FAQ  •  Login

Level 21

Moderators: UncleTimmy, mjpieters

Enjoying the challenge? Need a hint?
Make a donation and help keep the site running! -thesamet
<<

cicadia

Posts: 9

Joined: Tue May 10, 2005 2:48 am

Post Tue May 10, 2005 12:06 pm

Level 21

Very intriguing... an unannouced challenge :)

Why does everything about this seem backwards?
<<

UncleTimmy

Posts: 118

Joined: Thu May 05, 2005 3:44 pm

Post Tue May 10, 2005 3:49 pm

Re: Level 21

cicadia wrote:Very intriguing... an unannouced challenge :)

Why does everything about this seem backwards?


It will be clear when you get to level 22 :wink:.

Oh, OK, 21 and 22 have been announced. I'm still trying to climb the fence in 20.
<<

wiml

Posts: 6

Joined: Mon May 09, 2005 9:24 pm

Post Tue May 10, 2005 10:46 pm

there's something primal about this puzzle, or maybe I'm just fooling myself.

kids' game?
<<

dfsmith

Posts: 21

Joined: Fri May 06, 2005 10:50 pm

Location: Silicon Valley

Post Tue May 10, 2005 11:36 pm

I'm guessing there are supposed to be errors in the file. And that's what the game is. Did anyone /not/ get errors in the file?

[Oops---curl appends by default. Bleach! File is fine.]

Hey wiml, did you see any primes? Or are you just fooling us. (Would be appropriate given the RSA news today.)
<<

cicadia

Posts: 9

Joined: Tue May 10, 2005 2:48 am

Post Wed May 11, 2005 1:34 am

wiml wrote:there's something primal about this puzzle, or maybe I'm just fooling myself.

kids' game?

Thinking primally leads me to 2917 (and 41, and 2) but I don't think that that means anything.

There are some interesting things going on with period 7 at the start, but I'm stumped as to what that kind of regularily might mean.
<<

rspyc

Posts: 80

Joined: Sat Apr 30, 2005 8:15 am

Location: Flanders, Belgium

Post Wed May 11, 2005 2:18 pm

What game?

Does anybody recognize the game? It don't think it's a game I know, unfortunately.

I don't really see a pattern in the file data at the moment. There's something with period 7 in the beginning, but that's only in the beginning it seems. And the total size is not a multiple of 7 (and neither is it a prime, unless I made an error).
<<

rspyc

Posts: 80

Joined: Sat Apr 30, 2005 8:15 am

Location: Flanders, Belgium

Post Wed May 11, 2005 2:20 pm

Oh, now I see where 2917, 41 and 2 come from. Not that it helps though.
<<

wiml

Posts: 6

Joined: Mon May 09, 2005 9:24 pm

Post Wed May 11, 2005 7:26 pm

The small number of factors made me think about the old SETI message ideas, but that led nowhere. :/
<<

dfsmith

Posts: 21

Joined: Fri May 06, 2005 10:50 pm

Location: Silicon Valley

Post Wed May 11, 2005 9:29 pm

FYI: the char 255 is significantly underrepresented (over 4 standard deviations down). Hmm. (The peak char 171 is only 2 s.d. over.)
<<

UncleTimmy

Posts: 118

Joined: Thu May 05, 2005 3:44 pm

Post Wed May 11, 2005 9:36 pm

Chi-square

If you count the number of times each byte value appears, and apply a chi-square test against a uniform random distribution, turns out you cannot reject the hypothesis that the bytes are random at the 99% confidence level. (Ya, I speak this language the way you weenies speak HTTP protocols -- & if you did speak this language, it would suggest something obvious to you too :wink:.)
<<

dfsmith

Posts: 21

Joined: Fri May 06, 2005 10:50 pm

Location: Silicon Valley

Post Wed May 11, 2005 9:51 pm

Re: Chi-square

UncleTimmy wrote:...chi-square test against a uniform random distribution, turns out you cannot reject the hypothesis that the bytes are random at the 99% confidence level.


But the bytes aren't random (I hope). Anyway, sigma=4.5 runs about 1:300,000. Given there are 256 data points, that suggests only a ~0.1% chance that 0xFF is uniformly randomly distributed. I'd check your chi... B-)
<<

johnlr

Posts: 41

Joined: Thu May 05, 2005 4:17 pm

Location: kiwi in aussie

Post Wed May 11, 2005 9:57 pm

Judging by the name and the apparent randomness, I'm guessing it's likely to be a compressed file of some type. Do any common compression algorithms underrepresent 255?
<<

cicadia

Posts: 9

Joined: Tue May 10, 2005 2:48 am

Post Wed May 11, 2005 9:58 pm

dfsmith wrote:FYI: the char 255 is significantly underrepresented (over 4 standard deviations down). Hmm. (The peak char 171 is only 2 s.d. over.)


Other than that, (0xff, not so much 0xab, it's in the same league as 0xa4, 0xbf, 0xed, 0xf2) the file is extremely random. I've got all kinds of programs looking for cycles and patterns, and haven't found anything outside of the first 37 or so characters. I'm thinking more and more that ".pack" means something very specific, which I'm just not seeing yet.

0xff might be some sort of marker... the data is almost certainly comressed between those markers though

Grr. Back to research I go :)
<<

UncleTimmy

Posts: 118

Joined: Thu May 05, 2005 3:44 pm

Post Wed May 11, 2005 10:06 pm

Re: Chi-square

dfsmith wrote:But the bytes aren't random (I hope). Anyway, sigma=4.5 runs about 1:300,000. Given there are 256 data points, that suggests only a ~0.1% chance that 0xFF is uniformly randomly distributed. I'd check your chi... B-)


The chi-square calc is fine -- it's about 304 with 255 d.f. This is larger than the mean, but about 1.8% of truly random sequences would exceed that chi-square stat.

That doesn't mean "the bytes are random" -- it means what it means :wink: Note the file name. Also compare its size to the size of the file you extracted it from to get a third clue pointing at the same thing (is there not something surprising about how those sizes compare, considering the kind of file you extracted it from?).
<<

jtauber

Posts: 47

Joined: Sun May 08, 2005 10:20 pm

Post Wed May 11, 2005 10:14 pm

Re: Chi-square

UncleTimmy wrote:That doesn't mean "the bytes are random" -- it means what it means :wink: Note the file name. Also compare its size to the size of the file you extracted it from to get a third clue pointing at the same thing (is there not something surprising about how those sizes compare, considering the kind of file you extracted it from?).


zip, gzip and bzip2 all fail to compress it. That's pretty tightly packed :-)
Next

Return to Python Challenge Hints

Who is online

Users browsing this forum: Google [Bot] and 8 guests

Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.