Every HFT outfit has someone on the lookout for these kind of ideas. When this system is found, every idea in a commit in an obscure branch is going to be quickly reviewed, tested inside that outfit's backtest system (which likely has years of microsecond stamped tick data to test on, and a 1000-core strong compute cluster to do that quickly), and if it has merit -- applied before you have a chance to test it in trading yourself.
Your own success in trading this system is not required (and in fact, unlikely, even if you implement things that would work if done in a non-public way).
I'm not trying to get you not to do it - on the contrary, I love to follow interesting ideas in the field even though I'm not in it anymore. I'm just trying to give my sober, experienced, point of view - if you want it to succeed, this particular thing has to be closed source - you should only commit to public git after your edge has waned....
