Project Euler Forum

This weekend's change in criteria for acknowledging 'best performers' may encourage the top 100 or so enthusiasts, but is a real turn-off for those of us who doe not - for any reason - have time very early on Saturday mornings to make this their first priority.

I have been bragging to my kids that I had earned my way into the "Eulerian 100", and I am quite disappointed that the rules have now changed.

There might be at least 3 ways to recognize the diversity of your dedicated supporters:
(i) - As until yesterday or so, percentage of the last 25 problems
(ii) - As of today, the 'fastest 20 solvers'.
(iii) - Cumulative success, weighted by speed: award 100,000 points per problem, going back to #1: give bonus points to the top 20, then split the rest evenly based on the number of solvers, then add the scores by solver. Solvers of really hard problems will score highly, as will fast solvers of easier ones.

Aside from concerns already mentioned, I'm wondering if there is anything that can be done for problem suggesters who are also active solvers. So for example niino on problem 289 had no opportunity to score, as the problem suggester; in the old Eulerian list the effects of this would only last one week, but now the effects will last 200 days.

In my opinion, the new system is sophisticated and it is hard to make special criteria for problem suggesters. At least, I would not disturb fastest-20 list. I do not care so much that my score does not grow when my suggested problem is published, because I felt much honorable when I confirmed many solvers enjoy it and saw various algorithms in the forum. Anyway, I appreciate your warm concern for suggesters, cyclops

niino wrote:In my opinion, the new system is sophisticated and it is hard to make special criteria for problem suggesters. At least, I would not disturb fastest-20 list. I do not care so much that my score does not grow when my suggested problem is published, because I felt much honorable when I confirmed many solvers enjoy it and saw various algorithms in the forum. Anyway, I appreciate your warm concern for suggesters, cyclops

For the record, I'm in favor of the new system, and phsr's concerns don't personally apply to me. I happen to enjoy waking up at (or staying up until) weird hours trying to solve the latest problem. And I feel the new chart gives interesting information that would otherwise be somewhat hidden.

cyclops wrote:

niino wrote:In my opinion, the new system is sophisticated and it is hard to make special criteria for problem suggesters. At least, I would not disturb fastest-20 list. I do not care so much that my score does not grow when my suggested problem is published, because I felt much honorable when I confirmed many solvers enjoy it and saw various algorithms in the forum. Anyway, I appreciate your warm concern for suggesters, cyclops

For the record, I'm in favor of the new system, and phsr's concerns don't personally apply to me. I happen to enjoy waking up at (or staying up until) weird hours trying to solve the latest problem. And I feel the new chart gives interesting information that would otherwise be somewhat hidden.

I also happen to enjoy the new chart system. I do however mind that for one in six problems the posting time overlaps with a work day on friday. I would (rightly) get fired if I started working on those problems at work. I imagine this is even worse in the americas, so why not postpone the posting cycle with 12 or even 24 hours ?

Regards,

Michael

I think you have a point here.
The team are discussing what publication scheme would be best.
In any case we are planning to shift the publication scheme to somewhat later in the weekend.
However, it cannot be shifted too much as we have also a lot of solvers at GMT+9 ore more.
E.g. Sunday 13.00 GMT would mean Sunday 22.00 at GMT+9 (japan) or even later in Australia.
We will announce changes in the scheme a few weeks ahead.

Just out of curiosity, how are tiebreaks determined? For example, there is currently a 3-way tie for 13th place, with koen listed first of the 3, etc.

Member with the lowest ID first (that is oldest members first).
At least I think it was done that way. But perhaps Colin might better answer this question, because he programs the thing.