Possible flaw in codechef rating predictor

@vijju123

Today i checked my rating in rating predictor at

I found that another user (hh_algos) with rating 2149 and rank 65 is expected to gain 134 points while i, with rating 2141 and rank 58, will gain 94 points.

I’m neither accusing anyone nor blaming, but was interested to know whether the predictor is working correctly or not…

Because a person with better rank and similar ratings ought to gain more.

I have read the codechef ratings page but it didn’t helped me much understanding this strange rating prediction

Please helpalt text

1 Like

There is a concept called “volatility” used by codechef, which is in play here. Like in probability, we express “Expected value and variance”. Volatility corresponds to variance, or in other words, “How sure are we that he will be able to get the expected rank.”

Of course, if we arent sure about that, then him getting expected rank should lead to a bigger increase in rating (and our expectation next time :3). This was the reason I suffered a ~80 loss of rating in lunchtime despite <200 rank. :stuck_out_tongue:

Further, I think it was made by a friendly user of our community, not codechef, so I am not sure if @admin can do anything about it.

2 Likes

Well

That sort of suprise me…

Is there any way we know our volatility??

And by the way

As volatility sounds to me, does that mean

Suppose two users with same rating, user 1 has higher volatility that user 2

Then user1 will gain more rating than user2 if they get same rank ??

If this is how it works, surely admin needs to have a look i guess…

Exact mechanism isnt revealed, its complicated. No, you cannot know your volatility.

Volatility, in my opinion, was to take care of “someone getting huge boast after a lucky contest and having inflated rating”

But surely it wasn’t in place for my scenario??

Where i gain much less despite better rank??

What is putting you off? I will try to get in touch with admin to confirm the exact, but it doesnt play THAT big of a rule in the long run.

And about users getting lucky contest, surely that particular user isn’t bound to get lucky again and thus, will reach the level, their skills actually are after next contest

Yup, isnt that how rating should be? :stuck_out_tongue:

I know it matters not in long run, but this arose my curiosity, so i posted…

And yeah, thats how it should be :slight_smile:

In long run, the rating will automatically stabilise at level of skills of user, then what’s the need of volatility then??

I will have to ask @admin :3. Lets see when shes free.

There-

 it is to accelerate the convergence to the 'correct' rating. So when a person starts out, her volatility will be high, so that if she's like petr, her rating can jump up a lot in the first few contests itself. And then it tends to stabilize.

Please do let me know what @admin says…

PS: i too joined this june only

Suggestion: if it seems prudent, make volatility visible alongwith rating, this ought to give better idea of his coding skills

What admin said is above. I will try passing on your suggestion to him, although the chances seem sly.

I know that the chances are sly, but still do ask…