Let either the data or your inner voice speak to you

 Something I have been coming across lately is the loose debate between those who believe in more automated, quantitative trading strategies versus proponents of discretionary trading. I've commented about this in earlier posts, but it's worth revisiting a bit to at least ponder which one suits you better (or maybe a hybrid?).

Creating an automated strategy requires that you collect stats and back-test specific scenarios that are a function of very quantifiable variables such as closing price, volume, time of day, volatility, short interest, etc. What this means is that you're essentially leaning into data science to forensically find an edge from all the data that you're examining. Some of this data may even be more on the qualitative side, such as sentiment analysis based on Twitter feeds or your mood after a questionable night's sleep. But at the end of the day, your trading decisions still come down to an elaborate "if you see X, then do Y" set of rules (that can be captured in code).

Discretionary trading, on the other hand, is more akin to an apprenticeship. A surgeon, experienced engineer, or CEO doesn't have a precise blueprint or decision matrix. And yet all of these people are able to excel at their jobs in an often predictable, high quality fashion. Similarly, there are traders who "learn" how to trade, an admittedly broad term (but likely confined to a certain asset class, trading style, etc.), but one that implies learning a craft. To go one step further, many decisions require one to gauge the surrounding conditions, relationships and dependencies between different factors, and making tradeoffs on the fly. People are attemptimg to apply machine learning to different fields (including trading) to help reduce the computationally intensive nature of traditional rule or formula-based systems, but these still wind up being static models for specific scenarios. The field of trading is constantly demanding dynamic adjustment, so even applying an ML assistant doesn't quite cut it. 

The future of trading is likely to involve a hybrid approach where ML is used to identify potential winners, and maybe even apply position sizing and risk parameters based on predicted risk/reward, but then the experienced trader will still need to intervene to optimize execution or certain decisions. So do you want to put in the effort to create a statistically sound system backed by data and sit back and let the system execute, or are you motivated to develop an understanding and mastery of how (and to a lesser degree, why) markets move and apply risk management yourself? Both involve quite the journey, and will test your willpower and patience in different ways.