| GameFAQs Contests
|Started:||November 5, 2015|
|Ended:||December 15, 2015|
SMG2/ Mass Effect, Xenoblade/DKC2, SMB3/Pokemon,
Starcraft/ME2, Super Metroid/Majora's Mask, CT/Melee,
Super Mario RPG/MGS3, All Undertale matches
Fall 2015 Contest was the third Best. Game. Ever. tournament held on GameFAQs. The contest was first announced in spring by SBAllen after cancelling the Worst. Character. Ever. contest that supposed to run in the summer that same year.
Much to the delight of Board 8, there would be no divisions by era this time around, and all matches were 1v1 and lasted for 24 hours. For the first time, multiple matches were held at the same time. Four matches were held a day in Rounds 1 and 2, two in rounds 3-5, and finally 1 match a day for the semi finals and final match.
So two years after the fallout of Draven made us seriously question if we'd ever get contests again, the announcement came that we were getting Best Game Ever 3. The overwhelming response to Allen's topic was to get a fair 128 game 1v1 bracket, with no divisions by era and no BS.
The following days saw Allen make a bunch of math topics to spoil the format, which would end up being... a fair 128 game 1v1 contest! I can't really overstate how cool this was. Contest fans have wanted this format for 10+ years, and it really looked like we would never see it. Then vote totals fell off the map year after year, and the concern was such a contest would absolutely suck even if we did get it.
On top of that, the contest would end up being 4 matches per day for two rounds, then two matches per day until the semifinals. A lot of people didn't like this idea, myself included, because the fear was we would see low vote totals and have too many matches decided by collateral damage. The example everyone will jump to is that Mario 3 might have beaten Pokemon without the Undertale rally, though one could argue Pokemon would have won that one no matter what. In the end, the fears about the format weren't warranted and we ended up having a DAMN good contest.
Snubs was a huge theme of the pre-contest chatter, and there were a LOT of snubs due to Allen's weird contest prerequisites. Notable snubs: Sonic 3 & Knuckles (the guru nomination), Mario 1, Zelda 1, Link's Awakening, Final Fantasy 1, Mega Man 2, Street Fighter 2, Super Smash Bros. Brawl, League of Legends, Star Fox 64, God of War, and I'm sure others I've missed.
First and foremost, let's get Allen's data dump out there for archive purposes:
Allen: "Howdy all. I've been watching b8 over the contest and time and time again, I see people talking about vote stuffing, how things can't possibly be as they seem, and so on. I'd like to set the record straight and discuss these rallies. I'll show what they are, explain what they aren't, and hopefully that will put an end to it.
First, let's look at the referral traffic per match for this contest: http://i.imgur.com/1Ej81VQ.png
You can easily pick out the Undertale matches and can see the effect that the rallies had on both those matches and the surrounding ones. The reddit traffic is generally constant during the contest. It goes up during rallies of course, but not nearly as significantly as other sources. You can see that the Tumblr and Twitter traffic massively spikes during these rallies. Outside of the UT matches, those sources are very small. You can also see that FB became a pretty big rally player near the end once people started reaching out to other sources for rallying. The next 4 referral sources (SA, 4chan, Steam, Smashboards) added just a few sprinkles on top.
Let's just look at the most recent match of SSBM v UT. Here are the top four ISPs (making up over 35% of the total vote) and how they voted:
Comcast: 48.87% / 51.13%
Time Warner: 48.96% / 51.04%
AT&T: 46.30% / 53.70%
Verizon: 49.41% / 50.59%
You can see that these results are all pretty close to the final result. Undertale simply got more votes. It wasn't vote stuffing or any conspiracy, it was actual people coming to the site and voting. Let's actually look into that further.
To counteract bots and people using multiple devices from one location, we restrict votes to one per IP address for logged-out users, and carefully scrutinize multiple votes for logged-in users. For the Melee/UT battle, multiple votes from the same IP were 709 / 641 in favor of Melee, so that's not really an issue.
