Jungle-Sven
commented
1 year ago The logic behind calculating stationarity is quite simple. We analyze historical data to determine how often did prices revert to mean values and assume that this effect is going to stay in the future. This helps us to select the best markets for mean-revert strategies. As well as we can choose markets with low stationarity for trend following strategies.
If the sequence is stable, mean regression makes money. If it is not stable, mean regression loses money. Then why not use the strategy to directly test whether it is losing money or making money, instead of testing whether the strategy is profitable after measuring the stability