The ratings of each grade is determined purely from input responses provided by you, the
climbing community.
When you provide an answer to "Which is difficult?", the algorithm will update
both the grades' ratings accordingly.
The more responses from a wide range of climbers, the more accurate the ratings
will be (comparatively to one another of course)!
An invisible component of this each grade rating - sigma - determines the confidence level
of the value given to certain grades.
To put it simply, over multiple matchups, the sigma for each grade goes down which indicates
a higher confidence level of that numerical rating.
This, in turn, 'cements' the rating of that grade: it becomes harder and harder for it to
change rating from future matchups.
Ideally, this means that we will arrive at a fixed state of ratings eventually, ceteris
paribus.
However, since conditions of gyms gradings can change (e.g. changing setters, different
standards, change in business objective), a periodic reset of sigma will be carried out.
As of time of writing , this will happen on 2 occasions: a new year and when adding a new
gym to the platform.
Such resets wil be termed "Seasonal Resets" and will be reflected on the Rating's Submission
front page.
Current algorithm being used is TrueSkills, derived from ELO rating commonly used to rank
chess players and eSports athletes.
All feedback, queries or suggestions are welcome!
Contact me
