Skip to content

Method

How a portfolio of strategies is built.

  1. Hypothesis
  2. Out-of-sample tests

Limits

  1. Strategy
  2. CapitalPortfolio
  3. Order

Every quarter, five tests: kept, reduced or retired.

A strategy receives capital only with evidence; after that, limits on every order and tests every quarter.

Three questions before allocating capital.

Why might it work?
An economic hypothesis is checked against data, costs and tests outside the sample used to develop it.
When could it lose?
Adverse scenarios and the conditions that would weaken the thesis are examined before capital is allocated.
What does it add to the whole?
The exposure it adds is weighed, along with how it relates to the other strategies, under stress as well.

01Quantitative research

A tested reason, not a hunch.

Research is where Athena looks for that reason and puts it to the test, with systematic models, global macro analysis and statistical models of regimes and volatility. Every idea faces out-of-sample testing, free of look-ahead bias, and receives capital only if the evidence holds.

  • Multi-model research: systematic models, global macro, and machine learning
  • Factor research, signal extraction, and regime analysis
  • Multi-asset coverage: currencies, rates, equities and indices, through listed futures, options, ETFs and stocks

02Strategy design

A tested idea isn't a strategy until it can trade.

Design turns research into executable strategies — systematic and discretionary, each run through the same process — with defined entries, exits and sizing. The aim is a strategy built to trade in real markets, not to win a backtest.

  • Defined edge, size, risk budget, exit, and portfolio role before receiving capital
  • Robustness over optimisation: tested against adverse scenarios, not tuned to win the backtest

03Portfolio construction

Real diversification doesn't come from owning more things.

It comes from deliberately combining strategies whose returns are uncorrelated with each other. Each strategy receives a risk budget; combined, they seek to fluctuate less; and the portfolio is brought to the target volatility the client chooses: 6%, 9% or 12%.

  • Strategies combined as return streams uncorrelated with each other, a relationship also measured under stress
  • Built deliberately as a whole, not accumulated trade by trade
  • Market exposure is a decision, not a consequence.

04Risk architecture

Risk is managed at every level: position, strategy, portfolio and firm.

Exposure is measured with methods such as Value-at-Risk and conditional VaR, and reviewed as markets move. Risk management is a core discipline of the process; limits do not remove market losses.

  • Written limits in each mandate: loss, margin and leverage, by risk level.
  • Every order is checked against the mandate's written limits before it is sent, and Athena's records are reconciled with the broker's every day: Athena's own systematic controls
  • Scenario analysis and stress testing
  • Target volatility: if measured volatility rises, exposure is reduced in proportion.

05Execution discipline

A good decision can still be lost to careless execution.

Each order's size takes the instrument's liquidity into account, and trading costs are measured against what each strategy contributes. Athena trades listed instruments (futures, options, ETFs and stocks) and spot currencies, with no OTC derivatives.

06Quarterly review

Each test has a written consequence.

  • Two nearly identical strategies: no new capital.
  • A systematic strategy is retired when it crosses its written criterion, not for a bad period.
  • Off-model decisions over the threshold: reduce or close only.
  • Too heavy or too costly: it stops receiving capital.
  • Cost consumes the benefit of combining: the one that adds least leaves first.
View thresholds in the model mandate (in Spanish) →

The account is in the client's name at Interactive Brokers, which holds it in custody. The broker's custody is not supervision of Athena.

How Portfolios works →

Talk to Athena.

Every question, comment or proposal to collaborate is welcome.

Contact