Guys, any one know of a reasonable tool that optimizes DCA settings for the ASAP modes based on a coin you’d input? I.e. you put a coin inside of the tool, and it does some backtests and provides a few good DCA settings.
Like an automated strategy optimizer? We have plans to make one, but it is not so simple, as we also need an overfit test. Strategy optimizers test thousands of configs and are very likely to end up with an overfitted result.
Maybe instead of a brute-force testing, there could be an algo approach to proposing a reasonable DCA for a coin, based on its past 30-day behavior, let’s say using ATR?
I don’t know any other way a program can test other than brute forcing. ATR or whatever it is you need to tell the program what to test, and create many variations of it.
Here’s an idea. Start some kind of a machine learning to discover the best way of discovering, i.e. first discover a strategy for discovering an algo that would produce an optimal setting. This can later become proprietary algo of Gainium. Once you teach your AI a most optimal way to discover DCA, instead of brute-forcing, it can then be applied into a select pair.
I think is not easy to do something like that and it would take a company specializing on AI to train a model that can understand financial data. The model needs to understand all strategy KPIs, calculate them, and run monte carlo analysis or similar to detect overfitting.