We periodically improve our calling algorithm and reference data, and a newer result is generally the more accurate one. Occasionally an earlier reading is corrected because it rested on an inconsistent marker pattern — for example, a downstream marker appearing without the upstream marker it must sit beneath, which isn’t biologically possible. Correcting that moves you onto the branch your markers actually support.
Why did my result change from an earlier version?
Why a newer result is usually more accurate as we improve the calling algorithm and reference data.