>>/127/
next step is training it to improve the chains based on reaction level with a neural net