Lin Wood is at the top of his game, but elections are not his area of expertise. I would guess that he is extrapolating from information he knows and getting the number wrong (probably by one order of magnitude). I think the statistical anomalies are pointing to a fraud that is under a 10% change.