I'm afraid the contest and this website is all about coding. And I'm afraid it is correct that it should be. You have to decide how to make movements in and around the standard deviation bands a "factor". The optimisation process then minimises or maximises that factor. In other words it chooses a portfolio, a vector of stock weightings, based upon how well stocks reflect this "alpha" factor.
You might find it helpful to read up on optimisation methods. Gradient descent. Calculus. Machine learning.
I don't think there are any short cuts really. For better or worse Quantopian have decided upon a specific investment approach and it is down to you to understand that approach and how to code it.
I'm not wishing to sound unhelpful but systematic investment is not something you pick up overnight. And portfolio optimisation is not that simple a topic.
But if arts clowns like me can pick it up, anyone can. And don't get put off by my Ecclesiates cynicism. I'm sure there is gold to be discovered in Quantopians approach somewhere.