If you missed  many reads, for example if your system was setup incorrectly, you can use this function to interpolate missing times.

When using Splits then the Predicted Times for Splits are used, in this way a manual raw data record is created at the predicted time. 

When using Results, you need to select the result you would like to interpolate and a "helper" result. If for example you interpolate the start with the help of finish, the following will happen:

For each participant missing the start time, the system looks for the 5 participants that finished right before him/her and the 5 participants that finished right after him/her. The system then calculates the average start time of these 10 participants and uses this average as an interpolated value.