DDR BRACKET 1 RESULTS

[will-i-am]

Full Judging Results - DDR Bracket 1 Spreadsheet [link - edited version]

Thanks to all who submitted reviews, and to the rest of the reviewers for submitting well-written critiques. This competition wouldn't have run so smoothly without you guys.

Without further ado, here are the Semi-Final qualifiers for the BEST-mix FINAL DDR Division Bracket 1 (by bracket):

Bracket 1
Tweet Tweet
Teknopathetic
Fascination eXstacy

Bracket 2
LOVE IN THE SHADOWS
SPARKS FLY
FIREFLIES

Bracket 3
Russian Unicorn
I'll be there for you Zero no tsukaima F
Bubble Pop

Bracket 4
Space Truckin'
Vertex (Stage 1)
Trial Machine PheoniX WrighT

BLueSS EDIT:

Bracket 1 Results are being amended. For details on why, please see this post.

Bracket 1 will be rescored without hellrazor's reviews.
All of the files that originally moved on will still move on.
Any files that would have moved on will ALSO move on.
There will still be a wildcard.

Songs additionally moving on to the next round include:
Totalimmortal - DDR Bracket 3 (bumped up to 2nd place)
I'll be there for you Zero no tsukaima F - Will still move on

Gangnam Style - DDR Bracket 4 (also bumped up to 2nd place)
Trial Machine PheoniX WrighT - Will still move on

Congratulations to all files that advanced! Kudos to those who just missed out. And thanks to all who bombed for trying (mostly)! If you don't see a file here that you thought should have advanced, but didn't vote because you were too busy with "life", blame yourself. It's ALL YOUR FAULT.

[Hainaut]

Nooooo! Gangnam Style didn't move on? :(

[Zounder]

I already sent will a PM, but my math has Gangnam Style moving on and I just rechecked it, unless the scoring system works differently that I have it.

[Juub005]

Ugh, I forgot to reset my timezone on this site, so I thought the deadline for reviews was two hours later than it actually was. FAIL.

Edit: Although my review didn't count for B1-2, the results would have been the same. Phew.
Also, missing next round by 2 points: :(

[k//eternal]

for transparency do you mind showing the score totals (like, the excel sheet or whatever you used)? i know all the information is out there, but you've already added it all up so why reinvent the wheel :)

[will-i-am]

You have brought up a strong point, my good sir, but the way you calculated the results put tremendous favor in jammitch!'s review because there was only 1 partial review for sub-bracket 4. jammitch! actually has more say than any of the complete reviewers by your system.

I will quote your math in case anyone else is wondering the same thing. I hope you don't mind. This is evidence for my case.

Complete Reviews:
GS 25 21 12 AVG = 19.333
TM 23 20 19 AVG = 20.667

Partial:
GS 15
TM 11

Approval:
GS 3 out of 10, so 15/10 = 1.5
TM 4 out of 10, so 20/10 = 2

GS: 19.333 + 15 + 1.5 = 35.833
TM: 20.667 + 11 + 2 = 33.667

Each approval voter apparently has 0.5 points per vote, which is ridiculously low. The reason why I did not average each review/vote category was because this imbalanced of power needed to be avoided.

It's not the way people vote that matters. It's the number of people that vote. AKA democracy.

Case closed.

And yes, k//eternal, I'll edit the first post to make that clear. :)

[Zounder]

How was I supposed to calculate it?

The other option is that the scores just get added and then divded by the total, but that wouldn't make sense, since dividing at that point is pointless. The rules + guidelines post says the scores get divided, so I assumed each one got divided separately, because otherwise, why divide at the end? It doesn't change a thing.

[k//eternal]

i think it's just divided to keep the final number from being goofily large. i thought that was kind of funny too, but the rules seemed pretty explicit about what's intended so i didn't question it

i'm curious about the other files' scores too, since i was kind of only following our bracket in terms of what was winning

[will-i-am]

I didn't divide at the end. The system is actually flawed if you average it out. Sorry guys!

I know that wasn't stated earlier, but you can clearly see how averaging all complete, partial, and approval votes in their categories and then adding the averages all together would be an imbalanced system. Because most people who reviewed submitted complete reviews, surprisingly enough, less power was put in their hands when only one or two partial reviewers submitted a review. I couldn't average the categories for reviewing because it's unfair to put more weight on jammitch!'s review rather than either #Infinity, mine, or hellrazor's.

At the moment, BLueSS and I are discussing the system alteration. These results are tentative for the time being. I apologize for the inconvenience.

[Zounder]

Regardless of which of those methods are true, approval voting seems to hold more influence than it should. Your deviation is guaranteed to be 5 with approval voting, whereas if someone does a complete vote and all their scores are between 20 and 25, their vote actually counts less.

I don't really know where I'm going with that thought since it's not really pertaining to the topic, but yeah.

EDIT: I do understand the flaws of averaging each separately, I just thought that's the way it was, or that you guys anticipated more people doing partial/complete reviews so that we didn't end up with a jammitch! situation.

[BLueSS]

The scoring algorithm wasn't the final version of what I intended, but due to miscommunication will-i-am didn't see that.

I am reviewing the results now, as I did not have time to do so earlier. As will-i-am stated, these results are tentative.

[Valex]

haha Space Truckin got it in. <3 Telperion

Also woot my file actually moved on YEAH

[Spork!]

Well, I guess I predicted the immediate failure of FeXt incorrectly xD
Maybe someone will play the doubles on it eventually. It has silly 192ths :P

Good job everyone, can't wait to see how everything does in B2.

[will-i-am]

Regardless of which of those methods are true, approval voting seems to hold more influence than it should. Your deviation is guaranteed to be 5 with approval voting, whereas if someone does a complete vote and all their scores are between 20 and 25, their vote actually counts less.

Right. That is the flaw at hand with the overall system. Because people are allowed to only vote for 1 file by approval voting, they can bump one file 5 points ahead of every other file just like that, without giving any bonus points to files they thought should place 2nd and 3rd.

Do remember that sub-bracket 4 was really tough. I mean, look at Space Truckin'. A perfect 10/10 in approval votes. That's 100 points straight from that! Conversely, Space Truckin' gained a whopping 74.75 points in the complete reviews from just 3 people. That's a lot of power given to those guys versus the people who simply submitted a vote for their favorite songs without considering all the technicalities.

At the moment, the current question is: How many points do approval voters really deserve to award? Voting for just one file is a huge set-back to every other file in the sub-bracket, but that's the whole point of voting for just one file. You vote for just one because you really want it to move on, so we have to respect that. Same goes for public reviews. If you really like a file, you'll give it more points, but like you mentioned, reviews rarely set apart overall files by more than 10 points.

Also, Valex, if you head over to DDREvolved you'd be informed that Telperion is actually a consistently phenomenal DDR step artist. He won the Simfile Olympics, after all.

[Juub005]

With approval voting you might consider holding a re-vote.

The main problem, in my view, is that the approval voting is not ranked. If you implement a ranking system, you remove the all-or-nothing emphasis that causes the problem in deviation. (I want to do some work on the math necessary for such a revision in the scoring system but I need to get something to eat, so I'll just post this first.)

In regards to the unanticipated high number of approval voters, well, to a certain extent it is inevitable that the amount of effort and/or time people put into the voting will reflect in the scores.