Back to all discussions

  •   over 2 years ago

Clarifications regarding the outputs

I would like some more clarifications regarding the expected outputs. Regarding the tournament rankings, will we be expected to give outputs for tournaments dating back to 2020, or is there a cutoff for the time? (e.g. we will only be asked to give outputs for tournaments that occurred in 2023)
Additionally, how would this output be judged? Would it be judged based on how the rankings reflect the actual standings of the teams in previous tournaments? (e.g. if our model predicts the top 3 teams at MSI 2023 to be JDG, BLG, and SKT in that respective order, it is in good standing)

Also, if possible, I would like to understand how the judges determine if our ranking system reflects the state of the competition. For example, if our model predicts LNG to be the best team in the world, how would the judges decide its accuracy when this year's Worlds has not finished yet?

Thank you very much for your time.

  • 1 comment

  •   •   over 2 years ago

    Hi Yong Seok - appreciate your patience here!

    For your first part of the question, please see the clarifications below for tournament and team rankings:
    Tournament rankings: your method should return a ranking for any of the tournaments from a MSI/Worlds qualifying region for 2022-2023
    Global and team rankings: your method only needs to consider teams that have participated in a MSI/Worlds qualifying region in 2023
    List of regions: LPL, LEC, LCK, LCS, PCS, VCS, CBLOL, LJL, LLA

    As for the second part of this question, our judges are currently working on the methodology to score the "realism" judging criteria. What we do want you to keep in mind is that the ranking system's ability to reflect the state of the competition is only one part of the judging criteria. We recommend that participants do not put too much weight on that criteria or over engineer their submissions based on reproducing actual results. If the model outputs some unconventional results, it is fine, so long as you include strong justification for the methodology for your ranking system.

Comments are closed.