Younger voters today have little to no memory of the Bush-Gore post-election quagmire. For 35 days we had no idea who would win the election and every twist and turn became a major headline. A few lessons from 16 years ago that we need to keep in mind today.

1) Whether certain votes should count was often left to interpretation. Logically, who you think won the election should be independent of who you wanted to win the election. Yet everyone from my friends and colleagues to newspaper editorials to supreme court justices argued for an interpretation that helps the candidate they supported.

Lesson 1: Never trust an argument by someone who cares about the outcome. As much as we try to mitigate it, that even goes for sites based on polls and markets.

2) Gore won the popular vote. Could we have a split popular vote/electoral college again this year? Using the markets on Betfair (as of the morning of 9/30) I computed the probability that Clinton would win the election but Trump would win the popular vote at 6% and the other way around at 9%. There are some large bid/ask spreads so these numbers are very rough.

A Trump popular vote win with Hillary taking the electoral college might cause some to claim the election process was “rigged”, but it isn’t. The electoral college is the constitutionally defined method to elect the president. If we had a popular vote model the campaigns would have acted differently so it is unfair to look at a measurement that the candidates weren’t trying to optimize. Any more that to say it is unfair that a baseball team that got more hits than the other team lost the game.

It should take a constitutional amendment to change the process. There is a National Popular Vote movement where a state would agree to allocate its electoral delegates to the popular vote winner if the states who agree to this process hold a majority of the delegates. This could set up a dangerous scenario where one states delegates depend on the outcome of a state that hasn’t signed up, particularly if a recount is necessary.

The states that have signed up so far are generally democratic non-swing states. See Lesson 1.

3) Had even a small number of candidates who voted for third party Nader in Florida voted for Gore, it would have swung the election. Even the eighth party candidate’s votes would have been enough to swing Florida. A vote for a non-major party candidate has the same effect as not voting at all. You should never vote for a non-major party candidate if you live in a swing state and are not indifferent to whether one of the major party candidates would win.

The ideal solution is to have voters rank order their candidates and use a single transferable vote system so that if you rank order Johnson-Clinton-Trump, your vote would still count for Clinton if Johnson fails to get significant votes. Trouble is such a system is difficult to implement in a way that an average voter can quickly and easily figure out.

There’s vote swapping where, say, a Johnson supporter in a swing state like Pennsylvania agrees to vote for Clinton in exchange for an Illinois Clinton supporter to vote for Johnson. Since there is no verification, the voters may not fulfill their pledges and any logical voter in a swing state should not vote for a third party candidate in the first place. But psychologically it may allow the Johnson supporter to feel more comfortable voting for Clinton.

4) The final lesson is whether you should vote at all. People who didn’t bother voting in Florida in 2000 were kicking themselves. Analyzing the 2008 election data, Andrew Gelman, Nate Silver and Aaron Edlin gave the probability that a single vote matters at about 1 in 10 million in swing states.

This year’s election looks tighter than 2008 but not as tight as 2000. Even at 1 in 10 million the odds that your vote matters is about the same as winning the million dollar prize in the Powerball lottery. What’s more valuable, winning $1,000,000 or having your vote decide the leader of an $18 trillion dollar economy?

Lance Fortnow is a professor and chair at the School of Computer Science at the Georgia Tech College of Computing