Conceptually speaking, all three Washington governor ballot counts appear to have fallen within the margin of error for such machine and hand counts.  Each count of the several million ballots has spread over a swing of several hundred votes.  The last count seems well within any tolerance of error for whether for machine or human hand counting.  Scientifically, this would make the results ?inconclusive.?  It would not seem that an election could correctly result from inconclusive results. 

I would suggest that a mathematician/statistician review the ballot tally information in order to assess its actual value.  Scientific inconclusive results should render all three counts invalid.