Pasted

Here is blcd/Blade/Will's spreadsheet that show the new The Chest strategy for mapping:
https://docs.google.com/spreadsheets/d/1VJqmqntVUVIzW5l6XGHELbdt8G06kvDMcbFOUQ8i0Kk/edit#gid=919737978
 this confirms adam's propaganda as true (which was to always go left and right instead of down on the chest).
here's my post-hoc rationalization of why it is true:
 - we know that each floor is a grid of 13x13
- we know that on the grid, the starting room is always on point (7, 7)
- thus, from the starting room, the boss must be:
 1) above
2) above-right
3) right
4) below-right
5) below
6) below-left
7) left
8) above-left
 - the boss being exactly in the 4 cardinal directions is extremely unlikely, so lets remove them for simplicity. so now remaining directions are:
 1) above-right
2) below-right
3) below-left
4) above-left
 - blcd/blade/will confirms that from his testing, the boss is equally likely to be in any of these 4 directions.
- now, we can calculate door percentages for each of these possibilities (assuming the starting room has 3 doors):
 1) above-right - the exit door is probably right (since top is impossible)
2) below-right - the exit door is 50/50 right or down
6) below-left  - the exit door is 50/50 left or down
8) above-left  - the exit door is probably left (since top is impossible)
 - we can add up the probabilities:
 1) left: 0.5 + 1 = 1.5
2) right: 1 + 0.5 = 1.5
3) down: 0.5 + 0.5 = 1
 - so this is a hypothesis of why down is less likely. these proportions don't exactly match what is being generated experimentally, but they could explain the discrepancy.