Back in the (early?) 80's the NY lottery ran ads telling people to stop playing the diagonal patterns starting at the top left and top right of the playslips. I don't remember the exact numbers, but I'm pretty sure it was 8 to 10 thousand plays for each of them. Playslips for MM and PB vary a bit from state to state, but extrapolating from the NY figures it would be possible that half a million or a million people would play diagonal patterns that are all based on less than 50 combinations, except that they now have the option of using quick plays. Based on the reports that 70 to 80% of people play quick picks I'll figure that maybe 100 to 300 thousand players might play those various diagonals. Add in all the other patterns fortune cookie numbers, and birthday numbers and there might be several million tickets that are only responsible for 200,000 or so combinations.
I'm sure the lotteries track that sort of info, but AFAIK it's unavailable. Many years ago I stumbled across an ad for some kind of lottery players guide that supposedly had a lot of information direct from the lotteries about frequently played combinations. I'm pretty sure it admitted that the data was a bit out of date because the lotteries were no longer offering the information. I've tried looking for it again, but can't scare it up.
With a true random distribution I can know the exact probability of a winner for any given number of tickets. I could make a SWAG and figure that if 50 million tickets are sold the probabilities will be what we'd expect for 48 million random tickets, but it really doesn't matter. It's always possible that we'll have a winner when sales are very small, and it's always possible that we'll get a rollover when an enormous number of tickets are sold. So far there's never been a drawing when we didn't find out which way it went within 2 to 16ish(?) hours after the drawing.