i believe a system should be traded exactly how u found it, however analysis into each trades (analysis NOT involvement) could reveal important stuff and individual trades that you would make you reconsider the system, for example
are the gains equated for the most part where each trade and loser is about the same?
is there one trade or few trades that are responsible for a large amount of the overall profits or losses?
is there a time period where most gains or losses came from? if so u can eliminate that period all together
whats the worse loser? could be prevented? if so how many winners would it have erased with it? (for example if a major loser was because of a support broken, then before u say i can simply not trade if support is broken, u must look at the times where support was broken and the trade still worked and was profitable)
what ever changes you do it must be standardized in my opinion, meaning a change that is from the top to the bottom that the same standard is being applied to every trade. the more u stray away the more discretionary your becoming without knowing it.
keep in mind my trading is swing and back testing there might vary drastically from yours
on Brexit in one of the pairs i was trading i came in long and i got hammered that day but overall the pair was profitable
for now the one thing i do make sure is that if what iam doing doesn't lose me a lot when it stops working cuz its going to since nothing is forever, iam doing that by integrating the stops in the system and making sure the losses are significantly smaller than the winners, what u think bud?