I’ll likely be one of the concrete examples that could be given, and don’t mind being “out there” with the discussion. One of the challenges that OGS developers will have with this is discerning who really should be regarded as “above the limit”. When I joined the tournament that SuzieKyu and I am currently playing in, my overall rating was indeed at or below the 10kyu limit. This is not the problem we need to be focused on - the limiter works very well as I have found out when trying to join tournaments and not realizing I was at the moment above the cutoff. One thing I think that could be added though would be a check to see if a person’s rating (which rating to employ is my main concern though as you’ll see below) is above the cutoff when the tournament starts. Often tournaments are set up months in advance, and people can easily improve (or fluctuate upward in the absence of real improvements of the magnitude that seems to be) in that timeframe. Adding such a “tournament start checker” would be possible, I should think, but I don’t think is the main problem.
Because of some recent games, I am technically I am above the DDK cutoff of 10kyu because OGS looks at the overall rating - not the live or correspondence ratings - to determine who should be in or out. One of the problems with this is that bullet or live ratings skew the overall rating because of the method of calculation. I play a fair bit of rapid/blitz/bullet go because I can work on quick tactics that way and can squeeze it in during lunch breaks at work. However, since this is a large fraction of my games played overall, my rating is heavily skewed toward my blitz rating, which is about 7k. My live rating and correspondence ratings are well within the DDK range, and so to keep me (for instance) out of live and correspondence tournaments having DDK limits wouldn’t be appropriate.
The only fix that I can see (and I would strongly advocate for) is that tournaments should use the appropriate rating (and not the overall rating) - so correspondence tournaments should look only at the correspondence rating, and live tournaments the live rating. To do otherwise (as is currently done) seems to cause all kinds of issues, as SuzieKyu demonstrates with the original question, and to which I hope I have added some important qualifications.