last day (15 days later) » 

7:50 PM
The idea is to take the complement of the dual of the graph, see it's $(n+1)\times n$, and see that there is a left to right path in the graph iff there isn't a top to bottom path in the complement of the dual of the graph
This leads to $P(\text{path}) = P(\text{no path}) = 1/2$
 
Dual graph = graph made by the faces?
Isn't it $n\times(n-1)$?
 
Yeah kinda
Lemme do a drawing
Here's an example
Black are actual graph, blue complement of the graph, red dual graph and green complement of dual graph
 
I guess the edges on the left and right of the actual graph (and the top and bottom of the ghost one) don't matter
 
That's right
 
OK, I see
So, ignoring those, you're essentially turning all edges of the complement around 90 degrees
Or, perhaps, turning all edges 90 degrees and then flipping colors
And that's clearly a bijection then
Cool!
Like, 90 degrees around each edge's individual midpoint
 
8:07 PM
Yup, very short proof as promised :p
 

last day (15 days later) »