Trend following

• All entry signals follow a ‘buy on strength, sell on weakness' philosophy
• The system uses an actuarial stop-loss mechanism, exiting profitable positions via trailing stops, rather than at price objectives. We do not aim to predict where the market is going, and do not know in advance where or when a trend will end. Therefore, we only exit when our system indicates a strong likelihood that a trend it is over
• The system increases winning positions as unrealised profits grow, but never adds to losing positions
• In any given market the system can be long, short or flat: the entry and exit decisions are separate, and the system rarely reverses

These are desirable properties because periods in which markets trend are punctuated by price consolidations.

Diversified

• The broad diversification across all market groups serves to smooth P&L fluctuations
• All instigations and exits are executed piecemeal, thus reducing the importance of any single execution, or any single buy/sell decision
• The system actually uses a combination of separate sub-systems. Each has the same structure, but uses different settings for the two main timing parameters

Rigorous and 'general'

• A heuristic procedure is used for selecting the instruments we trade. The guidelines are driven by correlation and liquidity considerations. For example, if liquidity were to evaporate in a particular market, it would eventually be dropped. Otherwise, trading signals are followed without discretionary input and money management is completely formulaic
• Exactly the same parameters are used for every market

Long-term

• Long-term trends are considered to be of around three months to a year in duration. It might be argued that 'long-term' should apply to time-frames of a decade or longer (some technicians argue that the British pound has been trending down against the US Dollar since the end of the American Civil War in 1865), but even three months is a long horizon from the perspective of most participants in the futures markets
• On average, the system takes a position in each market once or twice per year and holds each trade for six months
• The low trading frequency helps keep transactions costs down; it is feasible to execute trades and monitor positions in a large number of commodities without requiring major human or systems resources

Good money management

• Position size is determined as a function of market volatility and pooled account equity. Individual market weightings are determined by proprietary algorithms drawing heavily on modern portfolio theory
• The initial commitment in any market is small: 0.50% or less of account equity
• The system takes bigger risks in profitable periods and conserves capital during drawdowns: during profitable periods, position size is increased at a slower rate than equity growth; during drawdowns, position size is reduced more rapidly than equity
• Stop-losses are a function of market volatility and are 'wide', which helps to lower the frequency of trading (reducing transactions costs), reduce position size (aiding execution) and avoid short-term market randomness (preventing us from being 'whipsawed')

Execution

• All signals are generated on a close-only basis and executed during the following trading session. This allows the system to avoid intraday noise and take advantage of maximum liquidity. It also means that the quality of executions can be very closely monitored versus settlement prices, and that the variance between hypothetical and real-time results should be minimised (in contrast to a system depending on intra-day execution, where slippage assumptions would be much more problematic)