Honestly, this strategy just like any other strategy will 100% fail in the long term which is why I advise taking your profits and running for the door asap in the short term. 400 sample size is way too small.
I have seen countless strategies fail in the long term just like the one you mentioned. Don't be greedy op. Take the short term profits.
I second this. 400 as a sample size is way too small and might give you a false comfortable feeling with your strategy. Chances are very high that you get rekt when you try your script with your own money.
Anyways its plain simple: The algorithms are usually bullet proof and cant be fooled with a "strategy" as far as there are no major bugs in the code. The longer a site runs the lesser is the probability that there are such critical bugs in the code because the house usually monitors its statistics closely and would see any flaw immediatly.