Hi all,
Happy New Year! 🎉 I trust everyone had a great break.
The final data points for weeks 52 (22–27 Dec) and week 53 (referred to here as week 1, 29 Dec–03 Jan 2026) have now been used to rescore the leaderboard based on your final predictions.
This final scoring is only based on the predictions for weeks 52 - 1, as all data up to week 51 were already public and presumed to have been used for model updates.
That was a bloodbath lmao
Damn Ikr. How bad were our predictions that we got negative scores? @AJoel could you release the actual values?
What a disaster! The score I was seeing on my leaderboard has been fattening me for slaughter. Could you explain how you arrive at the final scores?
Flames
I am seeing another shake-up. What is happening?
at least the selection for our last evaluation should have been our last 2 submissions
IMO, the multi-metric approach for this challenge was really messed up.
All low scores in the public leaderboard ended up performing better in the privete leaderboard.This is the first time I am witnessing this kind of complete reversal!