I go through cycles and spells where I like to test out new ideas. This site has literally been like giving crack to an addict, as you can't help but read a few posts and it totally opens me up to a whole infinite number of permutations for other ideas.

Teflon was discovered by accident. Some of the greatest discoveries in history were by people who were researching other problems.

To that end, I have discovered at least one viable strategy not by design, but through failure.

I normally do like most algo guys do, I come up with an idea while washing my hair in the shower, then I go research and learn a little more about it, then I go start doing some preliminary spot checks rearward in time to see if it's worth exploring further and then I code, attempt to backtest, then if it has promise, I'll forward test and decide if it's viable or if it deserves going into my bag of tricks.

While backtesting, I discovered that one of my ideas failed not only miserably, but really miserably, on a really consistent basis. So it gave me a grand idea (that I'm sure most of you have had before).....what if I "reverse the polarity?" i.e. if I'm getting absolutely kicked in the groin on both the long and the short side, what if I switch sides?

Low and behold, what I thought was a fairly good idea and a metric for a new strategy, turned out to be completely wrong, but the opposite holds true and with fairly good results.

Have any of you crafted successful strategies this way?

I would consider that "amateurish"...you don't even get a "cookie" for figuring out volatility regimes.....what else you looking for? I have yet to see a retail "algo" that I have even the slightest bit of respect for....
That just is not the game.....dumbasses with connections can raise hedge fund cash...those who can not...
Those who blow at their strategy and also can not raise funds get destroyed....
Exactly how money should flow.
It is a perfect market in some sense...