pipeline ingests tick-level price data, options chains, earnings call audio, SEC filings, and macroeconomic releases — processed, cleaned, and stored in a time-series database built for financial research.

Our backtesting framework implements purged walk-forward validation and combinatorial cross-validation — the methods developed by Marcos López de Prado — to ensure that what we measure in research reflects what we'll see in production. Lookahead bias is not tolerated. Survivorship bias is controlled for from the outset.

All strategies are paper-traded for a minimum of 28 days before they see live capital. When they go live, they start at 0.5% of capital and scale only if the out-of-sample Sharpe exceeds 1.5.