Quantitative Trading•寒蝉•3dOpinion

Choosing Grouping Criteria in Quantitative Backtesting

When conducting quantitative backtesting, applying the same set of factors to all stocks is not only meaningless but also affects the strategy’s performance on individual stocks. After all, different stocks have different characteristics, so so-called universally applicable strategy parameters should not be forced onto them. It is therefore necessary to perform rough screening and group stocks, but the screening and grouping criteria must have a causal relationship with the subsequent strategy choice; otherwise, the sample size may become too small to be statistically significant.