subliminal_Verses's blog

By subliminal_Verses, history, 5 days ago, In English

My current goal is solve a ,b under 30 min and c under 1 hour to get d .

 
 
 
 
  • Vote: I like it
  • 0
  • Vote: I do not like it

»
5 days ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by subliminal_Verses (previous revision, new revision, compare).

»
5 days ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by subliminal_Verses (previous revision, new revision, compare).

»
5 days ago, # |
  Vote: I like it +19 Vote: I do not like it

voting here is so strange more than Political Voting in Egypt

»
5 days ago, # |
  Vote: I like it +11 Vote: I do not like it

try to solve A in 2 hours

»
5 days ago, # |
  Vote: I like it 0 Vote: I do not like it

In every contest solve either C or D whichever u feel as an easier one. Then solve A, B and other problems. I am doing this. If i can't solve C or D then i won't solve A, B. I will keep trying until the end of contest.

»
4 days ago, # |
Rev. 2   Vote: I like it 0 Vote: I do not like it

I just do regular practice problems and try to do them as fast as possible, and usually get A and B within 20-30 minutes.

»
4 days ago, # |
  Vote: I like it +13 Vote: I do not like it

I noticed a lot of people focus on number of problems solved or problem letters. For example, it's common for there to be a comment like "first time solving problem D" on a contest announcement blog. In my opinion, it doesn't make too much sense to focus on that because difficulty fluctuates every contest. ABC could be the bare minimum to not lose rating or the goal to strive for depending on which contest.

Instead, a more reliable metric to match is performance. According to performance rating calculators like Carrot, generally top 40-50 in a Div. 2 is red perf, top 200 is yellow perf, etc. The heuristic I use for red perf is usually to solve every problem in contest with at least 100 solves in a reasonable amount of time (so if it's 6 problems in a global round, the split usually looks something like first 4 in first hr, remaining 2 in second hr). This heuristic is more adaptive; if a contest happens to be exceptionally easy and every problem is trending towards over 500 solves, then a candidate master taking the contest should be trying to AK to get a good perf instead of abiding by some usual A-D or A-E rule.

P.S. Of course, performance is moreso something to look at after the contest to reflect. During the contest, your focus should just be "try to solve as many problems as you can."

  • »
    »
    4 days ago, # ^ |
    Rev. 2   Vote: I like it 0 Vote: I do not like it

    Performance is useful after the contest but I believe score distribution is more reliable in predicting how well you should perform to reach X perf before the contest.

    Rough estimates: CM perf is probably solving all problems <=2250pts, Master/GM can be <=2500/2750, IGM+ perf can be <=2750+ etc.

    P.S. There are other cases to consider with score distribution like in scenarios with more than one problem having 2250pts for a CM etc.

    • »
      »
      »
      4 days ago, # ^ |
        Vote: I like it +8 Vote: I do not like it

      I would look at score distribution if they weren't so... arbitrary? Thing is, the raw magnitude of the score is meaningless because it completely depends on how many problems are in the set and how much weight the author wants to give to each problem with respect to the other problems. Just looking at the last two Div. 2s for example, the final problem in both sets were assigned scores of 3250 and 2750 respectively, but the problem with score 3250 received 39 in contest solves while the problem with score 2750 received just 17 (excluding out of competition). Another example, Global Round 11 flattened the scores of every other problem in order to make problem H = F + G, which was how we got a 2500 rated problem with a score of 1500. It's also just really hard to estimate relative difficulty to such a high precision (how do you know for sure that problem C is exactly twice as hard as A?), so the actual solve count distribution often deviates from the initial score distribution, usually by a lot in the tail end of the problemset. So in my opinion, score distribution is about as useful problem letters.

»
4 days ago, # |
  Vote: I like it 0 Vote: I do not like it

Well, you don't have to think too deeply about this. The best practice strategy for this is of course "as fast as possible"

»
3 days ago, # |
  Vote: I like it +2 Vote: I do not like it

Now-a-days I am creating mashup on my own where I am setting very small time and trying to speed solve 1500-1700

»
3 days ago, # |
  Vote: I like it 0 Vote: I do not like it

cheat(I'm just joking)