Election officials in Newport News, Va., examine uncounted ballots on Dec. 19. (Ben Finley/Associated Press)

A very simple principle guides the determination of popular vote-based elections in the United States: The outcome is based on a counting of clear and valid ballots. We make no assumptions. If a ballot is unclear – if there is any doubt at all as to the voter’s intentions – it is thrown out. Trying to infer the intentions of voters is guesswork, except for those who have magical powers and can read minds.

None of this has stopped the Virginia Board of Elections from attempting to use said magical powers. According to its guidance, we supposedly know that a voter who has written MY MAN over a candidate’s name voted for that candidate, whereas voters who wrote a candidate’s name without checking the associated box have not clearly expressed intent. The board’s 15-page manual is full of similarly questionable logic.

All of this, of course, pertains to the ongoing drama over the Virginia House District 94 race that at this moment is declared a tie because of the reintroduction of a disputed ballot as an intended Republican vote. If this ballot is not counted, the Democratic candidate wins the election by a single vote. Count the disputed ballot, the election turns into a tie and, absurdly, the winner is chosen by drawing lots.

Normally there would not be much fuss over the outcome of a single state legislative seat. But, in this case, the consequences are big. A GOP win keeps the Republicans in the majority of the House of Delegates by a 51-49 margin — and in control of the speakership, committee chairs and the flow of legislation. A Democratic win means the chamber is tied 50-50, and the parties likely have to agree to a power-sharing arrangement.

The disputed ballot contains what is called an “overvote” — the voter filled in the bubbles for both major-party candidates in the House race. The voter, perhaps realizing this mistake, then put a line through the bubble of the Democratic Party candidate. Because the voter chose Republicans for other offices on the ballot, a panel of three judges ruled that he or she had intended to vote for the Republican for House too.

This judgment seems to violate the Board of Elections’ guidance, which states that to override a previous mark the voter must have “partially erased, scratched out, or otherwise obliterated” it. Is one line enough for a “scratch out”? Does it matter that the graphic examples provided by the Board of Elections all involve much more vigorous scratching? How many back-and-forth motions of the writing utensil constitute a scratch?

On the other hand, the board says that a vote can count if two bubbles are marked but one is further accompanied with an “affirmative Mark,” such as a check. Could the one line in question be a check or other mark indicating intent to vote for the Democrat? Perhaps it was a stray mark.

Admittedly, the odds that a guess about a voter’s intentions is accurate may increase substantially when the disputed ballot shows a propensity to vote entirely for one political party’s candidates, as happened here. (Perhaps not, though, as the vote for Republican gubernatorial nominee Ed Gillespie has an X over the marked bubble, which may be interpreted as an undervote). But even so, trying to divine intent is a dangerous slope. Elections officials should avoid projections or inference of any sort.

Consider: If you filled in two bubbles by mistake, do you put a line through the bubble of the candidate you favor or the one that you oppose? The Board of Elections assumes you are crossing out the one you do not want to vote for. Some voters might view it differently. Another voter in the same situation might try to affirmatively write out the name of the preferred candidate, which the board takes to mean it cannot infer a preference.

Finally, consider that a poll worker may have alerted the voter that the scanner captured the overvote and offered the opportunity to cast a proper ballot. Thus, rather than taking a do-over, the voter may have chosen to leave the ballot as-is, even with the knowledge that the vote would be disqualified. Even without the intervention of a poll worker, we’re all left guessing what happened and what the voter really intended — and that is the whole point: There is no way to know for sure.

To reintroduce the ballot into consideration violates the integrity of a voting process that presumes basic citizen knowledge of how to properly cast a ballot and that every ballot is counted under a clear set of rules and procedures.

The ballot should be tossed and the scheduled picking lots to determine a winner called off. Then the General Assembly should pass a bill making the standards for vote counting clear, fair and not subject to guesswork.

Mark J. Rozell is dean of the Schar School of Policy and Government at George Mason University. Whet Smith, a lawyer, ran as a Republican for the Texas House in 2012.