Earlier quoted context omitted.
> The thing is inactive voters are just that, inactive. There is very small risk that some of their names might get used for voter fraud. But every time this is researched, only a handful of incidents are discovered and it takes thousands to swing an election. You claim the risk is small, but I'm not aware of anything that actually supports this assessment. This form of manipulation is nearly impossible to detect in…
> [Voter fraud] is nearly impossible to detect in a meaningful way. If the "victims" of voter fraud truly cared enough, they could audit the voter records for any election. Unless I am wrong, there exists a record of whether or not each person on the rolls actually voted. The intersection of the people who voted with those that are deemed truly inactive would be the extend of the voter fraud (at best). No one seems t…
To be clear -- what do you mean by "truly inactive"? Usually "truly inactive" means they haven't voted. So people who voted are by default not "truly inactive". You would have to take some sort of heuristic estimate of how likely it is that a voter is inactive, and cross that with the voting record, and then attempt to individually contact the voter to verify that they did vote. The quality of such a heuristic would fundamentally affect the results. In other words, the experiment would either confirm widespread voter fraud, or, way more likely, provide no evidence for or against voter fraud.