We prefer not to use the simplified methods, but to use simulation-based methods for estimating WTP.
That said, extreme outliers are known to occur in respondent-level data. Any respondent who has very small differences in their price utilities will have very large WTPs. For respondents with utilities of 0 for each price point, WTP will be so large as to be undefined. It's simply the mathematics of dividing by a number that can sometimes be very, even vanishingly, small.