Here's a good example of overreacting to a small sample of results from another Indiana election ...

Results are starting to trickle in from the first few counties in Indiana and Kentucky! However, they're incomplete, and it's far too early to draw any conclusions from them. My general rule of thumb is to wait until at least 10% of the expected vote is reporting before taking those numbers seriously.

On election night, information overload is almost unavoidable. With this in mind, 538 is here to offer you a guide for knowing whether a state might be in play as votes are tallied. We estimated county-level benchmarks for every state (save Alaska, which doesn't report votes by county) to offer a back-of-the-envelope gauge for how the election is going.

The benchmarks estimate what percentage of the vote Democrats or Republicans would need in each county for a statewide race to be roughly tied. That means if a candidate is performing consistently better than their party's benchmark as we get a meaningfully large number of returns (keeping in mind that the partisan split from early returns may shift significantly as the count continues), it's a decent sign that contender is on track to come out ahead in the final results. In the interactive, you can look for any state or county in the search bar above the map.

To explore how to use the benchmarks, let's take a look at one of the key swing states, Pennsylvania. On its state page, you'll find a county-level map and a table with benchmark figures, along with the share of the statewide vote that each county contributed in the 2020 presidential race. Above the map, you can toggle between the statewide races, choosing among president, Senate and governor (where applicable). And above the table, you can toggle between viewing benchmarks by margin (how much a candidate leads by) or by party vote share (what percentage of the vote each candidate has).

The map displays both the benchmark margin and the 2020 share of the vote from each county. The bluer a place, the more Democratic it would be in the case of a statewide result that's about tied, and vice versa for red and more Republican. And, the larger the bubble over a county, the more votes it contributed in the 2020 presidential election.

Take Philadelphia. Heavily Democratic, it contributed nearly 11% of Pennsylvania's statewide vote in 2020, so it's represented by a big, blue bubble. Based on our benchmark, we can expect the statewide result to be very close if Harris is leading in Philadelphia by about 62 percentage points (80% to 18%). If she's doing a bit better than that, it could signal that she's running ahead of where she needs to be in order to win; if she's doing a bit worse, it could be a sign that Trump has an edge in the state.

Of course, you should never rely on just one data point as determinative. Instead, you'll want to consider what the returns are saying across a large number of counties, and only once a large portion of the vote has come in from each of those places. Additionally, you might look at a place like Erie County in the state's northwest corner, which has a reputation for being one of the state's bellwethers. Tellingly, our benchmark there is basically a tie, so if one candidate is running a few points ahead, that is probably a good sign for them.

We hope these numbers prove to be a helpful shorthand for who might be leading in close races as the returns come in. And even in the uncompetitive states, they could provide a hopeful Republican a sense of what it would take for their party to carry California in this day and age, and a hopeful Democrat a clue as to what their party would need to compete in South Dakota!

It's 6 p.m. Eastern, which means the first polls of the evening have closed! Voting is now over in most of Indiana and in the eastern half of Kentucky. We should see initial results shortly ...