Meet Aidan Ronan — Founder
Behind the foundation,
a single conviction.
My name is Aidan Ronan. I didn't come to finance through a traditional path. From a young age I was building and running a consumer brand managing cash flow, international contracts, fulfillment, and margins. What that experience did was expose a mindset. An obsession with risk instead of reward.
What I learned running a business had nothing to do with the products. It had everything to do with structure. You can't afford guesswork when real obligations are on the line. Assumptions are the most expensive thing any operation can carry.
When it was time to move forward, I locked onto what I knew. Markets. Over years of direct experience and deliberate refinement, what started as a simpler operation evolved into something more considered. The volatility of early decisions made it clear to me that the durability I sought required a different foundation. It required systems built from logic, not discretion. That process became Bayfall Morrigan.
Bayfall Morrigan was not built in spare hours. There were no spare hours. At some point, this stopped being something I did and became something I lived. The transition was invisible.
Bayfall Morrigan is built around one conviction: that the most dangerous thing in any system is an assumption you haven't tested. We don't chase outcomes. We build structures designed to hold across changing conditions.
I am not building Bayfall Morrigan for what it is today. I am building it for the people who come after me who will never know what the beginning looked like but will inherit what the beginning made possible.
That is what this is for.
Published Writing — LinkedIn
On markets, methodology, and the assumptions underneath both.
Model Risk & Epistemology
For most of human history, you could not look inside another person's skull and check their competence. You read it off their outputs. Complexity was a proxy for skill, and for a very long time it was a good one. The two traveled together for so long that we stopped treating complexity as evidence and started treating it as competence itself. The problem with this is that proxies drift.
Read on LinkedIn
Bayesian Inference & MCMC
Most people let MCMC define their model, treating whatever the posterior finds as the answer. I use it the opposite way: feeding the sampler the exact cases my model was already built to capture, so it can only confirm what matters, never decide it. That looks like an overfitting machine on paper, but it isn't, because the sampler is never given the authority to define anything.
Read on LinkedIn
Structure-First Methodology
Most quantitative researchers are using machine learning the same way an eight year old uses laser eyes: impressive until nobody tells him when to stop. This article is about what separates a system you can maintain from one you can only replace, and why the difference has nothing to do with how much data you have.
Read on LinkedIn