Eight out of 10 times, the non-incumbent party’s candidate — that’s Trump this year — gained in the polls after the first debate. That includes each of the last five times. There are various theories to explain this. Some people think, for instance, incumbent presidents do poorly in first debates because they’ve had four years to grow unaccustomed to being challenged so directly, or that the challenger benefits simply by being on an equal playing field with the sitting president. Those theories don’t apply this year. That said, there are other reasons to think Trump has more to gain. He is currently winning a lower percentage of self-identified Republicans than Clinton is getting Democrats, so perhaps he has more lower-hanging fruit than Clinton: More Republican voters may come home after seeing both Clinton and Trump in action.

Another pattern that jumps out from the table above: First debates haven’t moved the polls all that much. Only twice did the leader change: Ronald Reagan moved ahead of Jimmy Carter in 1980 and won the election easily, and George W. Bush overtook Al Gore in 2000 and won the Electoral College (but not the popular vote). The average change in the margin between the two major-party candidates has been about 2.6 percentage points. Of course, with Clinton up by only 2 to 3 percentage points nationally, even a modest-sized Trump bounce could give him the lead.

There are also a couple of factors at work this year which could allow the first debate to swing the polls more than usual. First, it could be that at a time when social media can cement a narrative within minutes, a good debate performance matters more than ever. The largest first-debate gain since 1976 occurred in 2012 when Mitt Romney picked up 4.4 percentage points and drew into a near tie with President Obama in national polls. Four years before that, in 2008, Obama expanded his lead over John McCain by nearly 3 points after the first debate.