EARLY MORNING BANK
FEBRUARY 2026
Not luck. Structure.
A curious coincidence that wasn't.
Flying FSD->MSP->DCA on a CRJ, I landed at Gate G3 and departed from Gate G3. The same gate. At a hub airport. On a connecting itinerary.
Most people would call it lucky and move on. Instead: what is the actual probability of this? And why is the answer not 1/K where K is the number of CRJ-compatible gates?
"A seemingly trivial passenger observation reveals the structure of a distributed, real-time, stochastic optimization system operating near equilibrium. And sometimes, the system remembers you."
That question became Episode 01 - a conditional probability decomposition across gate selection, temporal feasibility, and assignment stability. With an entropy interpretation. And a loyalty-tier extension that shows a Diamond flyer sees 50x better odds than a base-tier traveler.
Skyhounder is what happens when a graph theorist gets stranded in a hub airport and starts asking the right questions.
The Series
ONGOING - OPEN SOURCEThe math is the point.
Not decoration.
Every episode is built around a real formulation - not hand-waving. The gate symmetry result emerges from a multiplicative decomposition under conditional independence.
GitHub Pages lets us render it properly. No LinkedIn. No compromises. KaTeX inline with live NetworkX computations.
P(A(e2) = g | e1, RJ, W)
. P(Tg <= D)
. P(no reassignment | W)
~= 1/K . 0.90 . 0.90
// K=15 -> P(S) ~= 5.4%
// Not luck. Constrained optimization.