Hey everyone, happy Valentines day!
Back in December I estimated that v3.0 would be gearing up to release in late January or early February, so it’s time for a little update! While I have missed that estimate, v3.0 is very close to being ready. I now estimate that it’ll be releasing around the end of the month, maybe the very start of March.
Yea this is important to emphasize, data will definitely lie. Newbies often use rogue and huntress which tanks their winrate, but that doesn’t mean they’re worse than warrior. And players who know a lot about the game will try weird things like upgrading bad rings and throwables (and win), but this doesn’t mean that they’re good upgrade options.
This isn’t a game with massive daily players, and if data is from players who won once then it’s a very small sample that can be easily manipulated. Sometimes I’m scared to win with a bad class + item because I know it’ll muddy the data :/
I only consider gameplay data from players who have at least 1 win already, specifically because new players would have an outsized impact on winrates. However, the Warrior is the first character to unlock, so considering only newer players actually tanks his winrate the strongest, as absolute newbies have him as the only option.
You’d also be surprised about the number of runs in total. I won’t give away specifics, but even when only considering 0-chal runs from people with 1+ wins it’s on the order of thousands of runs a day. I actually specifically ignore runs from higher chals because at that stage the number of runs becomes too few to make meaningful inferences.
I know about the 1-win thing, it’s awesome to hear that shattered has decent daily players and it’s going strong, it’s harder to manipulate with those numbers. The meaning of data is subjective, and a bad actor could try losing or winning a lot with a specific build to manipulate it. But it should be harder to do if those daily numbers are true.
Data isn’t perfect, but it’s objective.
Play however you want, it’s a game, and it’s fine lol