The best positive way to deal with chess cheaters is by a strong word of mouth among fellow TDs. Unfortunately, human gossip always leads to a host of negatives which can be accomplished so easily, not the least being false accusations, so this method can’t ever be solidly recommended.
I do not recommend the following proposal, but I think it would help curb cheating. The dangers I anticipate are still a few false accusations (deliberate or not), an increased target on the USCF computers, and an increased climate in mistrust. There is also probably significant infrastructure cost. In my opinion the dangers/costs are not worth the effort that it would take now, but maybe this is a path we might want to visit in the future.
Have in the TD area a so called cheating form, only accessible to Senior-level TDs and above. Lower level TDs and non-TDs can try to convince a higher level TD to lodge a report that they can credibly substantiate. The form would be basically TD info, player info, case # report, and an urgent flag. No one can ever see anyone else’s submissions except the USCF Technical Director and maybe 1-2 others, if the Director feels there is a need. Basically, if a player is marked with an urgent flag (guidelines can be established as to what is urgent) or accrues, say, 3 reports, the system automatically notifies the Technical Director who looks into the merit of the report(s). If the Director thinks serious, credible reports of cheating have been lodged, he can notify the player, the submitting TD(s), and any witnesses for an internet/phone chat/email log to give the player due process to explain himself or submit people who will testify to the player’s good character. Then the Technical Director can choose to either exonerate, monitor the situation further, or find that the corroborated reports are meritorious, in which case he applies penalties including adding the name to a cheater’s list and submitting written notification to the player’s state president.
Laborious? Yes. Risky? Yes. Costly? Probably. Good enough? Probably not. Alternatives? Do nothing, which unfortunately is maybe best by test.