How to Design a Rules-Based Automated Investment Strategy

This is not financial advice – it reflects my own approach and research.

I’m sure I’m not the only one who has wondered how to take advantage of AI to improve their personal finances ;). For me, this took the form of a “vibe session” describing my situation in the Netherlands with the aim of building a low maintenance rules-based automated investment strategy. Low maintenance and automated because I’d like to spend my time on other things. Rules-based because ad-hoc investing can be a quiet tax on mental bandwidth and portfolio returns.

Without a rigid framework, it’s easy to buy on sentiment, pause during volatility, and reverse decisions. As a developer who builds automation systems, applying systems-thinking to personal finance felt like an obvious move – and the result is the structure I’m documenting here. This is also the foundation for the 88capital content brand, where I document the rules-based automated investment strategy as a running experiment. It is an automation template, not an execution engine or financial advice – it reflects my own approach, research and implementation, which you can adapt to your own situation.

The Rules-Based Automated Investment Strategy

A common recommendation for personal financial budgeting is the tripartite 50/30/20 split. This investment strategy applies a variant of that rule to (a subset of) the “saving and investing pot”. 50% of that budget goes into passive index anchors, 30% into thematic allocations, and 20% is reserved for active trading when triggered. All layers carry different risk and reward profiles.

The 50% and 30% layers are passive buy and hold layers, to benefit from average market returns. The 20% active portion is not designed to beat the market or generate outsized returns through speculative trading. It is a structured mechanism, running on an n8n platform to identify and provide a controlled outlet for opportunistic short or long term capital deployment when conditions are right, e.g. during a market drawdown, without relying on gut feel.

Don’t be a collector of funds; be a builder of positions.

The passive 50% and 30%

The passive portion splits across two distinct environments to balance tax efficiency with thematic exposure: 50% in Northern Trust (NT) funds, 30% in a basket of thematic ETFs.

Northern Trust funds and Dutch tax efficiency

For the 50% core index allocation, the strategy uses Northern Trust index funds. These are Dutch-domiciled FGR (Fonds voor Gemene Rekening) structures, which are worth exploring if you are tax-resident in the Netherlands. Typically, when U.S. companies pay dividends to international funds, 15% is withheld at source, which in some cases might not be reclaimable by retail investors. Dutch-domiciled FGR funds however, can be structured to avoid that dividend withholding leakage, a small annual amount that compounds over the long term.

Thematic ETFs

The remaining 30% goes into a small basket of thematic ETFs. The brokerage choice of Interactive Brokers (IBKR) was based on the availability of the features I needed, such as the ability to schedule recurring trades, broad instrument access, cost transparency and the potential for further automated integration in the subsequent stages of this experiment. The purpose of this thematic layer is to target trends that are under-represented in standard market-cap-weighted global indexes, for example Quantum computing and technology.

Low cost and structural tax efficiency are not optional optimisations here. They are some of the levers that passive investors can control, and protecting compounding returns over long horizons starts with not leaking capital when this can be avoided.

The active 20%

The remaining 20% is reserved for signal-driven dip-buying. Not as day trading or an attempt to call market bottoms. It is a structured way to deploy cash reserves into established large-cap companies during temporary market distress.

The multi-tier signal criteria

Capital is only deployed when a specific set of technical conditions is met. The evaluation follows a strict filtering hierarchy:

• Tier 0 gate: The asset’s current price must be at least 15% below its 52-week high. If this condition is not met, the asset is immediately discarded from evaluation. This prevents wasting analysis on minor market noise.
• Secondary evaluation: Once the Tier 0 gate is cleared, the system assesses the 14-day Relative Strength Index (RSI) and the 200-day Simple Moving Average (SMA). The RSI must be below 35, indicating oversold conditions, and the price must sit below the 200-day SMA.

These parameters are designed to isolate periods of genuine capitulation rather than standard daily fluctuations. They do not eliminate false positives, but they do enforce a barrier to entry. This effect of following this strategy is being tested using a paper account at IBKR.

The automation and execution split

I use n8n to run scheduled automated workflows that implement these rules against monitored portfolio and watch list assets. The workflows periodically fetch market data, calculate the indicators, evaluate the conditions, and – if a signal fires – send an alert to a Slack channel. The automation handles monitoring only – it does not execute trades. When a signal fires, I review the alert and decide whether to act or not.

Try It Yourself

A rules-based system’s real value isn’t finding better opportunities – it’s the guardrail that keeps you consistent and honest about your own decisions.

This is the first post in the 88capital series — next up is the signal logic in full, and the set of workflows that realise the whole thing. If you want to replicate the stack, IBKR is the broker behind the thematic layer and n8n runs the automation. Follow the 88capital experiment page for the rest of the series as it publishes, or skip ahead with the 88capital workflow bundle.