There is a lot emphasis these days about developing a quantitative skill set for markets. Learning Python for example helps you do number crunching on large datasets. Having a grasp of statistics enables you to decipher the relationship between various market variables. All this stuff is important if you want to develop a trading strategy. It requires time and training to pick up these skills. If you want to delve into high frequency trading, you could argue that the technical aspects are even more most important, being able to write code which executes very quickly. Having code which executes slowly, cause result in excessive latency, is not great if you are having to market make at very high frequencies.
However, at all trading frequencies, when it comes to developing systematic trading strategies, I think that a quantitative skill set is not the only thing you need, even if it is an important prerequisite and building block. When you are trading regularly, having an understanding of market liquidity is crucial (or if you are trading large sizes too). What might look theoretically “nice”, can often be killed when appropriate transaction costs are factored in. Furthermore, we need to understand the differing levels of liquidity in different assets. Something very high frequency, which might work in EUR/USD, might not necessarily work in NZD/USD, because of the disparity in liquidity. There’s also a question of how deep that liquidity, which is key, when we are trying to scale our strategy.
What might look on paper (with improper assumptions) could be untradable in practice. This can unfortunately be the result when quants are totally disconnected from the market. It’s also the case that the return profile of a strategy can make it difficult to trade in practice. If it has repeated drawdowns, but “works” long term, it can be very difficult to sit on your risk. The skill set to risk manage exposure is different to the skill set involved in developing a trading strategy. A good strategist won’t necessarily make a good trader.
At lower frequencies, we still need to understand the vagaries of market microstructure, when it comes to execution. We also need to consider the economic factors which drive price action, and the way they interact with the market. How do long term shifts in capital flows impact FX for example? We can try to “data mine” relationships, but it seems more intuitive to have an understanding of how markets work, and then go from there. After all, we do not wish to find spurious correlations in our analysis.
There isn’t a standard way to learn how to build a trading strategy. I would argue though that whilst it’s important to build your quantitative skill set, just as important is getting the intuition of how markets work, to avoid creating strategies that are overfitted and cannot be executed in practice. Think the second part is unfortunately not emphasised enough.