Now, a few people have mentioned that it's "easy" to stuff votes by simply using different devices (or clearing your cookies) and changing your IP address, even by driving around and using multiple open wi-fi signals to vote. Well, of course a few people are going do that, just because they can, and it's impossible to tell the difference between one person using two devices in two locations and two people doing it.
But, how many votes can one person could actually get into the system doing this? And remember, by stuffing like this, you'd end up creating clusters of votes from a single area or a narrow range of IP addresses, which are easy to spot (and happens naturally all the time, thanks to cell phone companies re-using narrow blocks of IP addresses, typically evening out for both battlers).
So, just in case, if we take all of the potential vote "clusters" (more than 10 votes from a single block of 255 neighboring IP addresses) and add them up... it's 2414 / 1897 for Melee.
Finally, just for laughs, on Friday we had one last thought: What if there's some rogue botnet that's forging votes through thousands of compromised home computers? So, we set up one last gasp attempt to catch this phantom vote-stuffer, throwing in an additional hash into the voting script right in the middle of an Undertale rally that would have immediately exposed any voters not actually using a live web browser. And it didn't catch a thing. Well, it did catch one registered user (who we refuse to give any additional publicity to) that apparently tries to get the very first vote each day via curl script.
So, from a technical perspective, there's nothing suspicious about the incoming votes, and definitely nothing favoring one side over the other. Real vote stuffing (where a small number of people vote a large number of times) leaves clues and signs, and there aren't any. We've chased down every lead, and explored every possibility. We didn't believe Undertale's fanbase was capable of this at first either, and nobody else saw this coming but the 124 people who picked it to win the contest (making it the 17th most popular pick).
And remember, it's all just a big popularity contest, and we all know that the actual Best. Game. Ever. is Final Fantasy IV. Tomorrow will be a fun day of Grudge Matches, so hopefully you'll all enjoy that.
During the contest, we are pretty tight-lipped about these types of things because we don't want people looking for patterns and trying to find weaknesses. With the contest basically over at this point, there's no concern, since the methodology changes (and improves) from year to year."
Nanis: "Can you tell us how Undertale matches go for registered - unregistered users? Can be funny if it's something like 80-20 against Undertale"
Allen: "Sure, why not. It should be a surprise to no one that UT would lose every time if it's just logged in users. But a contest with only logged in users would be predictable and boring.
Undertale vs. Mass Effect 3 - 29.49% 70.51%
Fallout 3 vs. Undertale - 75.67% 24.33%
Undertale vs. Super Mario World - 19.28% 80.72%
Pokemon Red/Blue vs. Undertale - 76.30% 23.70%
Undertale vs. Super Mario 64 - 26.21% 73.79%
Smash Bros. Melee vs. Undertale - 69.59% 30.41%
Undertale vs. Zelda: Ocarina of Time - 33.44% 66.56%
SSBM would have lost to CT and FFVII.
Smash Bros. Melee vs. Portal: 70.37% / 29.63%
Smash Bros. Melee vs. Pokemon Gold/Silver: 57.72% / 42.28%
Smash Bros. Melee vs. Kingdom Hearts II: 59.51% / 40.49%
Chrono Trigger vs. Smash Bros. Melee: 63.28% / 36.72%
Smash Bros. Melee vs. Final Fantasy VII: 42.92% / 57.08%"
Allen, in response to some of Lopen's ridiculous conspiracy theories: "Lopen, let's get real specific (and a little privacy-invading) for just one second.
For the city of (city where you live), using (ISP you are connecting on) the Melee/UT votes were 195 to 245 for Undertale.
Those Undertale voters? They're all around you, right now. They live. And they are watching you."
