if you do not understand that a valid model test is running a model from any start date and initial conditions to assess how well it forecasts real temperatures as they developed from there then there is no hope for you here's a little more of why spencer is misleading
