clist
clist copied to clipboard
aropan
Color coding rating on platforms without official color scheme [Enhancement, Minor]
Clist uses color coding for contest ratings on all platforms with a concept of contest rating. However, not all platforms have an official color coding scheme. Most notable examples include LeetCode, HackerEarth, and PrepBytes. I wonder how you decide which color scheme to use in such cases?
For some platforms like LeetCode the current schemes are standard and make perfect sense. However, for other platforms the cutoff numbers appear to be random. For example, HackerEarth cutoffs are 1250-1450-1600-1700-2000. However, their website only mentions cutoffs like 1600 and 1800. 1600 is the cutoff to win prizes in their easy monthly competition (subjectively Div. 3 difficulty), and 1800 is the cutoff between Div. 1 and Div. 2.
On Codeforces, Div. 3 ends with cyan, and Div. 1 starts with yellow, but in the current color scheme Div. 3 ends with blue, and 1800 is mid red. I would therefore prefer the following color scheme:
- 0000-1200 gray
- 1200-1400 green 🟢
- 1400-1600 cyan
- 1600-1800 blue 🔵
- 1800-2000 orange 🟠
- 2000-2200 red 🔴
- 2200-9999 target 🎯
It also should be noted that they changed the rating system somewhere around last June (it hasn't been publicly announced). It is (subjectively) much easier to gain rating ever since. For example, the number of targets more than doubled during this period. Hence I also propose raising the cutoff for being a target from 2000 to at least 2100, preferably 2200.
Ideally, we should ask HackerEarth administration if they plan to introduce official color coding for contest ratings, but I'm afraid we'll have to wait a long time for their response.
Sounds good. I set the color scheme for HackerEarth.
I usually use my inner intuition when choosing rating colors. 🙂
Thank you! I sent an email to their admins proposing the same color scheme, waiting for a response. Speaking of HackerEarth, it appears that the top ratings shown on https://clist.by/resource/hackerearth.com/ are way off from the actual values https://www.hackerearth.com/leaderboard/contests/rated. Perhaps ratings aren't updated properly for accounts that are not registered on clist?
How quickly is the rating updated after a contest?
I wait for the new rating while 5 days after contest. If I don't see it I ignore updating profiles.
Unfortunately, I cannot give you a single number that will be correct in all situations. It appears that 5 days is a very optimistic estimate. For example, they had a contest on the 9-th of April and the ratings are still pending.
Do you check for rating updates on user profile pages or are they somehow accessible grouped by contest? In the latter case, I have a workaround in mind.
My idea was to trigger pulling for rating updates in a contest whenever we see a new individual rating change in that contest. If I understand correctly, this [new individual rating change] happens when I remove and add my account back (this is how I kept my rating updated so far).
So basically any interested user can trigger a rating update for the entire contest. Fortunately, this cannot be abused, as the contest-wise update will be triggered only if we see a new change, meaning that one contest can be pulled at most once (there are no rollbacks on HackerEarth as far as I'm aware).