https://docs.google.com/spreadsheets/d/1FmANK72CcUfzGLLNtMkRUI50PSQ44HCC-xQhS0Vm5uI/edit#gid=0
Just dumping this here before people continue to complain about Splatfests.
Before Nintendo stopped leaving the real numbers of every Splatfest result in SplatNet 2 (the Nintendo Switch Online App for all intents and purposes) we got perfect accuracy on the results of every Splatfest, how many people picked which team, how many wins each team scored in Solo and Team, so just open this doc and take a look. The last Splatfest that had this data left behind was Action vs Comedy, which was a very long time ago at this point. But the point still stands, on so many of these Splatfests one team had less than a 1000 wins lead, which leads to these 50.8% win factors and the score for that category. Splatfests are working, but Nintendo cannot odds the fact that the unpopular team barely squeezes out enough wins overall to take both battle victories and win the Splatfest.
Let's take 2 of the European Splatfests for this argument,
Warm Breakfast vs
Cold Breakfast and
Film vs
Book:
So here's the data for Warm vs Cold if you didn't click on the link:
Warm Breakfast won this, but the data you saw in game was as follows:
53%-47%
51%-49%
49%-51%
This was before they made it show decimal places but this Splatfest could have very easily gone either way. Cold took Team wins by 28 wins whilst Warm took Solo by 431 wins. 15 matches decided the outcome of Team wins here (yes this doesn't affect the overall victor, but follow through). 1 win for you is 1 loss for the enemy (because these have ended we'd remove 1 for the winning team and 1 for the loser, so a change of 2 every match), so by taking 15 from Cold and giving those to Warm it would have been a 3-0 KO. In perspective that's 14/15 matches out of a total 4274, so that's not even 0.4% of the total matches played in teams (for the final results score less if just matches). If we go to solo we'd flip the results by having Cold win 216/217 matches and Warm losing them we get an inverse:
79'575 - 216 = 79'359
79'144 + 216 = 79'360
Cold wins by 1 match, scrapes the Solo win with the minimum requirement in this scenario (Team stays unaffected) and wins a 2-1, then we get another "Oh the popular team always loses, Splatfest is rigged" nonsensical argument thrown around (I understand this isn't one of those, but here we are).
(Throwing this here to demonstrate how close this was for Book winning, not going to disect the information again.)
Does this justify the current system of Splatfest results? No, but it's a long way off being a broken system as people keep claiming it is. Splatoon 1 started out with the Popularity issue, despite the whole "wins is worth twice of popularity" because popular teams had like 20 point leads on the unpopular and a 6% win advantage only works out to 12 extra points, which is why the multiplier became 4x the amount, which was perfectly fine, but then for some reason they shot it up to 6x and broke everything. This system is by far fairer because you get 3 categories and 1 point for each, with the requirement being "win 2 of 3" for your team to become victorious.
In an ideal world team imbalances like Film vs Book shows wouldn't be an issue, but where there's a topic there's a popular opinion, and where there's a popular opinion there's a massive wave of people bandying it about. This is further enhanced by the idea that Splatoon 2 is a children's video game and a theme like Pulp vs No Pulp will be inundated with kids because they absolutely hate bits in drinks and other liquids (not all, but most) so they skew things more and people avoid that team to not get terrible matchmaking times and bad experiences with a low skilled team.
THIS is where the biggest issues with Splatfest arise, and no amount of fixes will alleviate this. I might as well refer back to Splatoon 1, because of the state of popular vs unpopular Splatfest Power was brought in to separate the S/S+ players (which I was in) from the A ranks, B ranks, C ranks and unranked (those where the 6 starting power requirements, S+, S, A+/A/A-, B+/B/B-, C+/C/C- and no rank). This did separate us because those who were in S/S+ recall the bloodbath Turf War Splatfests of the last few fairly well, and it also meant that team popularity was less noticeable because higher skilled players may have picked a lesser popular team whilst the kids and casuals picked the popular, meaning the win percentages dropped down and made 6x less powerful and popularity scores were worth a little more once again. Splatoon 2 uses placement matches for Splatfest Power, but you still get matched with anyone even when your power reaches 2300+, you even get placed with people in their "seeding" matches which scuppers the whole implementation of Splatfest Power.
A massive essay, but just explaining how it works. I don't want to return to Splatoon 1 where popularity decided way too many Splatfest victories, because winning matches should mean way more than having more players on your team, but Splatfest Power needs adjusting again. I want to see the old system, and make it so the starting power will be decided on if you're Rank X, S+, S, Any A, Any B, Any C or unranked. It's not hard for them to average out your ranks when there's 4 and give you a power, then make power deliberately separate players. The only faults here are players going to Team instead and having a larger spread of areas to search for, so longer queue times occur, but it should at least make scores closer and give the popular team a better chance of winning one of the 2 categories.
TL;DR: The data shows things are fine, but fixes are needed to matchmaking and Power to balance things out, do not go back to Splatoon 1 where popularity was too powerful at the start of the game.