Shuffling with ridge regression leads to identical cross-validated training and test error?

Question

I am using Ridge Regression to infer housing prices, but i am facing an issue when shuffling data:

this is the plot of cross validated risk estimates (and training error)