I have a question regarding the scoring method. @allawayr have said in [this discussion](https://www.synapse.org/#!Synapse:syn20968331/discussion/threadId=6468) that the score is negative log~10~ p-value, so the bigger the better. And according to the leaderboard and @allawayr 's reply, the baseline method posted by the organizers has the best score so far. To my understanding, participant's scores should be better than the baseline score, technically speaking. But it does not seem to be the case here. Can you kindly clarify to me why is that? If I am mistaken, I sincerely apologize for that. I request you to reaffirm the 'bigger the better' statement. Thank you.

Created by Senthamizhan V SenthamizhanV
Thank you @allawayr . It is now clear to me.
Hi there - Just a note: it's actually -log~2~, not -log~10~ - apologies, that was a typo on my part, I've corrected this in the wiki. Regardless, your question still stands. The baseline was provided by @SzalaiB and represents a relatively straightforward approach to predicting targets. It's *not* a random model and therefore predictions could be either better or worse than it. So yes, bigger scores are better. Thanks, Robert

Question regarding scoring method page is loading…