The setup here is a classic probability puzzle. Imagine a ball dropping through a triangular pegboard (a Galton board), bouncing left or right at each row of pegs until it lands in one of the bins at the bottom. Normally each bounce is independent and fair, producing the familiar bell-curve distribution of balls across bins. This paper changes the rules: after the very first bounce, each subsequent bounce tends to repeat the previous direction with probability p, making the ball's path "sticky" or correlated. The question, originally posed by Kagey, is: for what value of p does the middle bin collect exactly the same number of balls as the two outermost end bins? That is a surprising and non-obvious condition to ask for, since ordinarily the middle bin is the most likely destination.
The authors work out that the critical value of p corresponds to the ball changing direction roughly a logarithmic number of times relative to the total number of bounces. In other words, as the board gets taller and the number of bounces N grows, the "tie" value of p shifts in a very specific, slowly changing way. To make this precise, they connect the probability of landing in the middle bin to a classical mathematical object called a Bessel function, and they carefully control how large the error in that approximation can be. Crucially, their error bounds are explicit and uniform across all values of N, not just valid asymptotically for very large N.
This explicitness is the paper's main technical achievement. For boards with 175 or more bounces, they establish the tie interval analytically with guaranteed bounds. For smaller boards, they verify the answer exactly using integer arithmetic, and they double-check those computations with the formal proof assistant Lean, a software tool that certifies mathematical arguments machine-by-machine with no room for human error. The paper also extends the analysis to find the tie conditions for other bins besides the middle one, and examines how the picture changes when the board wraps around with periodic boundary conditions rather than having hard edges.