Arcstar
  • Welcome page
  • Tournament
    • Getting Started
    • Data
    • Model
    • Submissions
    • Scoring
    • Leaderboard & Payout
    • API Key
    • Profile Page
    • FAQ
    • Support
    • Socials
    • Live Score Computation Process
  • $ARCSTAR
    • How to add the Token
    • How to use your Wallet
    • Tokenomics
    • Audit
  • Roadmap
    • General roadmap
Powered by GitBook
On this page
  • Active datasets lookup
  • Listing of primary submissions
  • Live Score computation
  • Mean by dataset
  • Mean of mean
  • Normalisation
  • Computing global leaderboard
  1. Tournament

Live Score Computation Process

PreviousSocialsNextHow to add the Token

Last updated 2 years ago

Computing the live score each weeks is a very computationally intensive process.

Targets computer have to be made to make sure everything works as expected.

Active datasets lookup

This board allows you to control which dataset is currently active and score with live data.

Listing of primary submissions

Primaries submissions are the ones that you submitted and selected for live scoring.

Live Score computation

For each primary submission, the scorer will compute the spearman correlation score between your prediction for that target (30, 60, 90 days) and what happened live.

Mean by dataset

For each dataset your live score for this dataset is the mean of your last two participation at this dataset.

Mean of mean

Global mean score is the mean of all your live scores over the universe of strategy. If you miss a round you get a score of zero.

Normalisation

All round' scores are normalised to avoid some rounds to be more weighted in the overall score.

Computing global leaderboard

Arcstar users are ranked by their mean live spearman correlation score normalised over all strategies.