This report summarizes the results of fitting of Recycler optics model to the orbit response data taken in February and October of 2006 at two settings of the machine working point. The quadrupole gradient errors obtained with this fitting are different from previous results. The problem was tracked down to the fact that kick of some (‘LEP-style’) correctors used to excite the orbits appears to be 60% off from the design value. The new model presents an improvement in terms of capability to predict the optics performance.