@vsp4 please describe these formulas you have used so that we can make ratings predictor?
Can we be assured that we will get to use codechef rating predictor again ?
It will be down for indefinite amount for time because currently I’ve made the changes for submissions API which was my last blocker.
But now the next blocker is https://discuss.codechef.com/t/request-for-change-in-codechef-submissions-api/48480. I’ve got a mail yesterday from codechef saying that are stopping the support for API indefinitely. So I think it will take long time to comeback.
The only solution is I can start the predictor after the end of contest because parsing submissions during ongoing contest is highly inefficient because of above issue, I need to parse all submssions over and over again instead of new ones. It will take time to parse after contest since there is restriction for 2 submissions per second. Going by the December contests this is how long it’s gonna take
LTIME79A - 1920 submissions - 16 minutes
LTIME79B - 19560 submissions - 2 hours 43 minutes
COOK113A - 1050 submissions - 8 minutes 45 seconds
COOK113B - 10890 submissions - 1 hour 30 minutes
DEC19A - 29860 submissions - 4 hours 8 minutes
DEC19B - 213700 submissions - 29 hours 36 minutes
As you can see long challenges can take a day or two because of large number of submissions. But nowadays rating on codechef is updated in mere few hours(I might be wrong, correct me if that’s the case). So I don’t see much benefit but this may still help killing the curiosity for rating change.
If that’s the case then I don’t see any benefit of doing this. I believe we will get ratings change very soon after contest ends. It was beneficial for long challenges ( sometimes even for short contests) where we can predict ratings in between of contest to get motivation/happiness.
See , if we apply some thoughts. They most probably did it cuz people in Long Challenges might see how many problems they can solve, if they can solve ‘x’ problems, they check if its enough to get a rise, if yes, then they submit. But you have to be confident about your solutions before coding them.
Doing this is misuse of tools. One shouldn’t do this. One should participate without checking it. This will hinder one’s overall progress. Don’t participate for ratings. Instead participate for learning.
Maybe Codechef is trying to stop this
In my case I don’t check it before submitting. I solve as much as I can in one try. Then I check it to see what’s the ratings changes to get some motivation. I get motivation if I see I am getting negative ratings change or if I am going to miss being 6 star by few ratings. This is a good use of this tool.
I don’t think this is the reason. Might be one of many reasons like no/less resources available to maintain the API right now at the present moment.
Any way I feel predicting rating change before contest isn’t bad because after all these are competitive contests right? That’s why CF predictor has so many users - Firefox(1.4k) and Chrome(14.5k). But yes I understand, few people might see it as a bad habit. I feel it’s not bad until you break the rules. I agree with both opinions though.
It isn’t bad. What I am saying is not participating because you will not get positive change is bad.
i think we can do it just for short contests of div1.only takes 8-16 minutes.feasible
I was thinking along the same lines. Moreover if possible having a separate queue/scheduler for div2 which start after contest ends. Will give it a try. Ofcourse this way new predictor will be very slow compared to the old one.
It is feasible in long challenge for div1 users at least.
You can probably run it for 4 hrs every night so that we can get prediction after a day at least.
Or just keep running it for 10 days. It will be updated every 4 hours or so.
Something is better than nothing.
Only God knows when the rating predictor works and after every long challenge this post appears on discussion forum despite the fact that its under maintenance
I’m sorry for the let down. Should change the title lol. This post appears frequently because people keep asking and I’m unable to deliver.
@dushsingh1995 Not a problem bro!. you did a great work btw
Not your fault - if Codechef no longer provide the required APIs, then there’s not much that can be done
I personally agree with you on this point. When it comes to gain/loss, you cannot learn.
Thanks to everyone who have used the predictor and found it useful. This was one of the reasons to bring back the predictor and other one being to apply my newbie web development skills and to learn something new.
I’m sorry to say this, I won’t be able to bring it back again since there is no support as of now. Link - https://blog.codechef.com/2020/01/20/api.
Once again thank you everyone for your feedback and support for the predictor.
Sad to hear this, but I’m not an Admin and can’t do anything XD
rating predictor is Not working?