I, too, am for keeping the clones, but not ALL clones. The ones that are identical shouldn't be included. In that case, I think only the "main" version should count. A good criterium for determining if two clones are identical is by creating an inp on one and seeing if it plays back correctly on the other. Yes, I know that MAME stores the name of the romset in the inp, but that's easy enough to change when testing if it plays back on another clone. For instance, although I haven't tested any of this yet (and forgive me for bringing Crazy Kong up again, but it's just about the only game I play), I am pretty sure that the inps created by ckongalc and monkeyd (I'd call this version "Crazy Kong") are interchangeable, and I am pretty sure the same holds for ckong, ckonga, ckongjeu, and ckongo (I'd call this version "Crazy Kong II"). ckongs, dkong, and dkongjp are all distinct versions. This would at least cut the number of CK/DK variants down from 9 to 5.(posted 9595 days ago)As for minimum scores/points: Perhaps some script could be written that is run once a week or so and that purges the system from submissions with too few points. Obviously, submissions that have 0 points can always be removed, and as long as a submission doesn't take a top 3 spot, points that are below a certain number (5 or 10 sounds good to me too), should be removed then too. Sure, people may be very proud of their achievements, but, as said before, nobody really will download those, so the proud people should just keep the recordings for themselves (hey, maybe even put those on their own websites).
I also agree that games for which a maximum can relatively easily be reached (so that would not include Pacman) should be excluded. I have actually changed my scores for those games to 0 on the leaderboard. Now if only Zwaxy would simply delete the files...
Pattern games should be kept in. To some extent at least, lots and lots of games are pattern games. Most people have learnt those patterns the hard way (using trial and error while spending lots of time, money, effort, ...), and even if they hadn't, then still it requires skill to master the patterns. Pacman has been discussed so often and for so long, and patterns that work are common knowledge. But it still requires skill, and I think the fact that a perfect score on Pacman is so rarely achieved is proof of that. The same goes for those other games. If Steve Krogman scores over 20M points on Galaga, and his inps are readily available, then why aren't there lots and lots more of multi-million point Galaga submissions? Because it still requires skill.
I think that limiting the number of submissions per individual per day is not going to work. People can too easily just save the surplus for a day on which they will not reach the maximum number...
In addition to the point limit, there could also be a position limit. Keep only the top 10 scores or something for each game. Let's face it, nobody is going to download the number 17 recording. BUT... I have sometimes downloaded a recording that wasn't one of the top scores because the recordings for the top scores wouldn't play back on my version of MAME. I know that on my .35 version of MAME32, I can't play back any inps created on .34 or earlier versions (or even some of the .35 beta versions). And downloading an earlier version of MAME doesn't always work either, since lots of the roms have been renamed, and it's extremely hard to find older romsets. I don't know exactly what the solution is... as long as the emulation itself hasn't changed, a conversion tool (that converts older inps to a format that can be played back on newer versions of MAME) can easily be written. I may actually do that because there are some recordings created on earlier versions of MAME that I'd like to see and can't because I don't have those earlier versions and because my romsets don't work with those earlier versions. And I'm not even talking about merged romsets.
Just my thoughts. Cheers, Ben Jos.