One Synth Challenge #159: Surge XT (j5v Wins!)

[Andreya_Autumn]

I didn't put any thought into statistical analysis of voting, and I don't really mean to, unless someone can show me a good reason to.

I work with music and am fortunate to hear others music all the time, but it's somewhat rare these days that I feel like I'll want to listen to something again. Like, not many songs/artists go into those playlists that I listen to for my own pleasure.

Anything that I feel like could go in such a playlist is going to get at least a 4 from me, I'll give a 5 if I feel like it's truly well developed and finished. That's hard to do with only one synth even if that synth is Surge XT, but there were a few who impressed me!

I'm giving a 3 to anything that sounds good but not fully convincing somehow, and a 2 to anything that's just not really a finished song. I'd rather not give a 1 to anyone that I think actually made an effort, which everyone did. If someone submitted five seconds of a sine wave or something I guess I'd give it a 1.

[emikaela]

i am also done voting. trying to add a few more comments still

[liqih]

votes placed as required + comments as time went

[Leonard Bowman]

I finished voting; as usual I voted high and compressed to a narrow range because I don't need my last-minute listening to have a big impact on the final results.
Really enjoyed listening to these tunes!

[Rahodees]

Voted

[diaseis]

When sbdy uses a baseline of 1 or 2 for example - then the assumption that she looks for a comparative advantage might hold, but it is still only an assumption about someones motivation to choose a baseline ... because while choosing the base line the person does not know the final average at all...

I don't really understand why you're trying to imply I have no credibility since those sentence show that you understand exactly what I was trying to show. Now of course in OSC we do have a good idea of what the mean will be, since it's always pretty much the same. The rest of your message concerns assumption I didn't make.

Anyway, again, I see this as a fun math result, not something to be taken seriously in the context of OSC.

[diaseis]

Just think about it. You do a 2 people OSC. What do you want to do (if you want to win and not acknowledge the very high quality track of the opponant participant) ?

[diaseis]

Now you normalize (divide notes by the sum of the notes you put), you'll get a note of 1.0 for each participant, so with this technique, they will inevitably end up at equality, which makes perfect sense since this voting system couldn't work for 2 persons. By normalizing, the only difference it makes is that only the variation between the rating of one voter relevant, not the absolute amount of points.


[EDIT : I feel the need to precise again that this is not a request]

[diaseis]

oops (duplicated post)

[phankiejankie303]

Voted!

@ Yadrichik: All the best to you and your mother. Best wishes for a quick recovery

[phankiejankie303]

Regarding the voting system... Well I think some kind of a limited pool of points should be introduced but I guess that will require too much work from the people running the show.

For instance let's say you have like 10 participants then you would have like (5*10) = 50 maximum points and then you need to cut some out so you force spreading. So you say that there should be just a tad more than half top scores so you multiply that by 60% = 0.6 so you have like 30 points to spare on 10 participants. That should make you extra careful where you spend your points and you are also forced to spread them from 1 to 5 so if you are very generous to some then you will have to be very ungenerous to others.

I think a percentage like from 0.6 to 0.7 over the number that emerges from multiplying the total participants by 5 is the best. 0.6 a bit more strict 0.7 is a bit more generous.

So my formula for the Pool of Points should be (TotalParticipants*5)*(0.6 or 0.7).

Hehe my two cents.

PS: Another thing I would introduce to future OSCs is a limit to the synth instances. It's pretty odd to not allow modulation effects like Chorus but someone (provided he has the CPU power) can stack 20 instances of the plugin with slightly detuned SAWs or stack Sinewaves and reaching additive synthesis territory.

[Double Tap]

It is possible to affect the result, but the effect gets smaller as more votes are cast.

A gives 5 votes to E, and 1 vote to themselves.
E gives 4 votes to A, and 1 vote to themselves.

A gets 5, E gets 6. So E wins. Clearly there is an incentive for E to low-ball.

But as you expand the number of players, a low-marking player has less and less influence.

Let's assume A and E have clearly made the best tracks. A gives high scores but E gives low scores. The others give everyone at least a 3 and have the same average score, and the other players prefer A's track to E's by a ratio of 2:1.

A - 1A, 4B, 4C, 3D, 5E - 17
B - 4A, 1B, 3C, 3D, 5E - 16
C - 5A, 3B, 1C, 3D, 4E - 16
D - 5A, 3B, 3C, 1D, 4E - 16
E - 3A, 2B, 2C, 1D, 1E - 9

A - 18
B - 13
C - 13
D - 11
E - 19

E wins, with a very narrow margin. The variation of E's scores is identical to A's, but they win because they marked low.

But that's such an artificial situation that it would only take a slight change in the voting for A to (rightfully) win. And as the number of players grows, the number of 5s becomes much more important than an individual score. A player who has five more 5s than anyone else would still probably win even if a rival deliberately low-balled them.

So it doesn't really matter if a player is not generous when most other people are. It may gain or lose you one or two places in a tight contest, but nothing more than that.

[silverpants]

I didn't put any thought into statistical analysis of voting, and I don't really mean to, unless someone can show me a good reason to.

I work with music and am fortunate to hear others music all the time, but it's somewhat rare these days that I feel like I'll want to listen to something again. Like, not many songs/artists go into those playlists that I listen to for my own pleasure.

Anything that I feel like could go in such a playlist is going to get at least a 4 from me, I'll give a 5 if I feel like it's truly well developed and finished. That's hard to do with only one synth even if that synth is Surge XT, but there were a few who impressed me!

I'm giving a 3 to anything that sounds good but not fully convincing somehow, and a 2 to anything that's just not really a finished song. I'd rather not give a 1 to anyone that I think actually made an effort, which everyone did. If someone submitted five seconds of a sine wave or something I guess I'd give it a 1.

I put all my 5's into a personal playlist and I listen to them regularly at work on shuffle. It's great instrumental music and it makes me smile. So I suppose in my way of scoring - I only give 5 if I want to listen to it in future for pleasure. A 4 might be well produced/written but does not make the grade for my personal taste.

[Andreya_Autumn]

That also resonates with me. Realistically I'm not gonna listen much to this again. But again, that's only cause I'm listening to so much music all the time for work now. But this can and will change again, and all the fives and fours I gave are definitely for tunes that make me smile too.

It's really sweet how this challenge is designed. So much of the internet is full of folks who (understandably) want to be heard, but don't necessarily have the time and energy to listen to each other. The fact that we gotta make an effort to hear each other out in order to qualify is a brilliant counterweight to that. This months synths do not really inspire me at all, but I may just enter anyway because I like this culture. Big shout outs to those who designed the process!

[EvilDragon]

I wholeheartedly agree. Even as a simple bystander who never participated, I am full of amazement at how well-devised the rules of the competition are.