We spend a fair amount of time on this blog looking for trading systems that have some discernible and consistent edge over the market averages. And edge is important: if you don’t know where your profits will come from, you certainly shouldn’t expect to have any. But an empirically demonstrable edge is just one of the three components that comprise any good trading system. The other two, execution and allocation, are also essential to the success of any trading program or system. This is an admittedly basic, big-picture approach to thinking about trading. But it sure beats a vague or unreflective stance.
Edge, again, just refers to whatever it is about your system that generates positive expectancy or, hopefully, market-beating returns. We’re not dogmatic about alpha generation, but we don’t do faith-based trading, either. The key rule of thumb for claiming an edge is: if you can’t quantify it, it doesn’t exist. Ideally, finding an edge should be the least difficult part of developing a trading system, in the sense that cognitive biases, emotions, and human error can in principle be entirely eradicated.
The process of finding an edge tends to go something like:
- Intuition – some hunch or probing question arises about a relationship or tendency among various products, indicators, timeframes, etc.;
- Quantification – the intuition is formulated as precisely as possible so that it can be evaluated. What was previously a jumble of notions becomes an array of variables and conditional statements;
- Testing and optimization – both walk-forward and traditional backtesting are essential to ensure robustness and avoid curve-fitting, and there are plenty of other key constraints;
- Application – even the most cautious testing procedures aren’t a substitute for application under live conditions.
The quantification criterion (“if you can’t quantify it, it doesn’t exist”) isn’t negotiable. That criterion may exclude approaches to fundamental analysis in which the “story” of a stock or company plays a non-redundant role beyond the balance sheet review; and it may exclude large swaths of what passes for technical analysis, especially when it comes to non-classifiable pattern recognition and ill-defined support and resistance specifications. But the good news is that no indicator or technique need ever be dismissed on ideological grounds – if it can be tested, then it should be.
Execution used to be an area where some traders had enough of an advantage that execution was itself a source of trading edges. But trading floors have been emptying out for years precisely because trading technology and electronic exchanges have levelled the playing field, meaning that excellent trade execution is now a necessary but not usually a sufficient condition for success. Any decent brokerage will time its order routing in milliseconds, with some platforms claiming sub-millisecond latency.
However, better data feeds and faster routing aren’t the whole story. Traders still must attend to the execution of individual orders, and execution becomes especially artful (i.e., more art than science) when it comes to filling option spreads. Sometimes, for example, closing an out of the money spread becomes impossible if one of the legs you’re selling is at zero bid. And while execution may seem to be largely a technological question, psychology actually plays a major role here: should you give up an extra penny or two on your price in order to enter a trade now, or hold out at your price and work the order for the next several minutes? Again, the consistent application of some rules for trade execution will never be a profit center, but it will always be a prerequisite for successful trading.
Sound risk management can atone for many sins. Flawed risk management can be your damnation. This isn’t the place to discuss principles of trade allocation and risk management as such, but we should say that it’s worth thinking hard about how the characteristics of a given strategy can best be implemented in the context of position sizing and money management. A strategy that experiences wild swings in profitability shouldn’t be treated the same as a system that is a steady earner but with lower returns. Here again, insisting on an empirical approach actually makes the process much easier. The Kelly Criterion is helpful for starting to think about money management; Messrs. Sharpe, Sortino, and Treynor help us think about volatility and risk in assessing performance.