I think in general this was a well-written article about electronic valuation systems used by Zillow and Redfin.
But I disagree with the advice at the last paragraph, which recommends comparing the value of my house between what Zillow estimates and what Redfin estimates. Considering the article was about statistical accuracy, using a sample size of N=1 or 2 (my home and another I know well) to determine whether I should trust Zillow or Redfin is falacious. Valuation accuracy will vary not only by city but by neighborhoods, especially when substituting machines for human judgment.
Read the full article in NewsOK.