I completely agree but hadn’t found a way to put it to words. It could be the model too trained on optimized strategies