I don't see it. Being a person that never kept up with how "turnout" was measured in the past, upon looking at the issue it seems obvious to me that the correct methodology for measuring "turnout" is to see what percentage of registered voters showed up at the polls.
The idea of using people who never bothered to register as part of the equation would never have occurred to me.
But let us say for a moment that your argument is correct. What is the normal "turnout" for a presidential election measured the way you say it has always been measured?
About the turnout percentages:
You can calculate turnout as a percentage of registered voters, or as a percentage of all adult citizens who could have registered and voted. The latter percentage will obviously be lower. (For example, if there are 224 adults, of whom 158 register, and 138 of them vote, then the turnout is 87.3% of the registered voters but only 61.6% of the adult population.)
Each statistic will tell you something interesting. What doesn’t make sense is to use different calculations for successive elections and then argue that the change reflects something fundamental. It reflects nothing but statistical sleight-of-hand. That’s what’s going on when people claim that turnout in 2020 jumped from 60-some percent to 90-some percent.
The lower number is the one more commonly used. By this measure, turnout was up about six or seven percentage points in 2020, a year when many jurisdictions made it easier to vote by mail. Recent presidential election numbers:
2000: 54.2%
2004: 60.1%
2008: 61.6%
2012: 58.6%
2016: 60.1%
2020: 66.6%
Biden 2020 got about 15.4 million more votes than Clinton 2016. Does that show Democratic fraud? Well, Trump 2020 got about 11.2 million more votes than Trump 2016. Does that show Republican fraud? I prefer the obvious explanation: EACH party got millions more votes, because millions more people voted. Trump’s share of the vote actually increased, from 46.1% in 2016 to 46.8% in 2020. Biden won because he racked up an even bigger increase over Clinton’s share (51.3% versus her 48.2%).
That three-point difference could have been fraud. Or maybe a lot of people just didn’t like Hillary. The overall numbers don’t prove or disprove the allegation of widespread fraud. You have to look at the specific cases.