Best Doujin Contest Discussion

Originally Posted by Djinova

Does one track advance or two?

The two leading tracks advance.

At this point, I've grown frustrated with the muffled sound. There's already enough ambient noise going on around my room, I shouldn't have to strain to hear the tracks even at near full volume on my speakers. Headphones aren't helping much either. I want to hear the layers, not just get the gist of the melody.

Is there any possible way to get access to the tracks without them being mp3gained?

A possible problem with voting for one track, but two advances is genre bias. For example you have one good rock track, one really mega awesome one and two mediocre soft tracks. Then it could be that one of the two mediocre tracks advances, because people are all voting for the mega awesome one, except a few that don't like rock and votes for the soft tracks. The good one wouldn't get votes because, there's one that's better, something that wouldn't happen if voting for two tracks is allowed.
I'm just saying, in the end the best wins, so does it really matter whether the good track gets disqualified here or a round later?

Originally Posted by Dr. Uzuki

Like hell I need to re-examine anything. It's all basically a copy paste job to slap up the threads which is not a lot of effort, but it still ends up being time consuming and monotonous with just the way it is, and, hey, let's double the amount of it you have to do for a minute perk. My opinion would be it's not worth it, but maybe Orion's opinion is different. Not really sure why you're directing this towards me, I am posting none of these threads.

Because I'm the one who expended effort to code it in the first place. If you don't use it, fine, but it's kind of insulting when you go around telling other organizers that it's a pointless feature. Should be obvious enough, no? (And let's not even get into "minute perk")

I mean, you of all people should know what it's like to have your work dismissed.

Originally Posted by Cyrus

The two leading tracks advance.

Oh, OK. That's what I originally thought it was going to be, but since we were only picking one, I wasn't sure. As Djinova said, I find picking the 2 that you most like out of the 4 to be better.

I'm sure some will groan at this thought, but I move for restarting the contest with tracks that haven't been MP3Gained (MP3Gain has an option to remove gain changes, rather than everyone having to resend the files) and with everyone choosing the TWO tracks they most like out of the four in each pairing. I feel that would be better for the contest and the participants in the long run than just continuing with the way things are now.

The four-track brackets for round one were a feature people suggested early on, and it's the first time they've been used outside of prelims, which were the model here (4 tracks, 1 choice). It seemed to work well enough there; we'll be back to two-song pairings in the next round (and they'll be shuffled, too).

There is no undo function for the OS X version of mp3gain, so the tracks can only be re-gained at this point. I don't have the best ears, admittedly, but listening to my own nominations I don't hear any distortion or muffling.

It's not at all a pointless feature, it just works more practically on small scales. These events go into the ridiculous status as far as sheer volume of music goes. It's worked great for SotW and as a general tool for journals and posting. I'm just comparing the fact that that's one post a week (in SotW's case) vs. sixteen back to back including poll making (32 in the more standard formatting of first round pairs instead of groups). Two nights later, it's done all over again.

Originally Posted by Djinova

A possible problem with voting for one track, but two advances is genre bias. For example you have one good rock track, one really mega awesome one and two mediocre soft tracks. Then it could be that one of the two mediocre tracks advances, because people are all voting for the mega awesome one, except a few that don't like rock and votes for the soft tracks. The good one wouldn't get votes because, there's one that's better, something that wouldn't happen if voting for two tracks is allowed.

Yeah, that´s exactly what I was thinking.

I was iffy on the four track first round system, but that was before I knew two tracks advanced, now I think it will work out nicely. Plucking a pair out of a group instead of a one or the other type deal seems like a much better method of weeding out the fodder. Worst thing about 1st round has always been the random pairings of two tracks that could seemingly go deep in the contest. There's a lot less of a chance grouping four completely crap songs (or great ones) as there is of only two.

Plus, it could make for some epic rematches in later rounds. That'd be cool.

Originally Posted by Dr. Uzuki

I was iffy on the four track first round system, but that was before I knew two tracks advanced, now I think it will work out nicely. Plucking a pair out of a group instead of a one or the other type deal seems like a much better method of weeding out the fodder. Worst thing about 1st round has always been the random pairings of two tracks that could seemingly go deep in the contest. There's a lot less of a chance grouping four completely crap songs (or great ones) as there is of only two.

Plus, it could make for some epic rematches in later rounds. That'd be cool.

