Our ledger does not lie, if it shows adding to position is piling bigger loses then there can be no argument in favor of scale-ins.
Today, i tried adds after eyeballing a few charts. That method still worked fine - when i checked charts EOD. The issue was that I could not execute is accurately. By 10:45, i shifted to a slightly different method of adds, which was easier to execute, but those adds only kept eating profits of my initial position. Part of it was also luck, because when in the first uptrend, i had deployed 100% of my capital, and if not for a slight pullback, my account would have increased a lot...
I'll move the "add" experiments to later, until the itch to try it kicks in again...