PlaidCTF 2012 write-up: The Game

The following description is provided:

Robots enjoy some strange games and we just can’t quite figure this one out. Maybe you will have better luck than us.

The game offers a choice between two hex strings asking for the “bigger” one. Anyway, it doesn’t appear to be a way to determine which one is bigger just by looking at the values.

The challenge consists in winning at this game for 75 times in a row.

After several attempts, it’s possible to see some recurrent values. Indeed, it turns out to be a finite set of hex-encoded random 17-byte strings with a strict total ordering on them. We counted 500 different strings during our experiments.

The strategy for winning is to play while keeping track of the values and sorting them on a list. The following script succeeded in winning for 75 times in a row after about 3 hours.

Key: d03snt_3v3ry0n3_md5

UPDATE: is a multithread version of the above script with other minor changes. By starting 50 parallel threads, it managed to complete the challenge in 1 minute and 30 seconds performing a total of 13000 requests. Not bad at all! :)

Author: Marco Squarcina

Computer Science student and an Open Source enthusiast. My main interests are computer security (especially mandatory access control systems), Linux systems administrations and audio applications.

6 thoughts on “PlaidCTF 2012 write-up: The Game”

  1. Awesome!

    I do wonder though, if these are md5 sums of random integers or did the backend program have these values in an array upon which it just looked at the value of the array index and not really care about the md5? Also, I am curious how many iterations it took your script to get to 75/75?

  2. Ehya!

    To answer your questions:

    1) “hex-encoded random 17-byte strings” is a quote taken from irc of one of the organizer, so no md5 hash involved in this challenge :)

    2) this would be interesting to check. I said the script took 3h to get 75/75, but I didn’t mention that my connection has a not-so-good latency. I noticed there are other write-ups for this challenge, where authors claim that their scripts completed in <1h: it would be nice to compare all of them against a local game simulation.


  3. Your solution is far more elegant than mine was. I suspect overall I took longer to do the whole challenge.

    Someone in #pctf was saying they’d made a multi-threaded solution (after the CTF was closed) that connected several times. They were able to solve it in 4 minutes!

Leave a Reply

Your email address will not be published. Required fields are marked *