You can extend this system even more, by pairing 8 tracks together and let 4 advance. As SotW has shown it shouldn't be a problem to keep an overview. It would even rule out more weak tracks as a matter of statistics. 16 tracks + 8 advance might be overkill though, but why not?
Another possible advantage is that it can make the decision easier in case of tough pairings. Take an 8 pairings system with the following quality check: 9 9 7 7 6 5 3 3. If statistics doesn't like you, you'd have a 9vs9, a 7vs7, a 6vs5 and 3vs3 pairing to solve. The likeliness is not high, but with a multiple vs grouping it's still lower. If it was to me I'd make all inner pairings multiple.

Btw, how much time do we have for voting? I would have to compromise if it's only till Friday (for A).

Well, maybe for next time since I think it's too late to change anything now, the 1vote-2advance has a few possible flaws like genre bias, secondary voting, late voting advantage, something that is non-existant in a 2vote-2advance system, actually any x-advance system should be accompagnied by x-vote system, in my opinion.

Originally Posted by Dr. Uzuki

I'm just comparing the fact that that's one post a week (in SotW's case) vs. sixteen back to back including poll making (32 in the more standard formatting of first round pairs instead of groups). Two nights later, it's done all over again.

Ironic considering that your own calendar from the BSC was the inspiration

If people feel like the voting period needs to be longer (A and B end late Friday) there's no reason we couldn't extend the polls, though that would be a headache for the moderator that'd have to do it.

How long do you think is ideal for a voting round, if not a week?

... what . . . what just happened?

The poll for F01 is entirely wrong, and there's like, no spaces between words in song titles.

Originally Posted by wvlfpvp

... what . . . what just happened?

The poll for F01 is entirely wrong, and there's like, no spaces between words in song titles.

The lack of spaces is because the titles of four songs have to fit in the subject line, I think. The poll itself...well, yeah, that needs to be fixed with the correct track names.

Originally Posted by wvlfpvp

... what . . . what just happened?

The poll for F01 is entirely wrong, and there's like, no spaces between words in song titles.

Yeah, I'm sorry - Yesterday's database errors made me quite impatient and I messed up the poll.

The lack of spaces between the words in the song titles is due to the subject line limit - I could have used abbreviations.. but I guess that wouldn't be good either

Ah and .. yeah..

WHERE IS BRACKET H?! *requests Bracket H for the second time*

EDIT: Okay, Bracket H is still missing.. guess I should ask Miles by myself then.
I'm going to post all the H pairings in the G subforum for the time being ~

When H is up, be sure to message me and as soon as I see it, I'll move the threads. I'm sure Crash'd be perfectly willing to do it to.

Gosh. I'm so sorry. This week has been hectic for me. Bracket H exists now. -_-

Oh, it's very ok Miles. Thanks

Additional Spam:
Oh, and also btw:

Quote:

You can extend this system even more, by pairing 8 tracks together and let 4 advance. As SotW has shown it shouldn't be a problem to keep an overview. It would even rule out more weak tracks as a matter of statistics. 16 tracks + 8 advance might be overkill though, but why not?
Another possible advantage is that it can make the decision easier in case of tough pairings. Take an 8 pairings system with the following quality check: 9 9 7 7 6 5 3 3. If statistics doesn't like you, you'd have a 9vs9, a 7vs7, a 6vs5 and 3vs3 pairing to solve. The likeliness is not high, but with a multiple vs grouping it's still lower. If it was to me I'd make all inner pairings multiple.

Yeah, I do feel that's going overboard personally. Choosing one out of four, if the groupings follow a doubling pattern, is definitely what I'd assume to be the reasonable limit. 1 out of 8 could be torture. Not wholly comparable to SotW where you receive six points to spread and some weeks that was a hard enough decision to come to. Not to mention could only work if voter turn out is huge. That was something I was worried about seeing 4 per group, but it's been a solid turn out of around low 30s or so, I don't think we'll see boatloads of ties like I thought could happen. Or if they do, it'll be mostly between the leads it looks like, making them inconsequential.

Is there any way to do blind polls? I'm just not feeling public polls more and more as I read the late-stage votes.

Not that I'm dying here, but it's just annoying to read "Well, Track X already has a lot of support, so I'll vote for Track Y." IMO, that shouldn't be what the contest is about and only rewards people who hold off on their votes until the last minute. At that point, we're talking about the difference between various tracks advancing or not given how tight some of the numbers are. Even if something hasn't had any support, I've been voting for my personal pick of each batch, not doing pity picks.

At least a blind poll would work without those types of politics factoring in so much. Plus people could still post their thoughts, since they aren't indicative of any trends. Can anything be done, for the next round onward? If not, again, I'll live.