@WheatWizard Huh...? Just compare the current graph with the 6 reflections, if there is exactly 1 match then it satisfies.

Yeah, I'm just not having the easiest time figuring out where to reflect so that I catch the symmetries

Deadline passed for 3 minutes. Although it doesn't look like cairdcoinheringaahing is online right now...

It looks like they are in the room

> seen 9m ago, talked 11m ago

Because it takes 15 minutes after you went offline before you are marked left the room.

Well I'm thinking of writing this up as a separate challenge, since I found it rather interesting. Thoughts?

@WheatWizard If you want. What will be the winning criteria?

probably just code-golf

I'd like to do fastest-code, but I don't usually find those very objective

Harder challenges often doesn't fit code-golf.

That's true.

What's wrong with fastest code here? My code for such sequences are terribly slow.

I'm open to reccomendations.

@user202729 1) It varies from hardware to hardware, and my personal hardware is very weak and not very stable. 2) By the looks of things only the first couple cases are going to be feasible to compute, which will make it harder to compare between answers. Usually it is best to compare on a extremely large test cases

Because IMO "serious contender" criteria for fastest code is much easier than that for code-golf .

You can ask for someone else to test submissions.

Also... small number can be "extremely large test case" if the number of graph is large.

Yeah, but its pretty hard to get fair tests between people

@WheatWizard That's a reply to which question? "It varies from hardware to hardware"?

@user202729 My issue is I would want the test cases secret to prevent people from optimizing for the testcases. But when there are only a dozen that are feasable, it makes it so that I have a harder time hiding my secret cases.

@user202729 I just mean that even if I have someone with a powerful computer test it. Speed tests are not very reliable.

Then do fastest algorithm. But then there are cases of unknown-complexity...

I prefer fastest algorithm

But in my opinion speed test for sufficiently-long processes are quite reliable.

Example answer of "complexity unknown" ^

My issue is that sometimes optimizations will make things faster on your platform but not on the test platform, so its hard to know exactly how good you will do until you've already answered

Allow answerers to specify compiler version & flags.

I mean my computer has 1 core, so optimizations for more cores would be pretty useless if I tested it. On the flip side if I'm writing code I can't really tell if optimizations I'm trying to make for multicore processors are effective, because I don't have one.

Just limit the affinity to 1 core. That would be fair.

Yeah, but cores are not the only problem, there are a million hardware variables to worry about and optimize for.

I like it when you know your score before you post.

Oh, don't worry too much about it. Generally faster-algorithm-will-win-anyway .

(unless they get extremely close, in that case it would be problematic)

But I don't think that would ever happen.

@WheatWizard I posted my answer anyway.

Are you allowed to use Java 5? Is that only because the challenge is over?

Java 5 is not yet used.

Then why were you even using Java 1?

That's a mistake.

I should not have used it.

BTW... it seems that markdown counting is limited by int32_t.

Just because it has come up twice now, I might as well say it:

If I'm unable to accept an answer for whatever reason (such as being asleep), you have a bit of leniency. This isn't an official rule, but I've done it twice now, so I might as well make it known.

In short, I'm being lenient with finishing the challenge. If I know that someone is almost finished with a solution, I'll give it a few minutes. If I can't accept an answer for a reason, I'll do it when I am next able to. You have until then to post an answer. Otherwise, however, I will accept the winner when my alarm goes off.

Yeah it's good to encourage people to work on solutions, rather than being strict and immediately accepting. I agree with caird

BTW we should upvote user202729 for his fantastic work.

Personally, +1

almost done with next seq in Actually

and with a lot of spaces.

spaces are lame

Suggest shorter URL codegolf.stackexchange.com/q/133754 for room title.

Does different versions of C++ counts as different languages? (98, 03, 11, 14, 17, 20)

Yes.. 20 didn't exist right now. It hasn't been 2020 yet. @cairdcoinheringaahing ^

That seems to be too cheating. But they are indeed different languages. I hate using 98 and 03.

New sequence!

And hexagonal numbers!

lol you got ninjad

@NieDzejkob Me?

@Mr.Xcoder In the comment section in your answer.

I always forget the number :-/

Note that the snippet matches by regex. / Anyway to make the snippet faster? Look for a server and cache the data? How? Can fasten be applied in this case?

Why'd you want to make the snippet faster?

It's reasonably fast already

Because it's slow, of course. / I mean the code snippet in the question.

Probably just that you have fast internet connection. It works client-side.

yeah, most likely | it runs well for me

(new next seq)

What would a graph of answer posting time look like?

like deltas or time since challenge start? I might add it to the answer snippet...

Has Cpp been used?

Yeah :(