[theincredulousk]

Devising a set of parameters that backtests 100% means nothing - there are an infinite number of parameters and models (or more accurately, algorithms) that can backtest any data set like this 100%.

In that sense, the math and "indicators" are just an obfuscated, custom-fit version of another algorithm that backtests 100% - if price[now] > price[now+1] -> sell else -> buy.

Rather than looking up the prices via array in the data, clearly "cheating", a set of parameters is hunted down that encodes that relationship in a number space. It doesn't mean anything more than the meme: 79 beers - your age + 40 dollars = the year you were born (which backtests perfectly as long as you were born after 1900).

That said, by all means, keep running the algorithm on new data every day and see how long it stays perfect. I'd be interested in the results.