The treatment is considered working when it produces better results then placebo. When the treatment is undistinguishable from placebo, then the conclusion of trial is that "improvements are result of chance rather then treatment".
Placebo is used so that patients, doctors and anyone else is not affected by their expectations/wishes of whether treatment should or should not work.
The whole point is to measure actual physical improvements. You can't replace real treatment by placebos and expect them to have the same results - clinical trials are done so that this is guaranteed to not work with drugs that passed them.
The conclusion is most definitely not "improvements are result of chance rather than treatment". You've also used quotes here but there doesn't appear to be any reference so I can only assume you're quoting yourself :)
I assume you didn't read (or perhaps chose to disregard) the article I referenced in my original comment as that provides evidence to the contrary of what you are claiming. Another much shorter article that you may like to read states "Experts have concluded that reacting to a placebo is not proof that a certain treatment doesn't work, but rather that another, non-pharmacological mechanism may be present." [0]
Placebo is used so that patients, doctors and anyone else is not affected by their expectations/wishes of whether treatment should or should not work.
The whole point is to measure actual physical improvements. You can't replace real treatment by placebos and expect them to have the same results - clinical trials are done so that this is guaranteed to not work with drugs that passed them.