You know, since I'm a lazy faggot who'd never actually do anything about this I might as well post what I had in mind. Sorry for derailing your thread OP, if this will actually spark a longer discussion I'll simply create a separate thread and we'll move there.
The basic idea is obvious by now. I figured most people are more likely to pick up new ongoing shows than work on their backlog. If you think about it it makes absolutely no sense whatsoever. It's very unlikely that 10 somewhat-randomly selected shows from current season will prove to be any better than 10 old shows which you can pick from over 50 years worth of anime.
So yeah, first we'd select 15-20 old shows from various genres. 1-2 mecha, drama, romance, mahou shoujo, action, sports, mystery, psychological, whatever. The point is to have a lot of diversity so that everybody will find something for themselves. As to what exactly consitutes 'old' I have no idea. I guess older than 5 years or so.
We could create a pool where everybody could suggest as many shows as they like. Some shows would inevitably end up popping up numerous times and those would take priority. If we'd end up with 59 shows with 1 vote each there are some criteria which one could use to select one (which I won't name for the time being because that could cause some controversy and it's a pretty irrelevant minor detail) or you could simply pick one at random.
As for what is eligible (aside from the fact that new shows obviously aren't) let's say it has to be something outside of MAL's (or anidb's if you can sort shows by popularity there, the site doesn't matter much) Top 500 most popular shows (or Top 400/300 if 500 is too harsh). The point is it'd be better if the shows in question are something most (?) people haven't seen yet. Obviously it'll be really hard to pick something nobody has seen but since it's nigh on impossible to begin with there's no use worrying about that.
Let's say we have already selected the shows by now. Somebody should make a chart at this point to make it easier to figure out what's what. A chart like that is not something that would magically appear out of thin air but making it would take way, way less work than making a normal seasonal chart. You already have all the info you need. You know the name of the show and it'll take like a minute to look up the genres, studio and grab a short summary from somewhere. You could easily get it done within an hour provided you have a decent template to work with.
Just about the last thing that needs to be decided are the dates when the episodes 'air'. I think it'd be best to speed up things a little bit and make season one month long instead of three. So you'd end up with an ep of one show roughly every 3 days (since you have 31 days to work with at very best and one cour can have up to 13 eps). If somebody has watched ahead - that's fine. If they have already seen the show before the chart was even made - no problem. As long as they can abide by the rule that they shouldn't discuss events that will occur in eps which haven't officially 'aired' yet (thus spoiling them and in some case making the discussion obsolete, for example in mystery show where people are actually trying to figure out what's what).
Since the dates will be set somebody could even >stream the eps but I don't know if that's a good idea, I think it would kill the discussion on the imageboard in favor of low quality spam during the stream.
The thing that really disencourages me from even attempting to make something out of this idea is that /tc/'s current userbase is simply too small. You'd have to have about ~50 people who are genuinely interested in this (and who would actually go through the effort of posting in the threads every now and then) to actually make it worthwhile.