This is a post (hopefully eventually a short series of posts) detailing my analysis of the recently released D&D.Sci scenario by abstractapplic. I've decided to go through documenting what I do - if you intend to play the scenario yourself without help, you should do that before reading this. If you want to use this information in your solutions, go ahead.

Raw Data and Columns

I start by grabbing the raw data and saving it as a csv, then import it into Python and Excel to play with it. I'm not trying to do any particular analysis at this point, just trying to get familiar with the data and see if anything jumps out.

We'll go through the columns one by one, and make some charts in Excel for visualization^[1].

Longitude is relatively evenly-distributed from -180 to +180.

Latitude can vary from -90 to +90 but is bimodal, usually taking on values around 45 - we rarely land near the equator or near the poles.

And Shortitude and Deltitude look the same as Latitude:

This might be what you naturally expect from landing at a random point on a 4D sphere - certainly values near +-90 being less likely happens naturally, since those are 'the poles' and take up less area. I don't have a good enough intuition for this to know if values near 0 are also naturally rare once you add more dimensions, or if we're avoiding the 'equator' for some reason.

Strange Smell is often Somewhat present, but only rarely EXTREMELY present:

Air Tastes Like has a few different values, some common and some rare:

Feng Shui is generally Adequate, sometime bad, and very rarely good:

Weird Sounds has five possible values, of which we see up to three at a time:

Eerie Silence
Impossible Humming
Otherworldly Skittering
Unearthly Squelching
Unnatural Buzzing

Eerie Silence appears exactly when no other sounds appear.

Skittering and Buzzing never happen together (though they are the most common sounds separately).

Local Value of Pi is distributed in what looks like a very neat distribution:

3.141 + (1d41-1d41)/1000 isn't literally accurate, since we see more decimal figures than that, but it's probably a reasonable way of thinking about it?

Murphy's Constant has a very weird distribution:

I'm going to guess something like this:

MC is distributed with the same type of distribution we saw Pi have, with probability increasing linearly from 0 to ~5.5 and then decreasing from ~5.5 to ~11.
However, ou

...

NickSharp

NickSharp

NickSharp

A browser game about AI safety

NickSharp

NickSharp

NickSharp

A browser game about AI safety