Error term in Time Series Analysis

Schweser book 1, page 230 if Predicted value < Actual value, error term is +ve Schweser Book 1, page 220 If predicted value < actual value, error term is -ve there’s a conflicting remark??? which one to go with …

From Econometrics book by Wooldridge: residual (error) is the difference between the actual yi and the fitted value: error = yi - yhati. So in this case I would say that if the actual value is greater than the predicted, then the residual is positive. Think of it as the direction you need to move away from the regression line to reach the actual value. Also look at the the calculation on page 220 above the explanation. The error is positive there, 0.5

Thanks a lot.

what does the CFAI text has to say on this … any 1 ??

there is an error on page 220 (see errata from schweser website): should read: “since the predicted value of y for 2004.1 is less than the actual, measured value, the error term is positive” follows that if predicted value > actual value, error term is negative

Are we suppose to be studying quant this year???