Project Euler Forum

I think there's still a bug in the ranking page.
At this time 70 solvers shown in the page of 401-450 problems, but 71 in rank list.
There's a 1 offset since some days, I thought it will be fixed, but as I see it again, I share.
It's not very important.

Francky wrote:I think there's still a bug in the ranking page.
At this time 70 solvers shown in the page of 401-450 problems, but 71 in rank list.
There's a 1 offset since some days, I thought it will be fixed, but as I see it again, I share.
It's not very important.

Is the website hacked?

It looks like there are fictitious solvers among the top:

414: 1st place: Popeye
415: 1st place: Bugs Bunny
419: 1st place: Garfield
420: 3rd place: Fred Flintstone

Can this be the source of the mismatch in number of solvers?

That's what happens when someone deletes their account (in this case, sk7).

Marcus Stuhr wrote:That's what happens when someone deletes their account (in this case, sk7).

So, is this the source of the mismatch or is it an unrelated issue?

There was a solver (abc777) last year, who had lots of points and first places, and deleted him/herself. Euler decided that deleted account looks bad, so every deleted user has some name from a tale. If you move your cursos above the name, it will tell you "Real user deleted account" or something like that.

TheEvil wrote:There was a solver (abc777) last year, who had lots of points and first places, and deleted him/herself. Euler decided that deleted account looks bad, so every deleted user has some name from a tale. If you move your cursos above the name, it will tell you "Real user deleted account" or something like that.

The real question/problem is still:
"Is this the cause of the mismatch to the number of solvers on different pages?"

At first I thought the mismatch was caused by different cache frequencies of pages,
but even after hours have passed, the mismatches are still there...

I'm clearly missing something here... could you explain which lists are showing a mismatch?

Thanks for the PM, ignat_s, explaining what was meant. *feels stupid now* And you were spot on with his explanation in your PM too!

See new post below...

@Euler : on my browser http://projecteuler.net/problems;page=9, the last line shows (at this time) 85 solvers for PE421, but on fastest solver rank list, there are 86 people, and there's no sk7 (nor popeye, ...) in the list.

Francky: sk7 was originally rank 33 for problem 421

@Francky: Thanks for explaining it too.

The Fastest Solvers (FS) table is a separate table in the DB so it always captures - and keeps - a record of who solved the problem and when it was solved. The Problems table, however, shows the number in the current membership DB who have solved it. Since we've been allowing members to delete their profiles this meant that such an issue could occur. Obviously we couldn't remove the deleted member from the FS table because then the next person to solve the problem - possibly way after the table was filled up - would enter the table in the newly created gap. To get around this I fill any unidentified members (i.e. those whose accounts have been deleted) with randomised cartoon character names. However, recently all problems solved are timestamped in members profiles, so in theory we could figure out who the next person would have been. In fact, having separate FS tables in the DB is possibly a redundancy now. This won't fix the problem for older tables, but if I can ever find the time then I might modify the scripts to automatically fix this issue. At the moment that would be near the bottom of my rather extensive to-do list.

Dear euler,
Thanks for the extensive explanation! However, there is a little detail that falls out of the picture: for the problem 416, it shows 87 solvers as of now in both lists, and there is still a deleted account in the last position.

The FS table only ever updates if a new person enters it; once it's full it won't change unless one of the admins clears the system cache. The Problems table updates automatically every hour so there may be a time delay between the two. When I checked just now the FS table had 88 (which includes the deleted account) and the Problems table says 87, so it's all good again.