Contest favorites, aka the top 20 games that people picked to win the entire contest in their bracket:
1: The Legend of Zelda: Ocarina of Time - 4145
2: Final Fantasy VII - 1831
3: Chrono Trigger - 703
4: The Elder Scrolls V: Skyrim - 429
5: Grand Theft Auto V - 319
6: The Legend of Zelda: A Link to the Past - 281
7: The Last of Us - 251
8: Super Smash Bros. Melee - 239
9: Pokemon Red/Blue/Yellow/Green Version - 230
10: The Legend of Zelda: Majora's Mask - 227
11: Minecraft - 205
12: Super Mario Bros. 3 - 203
13: The Witcher 3: Wild Hunt - 203
14: Dark Souls - 171
15: Final Fantasy VI - 158
16: Super Mario 64 - 125
17: Undertale - 124
18: Metal Gear Solid 3: Snake Eater - 122
19: World of Warcraft - 121
20: Half-Life 2 - 104
I'm pretty sure I covered my thoughts on this contest and rallies in general during several writeups, so I'll skip right to the fun stuff.
Allen's grade this contest is a solid A. The only thing stopping me from an A+ is changing formats halfway through. Winning a division is worth having a poll to yourself, and letting the tournament quarterfinal be as originally intended would have meant two extra days for the contest. This is frankly unacceptable to do for several reasons, the biggest of which being that people might pick winners based on who the poll is shared with. The other, like I said, is that winning a division needs to be treated with the utmost respect.
That being said, this format is the best idea these contests have had since they were first thought up. It needs to stick around, with 4 matches per day for two rounds, two matches per day for two rounds, then 1 match per day for the quarterfinals, semifinals, and final. If this causes neat rallies and more random results in the future, good. This format is the best one we've had to date, and the whole 12 hour day/night thing has run its course. 24 hour polls are the way to go, and 4 polls per day for 2 rounds lets you blast through the boring matches, then gradually give the more important matches their own time. Most people seem to agree, which I'm happy with. That, along with rallies making contests fun, is my big takeaway from this contest season.
People look forward to my contest ranking every season for whatever reason, and here it is:
Fall 2015 (Games)
Fall 2010 (GotD)
Spring 2004 (Games)
Summer 2002 (Characters)
Summer 2013 (Characters)
Winter 2010 (Characters)
Summer 2003 (Characters)
Summer 2005 (Characters)
Fall 2009 (Games 4-way)
Summer 2007 (4-way Characters)
Summer 2006 (Female Bracket)
Spring 2006 (Series)
Summer 2008 (Characters 4-way)
Summer 2004 (Characters)
Spring 2005 (Villains)
Fall 2011 (2v2 Characters)
It's true (I also bumped Summer 2003 down a little). Best contest we've ever had, best winner we've ever had, most fun I've ever had writing a PCA, and best format we've ever had. This contest was really good, especially with the format being so neat. If nothing else, I hope Allen realizes how good an idea multiple polls per day for 24 hours really is. I know we were all worried about wildly different vote totals and such, but it clearly worked out in the end.
Someone asked if I enjoyed winning 2004 or cheering for Undertale more. My honest answer is Undertale. You guys have to remember that I'd been very jaded with video games for a very long time. I liked a lot of games, don't get me wrong, but I didn't love them. Undertale got me to love gaming again, and did a lot to help me get my life on a positive track. You can call that lame if you want, I don't mind, but I'm not alone in feeling that way. Undertale's community is one of the best I've ever been a part of, and cheering it on this contest will always be one of my favorite gaming memories. It was a neat thing to be a part of, and it was especially neat relearning what loving video games is like. I was really and truly worried that Brawl murdered gaming for me, but it turned out not to be a permanent thing. I don't think you guys really understand how huge that is. Game of the Decade is my #2 contest ever entirely because Brawl lost. Fuck Brawl.
Last, but not least... The Ultimate Loser™!
Bayonetta 2! Undertale > Ocarina > Super Mario RPG > Metal Gear Solid 3 > Metal Gear Solid > Shadow of the Colossus > Red Dead Redemption > Bayonetta 2.
Thanks as always and see you next time!