Algorithms — RO Global Capital

Algorithms · Service II

Your vision.
Our algorithms.

RO Global Capital's proprietary algorithmic framework — a decade in the making, rigorously walk-forward tested, and calibrated to your specific goals after the Wealth Clarity Discussion.

What is
Takumi?

In Japanese, Takumi means artisan — a master craftsman who combines deep knowledge with precise execution. That is the standard we hold our algorithms to.

Takumi is not a black box. It is a documented, four-layer quantitative process built on statistical edge-finding, machine learning calibration, and continuous reoptimisation — developed by Tomas Nesnidal over more than a decade of live market research.

These are not managed portfolios. Algorithms are licensed directly to clients who deploy them in their own accounts. You retain full control of your capital at all times.

Cumulative Return
+254%
vs +88% Nikkei 225
Walk-forward simulation 2010–2025
Max Drawdown
-16%
vs -32% Nikkei 225
50% less downside risk
Winning Years
88%
vs 75% Nikkei 225
More consistent positive years
Avg Annual Return
+16%
vs +12% Nikkei 225
Higher risk-adjusted profile

All figures are based on walk-forward simulation testing (2010–2025) on the Nikkei 225 universe. Simulated results are before transaction costs and performance fees. Past simulated performance is not a guarantee of future results. Please read the full disclaimer below.

How Takumi works

Every position passes through four sequential decision layers. Each layer removes a category of human error or market inefficiency. Together they form a system designed to find and exploit statistical edges that discretionary traders consistently miss.

Layer 01
Selection

Quantitative Stock Selection

A proprietary DQA algorithm (Dynamic Quality Allocation) assesses and selects which stocks have the highest potential for the upcoming time period. It objectively screens the entire market universe using multi-factor quantitative analysis — removing human bias from the selection process entirely.

Layer 02
Entry & Exit

Precision Entry & Exit

Once stocks are selected, the entry-exit algorithm monitors them daily and enters trades automatically when opportunities are spotted. It manages the position continuously — including trailing stop-loss to protect profits. The algorithm does not trade all the time. It enters at the calculated moment, rides the move, then exits when conditions change.

Layer 03
Sizing

Corrective Algorithm — Adaptive Position Sizing

Before a trade is entered, a corrective algorithm checks market conditions at the exact moment of entry. Based on last-moment probabilities, it re-adjusts the initial position size: highly favourable conditions increase risk slightly; less favourable conditions decrease it. This algorithm alone has improved simulated results by approximately 20% on average in testing.

Layer 04
Adapt

Continuous Reoptimisation

All algorithms are regularly reoptimised to adjust to ever-changing market conditions. Walk-forward optimisation keeps strategies robust and adaptive — minimising overfitting risk to historical data. The system is designed to adapt to new market regimes. Each portfolio is built from multiple trading approaches — the system determines which works best for each sector.

Institutional risk management

01
Risk Per Trade (Risk Fraction)
Defines how much capital to risk per trade — typically 0.5% to 2%. Algorithmically optimised based on risk-adjusted return objectives.
02
Initial Stop-Loss Calculation
Computes the proper initial stop-loss level for each trade based on volatility and price structure — not arbitrary percentages.
03
Position Sizing Computation
Exact number of shares is computed from risk fraction and stop distance — ensuring consistent risk across all trades regardless of stock price.
04
Constraint Checks & Validation
Position size is checked against configurable constraints — max capital usage, sector limits — and adjusted before final execution.

Walk-forward, not back-tested

Most developers use backtesting to simulate algorithms. Backtests can be highly impressive — and deeply misleading. We use walk-forward testing exclusively, a procedure much closer to actual trading, including regular portfolio recompositions and regular reoptimisation.

1
Monte Carlo Analysis
All portfolios and strategies are tested with Monte Carlo Analysis. We bootstrap all trades to assess potential max drawdown under simulated conditions (99th percentile) and configure the portfolio based on the client's requested max drawdown parameters.
2
Bootstrapping Calibration
Every algorithm is automatically evaluated by a thorough bootstrapping algorithm, which compares it against thousands of different strategies to assess whether it has a meaningful edge — and is not a result of randomness or curve-fitting.
3
Multi-Period Cross-Validation
For every algorithm, we automatically develop its sensitivity. We only deploy algorithms that keep producing similar performance with deviation from the initial settings — indicating a robust edge across multiple time windows.
4
Random Entries Calibration
We use a proprietary algorithm to calibrate trailing stop-loss methods using sophisticated sampling of random entries, then find which type and setting of trailing stop-loss produces best results among the majority of them.

Where Takumi operates

Takumi is currently deployed across three equity universes. The Nikkei 225 strategy is our flagship — with the strongest walk-forward tested results over 15 years. US and Asian equity strategies are available for clients seeking broader diversification.

🇯🇵
Japan — Nikkei 225
Flagship Strategy
Our strongest performer. 15 years of walk-forward simulation (2010–2025) with +254% cumulative return vs +88% benchmark. Best risk-adjusted results across all three universes.
🇺🇸
United States
Available
US equity strategies available for clients seeking exposure to the world's largest equity market. Full four-layer process applied to the US universe.
🌏
Asian Equities
Available
Broader Asian equity coverage for clients seeking regional diversification beyond Japan. Strategy calibrated to each market's specific characteristics.

From conversation to deployment

Step 01
Wealth Clarity Discussion
We understand your goals, risk tolerance, time horizon, and existing portfolio. The algorithm is calibrated around your specific situation.
Step 02
Custom Portfolio Design
Tomas Nesnidal builds the strategy mix to your profile. Multiple calibrations, optimisations, and stress-testing. Typically 2–3 weeks to deployment.
Step 03
Deploy & Monitor
You deploy in your own account. Continuous reoptimisation, real-time monitoring, and full performance reporting. You remain in full control.
Performance disclaimer: All performance figures shown are based on walk-forward simulation testing of the Nikkei 225 strategy over the period 2010–2025. Simulated results are before transaction costs and performance fees — actual returns will be lower. Simulated or hypothetical performance results have certain inherent limitations — unlike actual performance records, simulated results do not represent actual trading and may not reflect the impact of material market and economic factors. Past simulated performance is not indicative of future results. RO Global Capital Pte. Ltd. does not manage client funds. Algorithms are licensed for use by clients in their own accounts. All investment decisions remain entirely with the client.

"AI alone cannot crack financial markets. It needs human experience to understand the nuances and develop and calibrate algorithms properly. We use AI for implementation and fine-tuning — but every portfolio is carefully crafted and validated by humans with 20+ years of trading experience."

— Tomas Nesnidal, Senior Partner, RO Global Capital

Build your algorithmic edge

Every Takumi portfolio begins with a Wealth Clarity Discussion. We understand your goals first — then build the algorithm around them.

Request a consultation