LogFAQs > #952936277

Coming in early with the stats this year! After 36 hours, I think we have a good number of participants accounted for in the spreadsheet, so let's have a look at the support votes in play!

We currently have 66 nominators accounted for in the spreadsheet - as I wrote that, it changed to 67 but I don't want to redo the numbers. Each person gets an effective 35 votes to allocate, but how many will a song need to make the bracket? The lock threshold is currently 6 votes, but will the final cutoff be the same? Higher? Lower? If you do the easy math, we have (66 people)*(35 votes)/(256 slots) ~= 9...hey, what gives?! There's no way I'm gonna get 8 other people to hop onto my song!

The reality is that a significant portion of votes always end up unallocated to bracket songs. If the safe cutoff is 5/5, there are always songs holding votes at 5/4, 5/3, 4/4, and even 1/1! There are lots of reasons for this:
- People just don't have time to keep up with updates in real-time at the end. This is totally reasonable, and to you I recommend going for the lock deadline instead!
- People are just too stubborn to let their beloved songs go as the end times approach. Why are you still sitting on your 2/2 when the cutoff is at 4/3 and rising?! Your emotions make you weak!
- People do keep up with updates and they do shift towards a song with momentum, but...it just barely misses because the cutoff jumped at the last minute or they couldn't get that last support. These are the ones that hurt the most, the 5/4's and 4/4's gazing up at the 5/5 cutoff.

So, if 100% of votes don't end up allocated to the bracket, what's a better estimate? In VGMC 13, it was around 56%, with (69 nominators)*(25 votes) and a 4/3 cutoff. In VGMC 14, the locking system was added with a threshold of 5, and efficiency improved slightly - 60% from (66 nominators)*(25 votes) and a 5/3 cutoff.

What? Efficiently allocated votes were around the same, but the cutoff went from 4/3 to 5/3?! The answer is that the cutoff can move as high as you want if people coordinate - remember that we can reach a theoretical cutoff of 9 votes this year. More answers lie in the proportion of clearly dead votes that people didn't reallocate - in year 13, a whole 18% of votes ended on songs at 2/2 or below. In year 14, this was only 12%. Kill your darlings, people!

For this year, I'll use VGMC 14 as reference since it also had locks. We had 66 nominators both last year and this year (at the time of writing), so the math works out nicely! Some factors remain; new people still have time to join, and locking at 6 is probably harder to coordinate than locking at 5, so let's take VGMC 14's 60% efficiency and consider 55% and 65%.

If we take (66 nominators)*(35 votes)*(55% efficiency)/(256 slots), we get 4.96. That's a great deal compared to the locking price of 6! This would happen if people are very stubborn or otherwise do not actively shift their votes away from probably-dead songs.

If we try (66 nominators)*(35 votes)*(65% efficiency)/(256 slots), we still get 5.87. So it looks like there's a small premium paid to get songs locked at 6, but I take no responsibility for that claim! If there are still participants yet to join and they're more willing to compromise than previous years, we could easily end with a cutoff above 6/3.

But I don't care about the final cutoff! I'm just worried about how many songs will actually lock!
Okay, let's also look at how eagerly people went for locks last year. Exactly 40% of votes were used on that year's threshold of lock-at-5, so if we apply that to this year, (66 nominators)*(35 votes)*(40% allocation)/(6 to lock) is 154 slots locked. Honestly, if I tweak the 40% number things don't change much, so people would have to be really enthusiastic about locking to lock the entire field of 256. It's not just about coordination, but making compromises and concessions - if you think of your initial supports as scoring 9/10 for you, you would be shifting your votes to songs that are only 7/10 or 8/10 to help them lock. But hey, why should you have to be the one compromising here?! If you stay on your 9/10's and wait, that'll pressure other people to give up their 9/10's and join yours instead! Yes, yes, the perfect plan...and then no one budges from their 9/10's because everyone's thinking the same thing.

Hmm, how about adjusting the lock threshold? With 40% allocation, a lower threshold of 5 gets you 185 locks, and even lower at 4 still gets you 231. A higher threshold of 7 takes you down to 132. There looks to be plenty of wiggle room, so I wouldn't worry too much - we do have more people (supposedly) and +10 votes this year, but we're also adding 64 bracket slots that need 4-6 votes each to be claimed, which offsets a lot of the pressure.

That's all I have to share for now! Might give an update closer to the lock deadline if we see even more people joining or I notice a trend in vote coordination. Toodles~
---

Sometimes you just need a separate account for statsposting.
VGMC stuff: https://p0p0p0p.github.io