I think that the scalping problem is situated in two keys:
Avoid high-impact news, which is becoming more and more unannounced, and that to develop such a strategy, backtests are not useful due to the distortions that backtests generate with respect to reality.