From time-to-time certain maps get rolled that end up with fewer city-states than is standard for that map size (should be two city-states for each AI on the map). I think there are others, but for sure a Lakes map has this problem. It has been theorized that, because of the higher frequency of mountain tiles on the affected maps, a CS settler might spawn on a mountain, then not be able to settle its city. Nobody's sure what the underlying technical problem is, but when one finally discovers the entire map, "Dang, there are only 15 city-states on this standard map!".
As the HOF rules require the standard number of city-states, these anomaly games get red-flagged by the parser software, and thus have to be rejected. However, previous discussions about this revealed that some games do not get red-flagged, even though it had too few city-states. I cannot tell you why or how, but there must be a couple of manifestations of this issue, one that DOES trigger the red flag, and another that does NOT. In the case of "does not", it must be that the parser can tell that the right city-state count was created, even though they all didn't settle. (I am reasonably certain that the parser always knows if a player has adjusted the city-state count prior to rolling a map.)
While correcting the code is not likely to happen (I presume it would take the game developers to do that), in my opinion players should submit any of these games they choose to finish. If the parser sees that there was indeed the correct count at the start, (i.e. no red flags), then that game could be accepted. If the parser does red-flag the game, it obviously won't be accepted. I will check with Noble Zarkon, but I think those red-flagged submissions could be deleted rather than rejected.