Not a revolution. A quantum evolution in portfolio optimisation.

The starting point: data

A.L.F.R.E.D. (Adaptive Learning For Ranking Equities and Derivatives) is Avangard's proprietary investment system. It is 100% owned and built entirely in-house, and Avangard does not license a third-party quantitative platform. It draws on up to 100 years of cleaned historical financial data and processes roughly 28,000 daily price inputs across more than 2,500 ASX-listed securities and ETFs, refreshed and re-ranked every trading day. The scale is deliberate: a signal is only trusted once it has been tested across enough securities and market cycles to separate it from noise.

Finding signal in the noise

The research behind A.L.F.R.E.D. began 24 years ago with momentum, the tendency for securities that have performed well to keep performing well over the following months. That was the starting point, not the finish line. Since then, the system has evolved continuously. New techniques are researched and tested rigorously against decades of market data, and only adopted once they demonstrably separate genuine signal from noise. What runs today is not a single model from two decades ago. It is a technology stack that has been refined and extended every year since, and it keeps evolving as new evidence emerges.

The machine works for us, not the other way around

A.L.F.R.E.D. amplifies human judgement, it does not replace it. Outputs are reviewed by the investment team, and the system processes far more information than any analyst team could before filtering it into disciplined, explainable investment decisions. The process is systematic by design, but our portfolio manager retains oversight and judgement throughout. Rules drive the day to day construction of the portfolio, and a person remains accountable for the outcome.

From ranking to portfolio

Each trading day, A.L.F.R.E.D. produces a ranked view of the eligible universe, and the portfolio is built from that ranking within position sizing and risk limits set in advance. This removes the noise of daily chatter, broker tips and short-term sentiment from most decisions, while leaving room for the portfolio manager to apply judgement where it genuinely adds value, particularly around risk and unusual market conditions.

Managed risk measures, built in rather than bolted on

Risk management is part of the process itself, not a separate overlay applied after the fact. Position limits, sector and factor exposure controls, and portfolio-level risk budgets are built into the same framework that drives stock selection, and the portfolio can shift toward higher cash or more defensive positioning as part of that same process when conditions warrant it. This does not eliminate investment risk, and no investment strategy can guarantee protection of capital, but it means risk is actively and continuously managed rather than reviewed only in hindsight.

A genuinely different portfolio, not another version of the index

Because positions are built from a systematic ranking rather than an index weighting, the resulting portfolio has historically looked materially different from the broader market: different companies, different risk drivers, different return payoffs. This differentiated profile is why the strategy is designed to complement an existing Australian equity allocation rather than duplicate it, and it is a direct outcome of the process described on this page, not a separate feature layered on top.

What this means in practice

• Every decision is traceable. Each position can be linked back to the data and rule that produced it.
• The process does not get tired, anchored, or emotional. It recognises when market conditions shift and adjusts positioning accordingly, without ever losing discipline in doing so.
• The strategy is not infallible. No systematic process outperforms in all conditions. Periods where market behaviour deviates sharply from historical patterns can weigh on returns, and managing that risk is part of the process rather than an excuse when it happens.