How much sewage water can you add to drinking water before it becomes undrinkable?
1) If you assume hypothetically that there was a complete absence of fraud, then you would expect that the changed rules would have produced a significant increase in turnout, with the additional votes being divided between the two major parties.
Changing the rules was the first bit of fraud. The new rules allowed unsecured, no chain of custody ballots into the system. It allowed fake people, senile people, felons, illegals, etc to vote in the system.
2) If instead you assume that the Democrats were ready, willing, and able to commit fraud so as to steal the election, but you know they didn’t need to steal millions of votes to accomplish that, then you would expect that the changed rules would have produced a significant increase in legitimate turnout, with the additional votes being divided between the two major parties, but with a comparatively small overlay of fraud so as to oust Trump.
This presumes that all the different agents of the Democrat party knew how many they needed and where they needed them. I have no doubt that all the agents of the Democrat party were rigging votes in any state that allowed them to do it. Perhaps some of the higher level crooks knew exactly where they needed to concentrate their efforts, but the vast majority of the lower level crooks simply stuffed ballots into the system wherever they were.
In other words, the observable data (namely, the official counts) are perfectly consistent with either hypothesis. Therefore, looking at the totals doesn’t prove or disprove fraud.
I think the 7% increase in votes is itself proof of fraud.
Turnout is subject to many different factors besides fraud. Exact measurements of percentage turnout can vary, depending on definitions (e.g., trying to take account of U.S. citizens living abroad), so percentages given to the nearest tenth are pretty suspect. Keeping that in mind, the increase from 2000 to 2004 was about 5.9 percentage points, and from 2016 to 2020 it was about 6.5 percentage points.
Is every such increase, by itself, proof of fraud? If so, is every significant decrease proof that fraud had been reduced?
Turnout has been much higher and much lower than in 2020. In 1868, the first election after the Civil War, it was 80.9%. But in 1924 it was 48.9%. You can’t simply assume that the legitimate turnout is some fixed percentage from year to year. It varies.
Changes in the official totals certainly include changes in undetected fraud but also include changes in legitimate turnout. It’s obvious that all the changes adopted in several states in 2020 (such as no-excuse mail-in voting) would significantly increase legitimate turnout.
Your position is “I think the 7% increase in votes is itself proof of fraud.” History doesn’t justify that inference. If you think it does, we probably have nothing more to discuss.