Take it from a man who’s been around the Wisconsin voting system a few times himself, including a recanvass or two. Scott Walker explains the problems with scope and scale in recounts to CBS This Morning, and at the same time debunks a hypothesis floating around in social-media circles. What might actually help is a recanvass, a check of numbers from each precinct that might have introduced error into the state voting counts.

Had he ever seen a recanvass correction on a scale that might overcome a 20,000-vote gap? Actually, one did come close nine years ago, Walker explains:

“Typically when you have problems out there,” Walker explains, “it’s not corruption. It’s somebody made an error.” A recount would correct that, but a recanvass would take place first, and that would correct it more easily.

How about a recount? “Is there any record,” Walker was asked, “of any error large enough to actually change the outcome in Wisconsin?”

“Close as you’ve got in 2011,” Walker replied, “was about 14,000 votes. That’s shy of the 21,000, but it’s at least in the ballpark.” That came in a Wisconsin Supreme Court election, and it was an error that actually impacted the election outcome.

What about reports of votes in precincts exceeding the the number of registered voters prior to the election? Some people, including members of Donald Trump’s political team, point to those precincts as evidence of fraud. The answer is likely much simpler; as Walker explains, Wisconsin allows same-day voter registration, so this may not even be an error. Walker also points out that the Trump campaign really drove new voter registrations in their ground game and GOTV efforts, so that may in fact be part of where Trump’s strength originated in Wisconsin on Tuesday.

PolitiFact pointed out the same thing yesterday:

Wisconsin is one of 19 states (along with the District of Columbia) that allow same-day voter registration. That means the correct comparison is eligible voters, not registered ones.

More than 3.6 million Wisconsinites were registered to vote as of Nov. 1, according to the Wisconsin Elections Commission. That was two days before the election.

The number of people turning out across the state topped 3.2 million, the Milwaukee Journal Sentinel reported — the most votes ever cast in a Wisconsin election, with at least 71% of the state’s voting-age adults casting ballots. (Not quite the highest percentage-wise).

The Wisconsin Elections Commission weighed in on Twitter later in the morning, explaining that registration numbers reported by counties can be off because of same-day voter registration.

In high-turnout, high-energy elections, more people will register for the first time, and likely that will result in votes exceeding the previous number of registered voters. In same-day registration, we can be reasonably sure of 100% turnout among those newly registered voters, after all. There are reasons to worry about vote security with same-day registrations — say, people who are trying to vote twice or who are ineligible to vote at all casting ballots — but the above complaint is simply without merit.

Let’s get back to recounts, which are almost always extremely limited in terms of impact. They do matter when gaps are in the hundreds, but rarely have an impact outside of that scope. If a recanvass doesn’t turn up a data-entry error, don’t put much hope into a recount. Even Al Franken couldn’t change more than 500 votes in Minnesota when Republicans were taken unaware by the aggressive ballot challenges his team made. It’s just that 500 was all Franken needed to change.