Is it me - or is Carlos the only one whose voting pattern looks legit?
Not you at all, that's the conclusion everyone here finally realized several days ago. As Lord EEVblog said himself above, you'd have to be crazy to think these amount of votes are legitimate. I have seen Carlos absolutely plaster the entire film industry, on some lists and groups he has threads with hundreds of replies and engagements. He can't even touch their votes-on-demand.
Your rate chart there proves it. With Carlos it's clearly shown that when you have hundreds of thousands of views of your contest video out there asking for votes (which he does) there cannot be any sudden changes in vote rate. The voting/viewing sample pool is too large to be jerked in either direction, you can't coordinate 1000 strangers to start or stop voting.
Compare it to nilus and davids graphs - laughably all over the place, going from growth indicating 900,000 ad impressions one hour to indication of 0 in 30 minutes. It's not mathematically possible without an external mass dump of votes on demand.
If Keysight does anything other than disqualify both of them, especially after seeing that, I'm going to be extremely disappointed in them as a company. Even if it meant the scope didn't go to Carlos but instead some random guy who didn't need it at all, that is still worlds better than rewarding blatant fraud