🏅 Top traders

Okay, I made round 2. Hopefully, it will be a little harder, though not that much harder.

WyP\r@gaYqTW#*{\mHkx:^lhaAe|G`tz

The title and description of this market have both been encoded with the following method: The ASCII code of the nth character in the encoded text is the number between 32 and 126 (the range of printable characters in standard 7-bit ASCII) that has the same value modulo 95 as the sum of the ASCII codes of the first n characters in the original text. The modulus of 95 is used because that is the number of printable characters, so every string using only those characters has a unique encoding and a unique decoding. This market resolves YES if someone posts the decoded title and description in the comments before it closes. Otherwise, it resolves NO, and I will reveal the decoded title and description after it resolves. I will not post the answer beforehand myself, nor will I give anyone the answer before close. However, I may post clues in the comments at my discretion.

Not sure if I made a mistake on the title or if that was intended to be a random string.

But if I made a mistake other people have plenty of time to correct me, so my work here is done IMO.

@jskf The title isn't a random string. It's encoded using the same method as the description.

@JosephNoonan Are you sure you encoded the title correctly?

@jskf I'll check it when I get the chance. It was encoded with the same code that I used for the description, so it should be correct.

@jskf I went and checked and found that I indeed pasted the wrong string in. I made both an encoder function and a decoder function that performs the inverse operation, and I accidentally put the result of calling the decoder function, rather than the encoder function into the title of this market. I just fixed it with the correct title.

@JosephNoonan "Will someone decode this market?"

@jskf That's the correct title. By the way, how did you decode it?

@JosephNoonan
The first thing I considered was base64 encoding. There were a bunch of characters that are unusual in such an encoding. I figured it was possible you did base64 with a rotated set of characters, so I decided to check if the number of unique characters was at most 64. I found that the number of unique characters was 95, ruling out this hypothesis.

I thought a bit about where the number 95 might come from. I figured it was probably about the number of printable ASCII characters, so I looked at the ASCII Wikipedia page, which confirmed there were exactly 95 printable characters.

Given this, my first hypothesis was some kind of character substitution cipher. From a quick glance at the text, it didn't look like a simple rotation (62/95 ASCII characters are letters or numbers, in contiguous ranges. If it was a simple rotation, you would expect to see longer runs of these).

I decided to compute all the character frequencies so I could compare them to character frequencies in natural text. I noticed that the most common character didn't appear all that often (I believe 14 times). That's way too few for a character substitution cypher: a normal English paragraph of this length would consist of 100+ words, so you would expect whatever character mapped to the space character to occur 100+ times.

I figured you wouldn't post a puzzle that was highly unlikely to be solved, so I it couldn't be that much more complicated than a substitution cypher. The first thing that came to mind at this point was adjacent character differences. I didn't have any particular justification for this, other than the fact that this could in principle change the distribution of tokens, which was currently too flat. I noticed that the character differences indeed had a much more reasonable distribution, so I printed all 95 rotations. Two of these were mostly legible. (One of them was of course off by 32, and I actually saw this one first, which confused me for a bit.) One of them was the solution.
--

Thanks for creating this market! I'm slightly annoyed someone already did the more profitable yes-bet, but it was still fun to figure out.

@jskf In hindsight, going from the observation of 95 unique characters to the substitution cipher hypothesis the way I did was a reasoning error: it would be kind of odd for your plaintext to include every printable ASCII character.

wait this market has much clearer resolution criteria for anyone who was too confused by that one!