This looks like an "actual vs. predicted" plot, with the x axis showing the predicted value and the y axis showing the actual value.
I think your "polynomial" in really a logistic regression fit, but the idea is the same no matter what the model is.
To construct the actual vs predicted plot given input data A, response data B and a model function f, you plot the line y=x and the points {x,y} = {f(A),B}.
More details: