@VozMadridista Que Uruguay y Colombia hayan perdido sus respectivos partidos y que Portugal haya perdido también es robo de Argentina??. Abrazo Aurelio!
Systematic FX is not another asset in the traditional portfolio stack.
The point is not to claim alpha.
The point is structure.
Full post on LinkedIn →
https://t.co/KUmFz82pn5
Since 2022, the 60/40 hedge has become less reliable.
Bond-equity correlation has spent extended periods in positive territory, reducing the diversification investors relied on.
Full post on LinkedIn →
https://t.co/QAAuoBU3Hi
What causes most drawdowns in systematic trading? Poor signals? Usually not.
More often, exposure remains active after the conditions under which a signal was calibrated have already changed.
Full post on Linkedin →
https://t.co/i95A7eOSTH
Model decay is the actual problem, not the exception.
Most systems treat it as a bug. We treated it as the constraint that drove everything.
Full post on LinkedIn →
https://t.co/KgWNjMsSsd
Low equity correlation doesn’t tell you much if a strategy correlates with what’s already in your FX book.
Many systematic FX strategies ultimately derive a large part of their returns from carry or short-vol exposure.
Full post →
https://t.co/O1AFqfBrps
If your system can't classify why it stood aside, you have a governance gap you may not have noticed yet.
Full post on LinkedIn →
https://t.co/Km0ewqPpLS
The real problem isn't predicting the next move.
It's detecting when your own model has already gone stale.
Dynamic adaptation isn't an upgrade. It's the survival condition.
The edge in systematic FX isn't in predicting better. It's in unlearning faster.
That's what keeps the equity curve decorrelated.
More on LinkedIn →
https://t.co/nXKiZAPNrY
Every model has a quiet flaw: assuming the structure behind past data still holds today.
In FX, it breaks constantly.
An AI built for institutions must know when its own accuracy has expired.
Knowing when not to act is architecture.
Full post 👇
https://t.co/uXrXoDi0tZ
$9.6 trillion in daily FX volume. 89% of insttutons see ROI from AI.
But most systems stll can't explain their own decisions.
Regulators are done waiting.
The edge isn't speed. It's traceability.
Full post 👇
https://t.co/FBbbwHlrNG
El alfa no es acertar más. Es mantenerse sólido cuando el régimen cambia.
Lightstorm construye inteligencia adaptativa para instituciones que operan en mercados que no paran de moverse.
When AI meets real institutional capital, the constraint isn't generating signals.
It's standing behind them. What was observed. What was filtered.
What let a trade pass.
You're judged on how a system behaves when it doesn't work.
https://t.co/i8WCoCoMik
You can spend years building a system before you understand how it behaves.
The question isn't "does the model work?"
It's: can the organization stay coherent long enough to find out?
Full post on LinkedIn
https://t.co/Y575CMR0p7
The market hasn't broken. It's just operating under different rules.
The models that survive won't be the ones that predict better. They'll be the ones that know when to stop.
Read the full post here 👇
https://t.co/mzy5fwLab5
95% of hedge fund managers use generative AI.
Before: "Does AI work?"
Now: "Where does it create the strongest edge?"
Machine-assisted research is becoming core investing infrastructure.
Full post 👇
https://t.co/QbLg8soGZP
Trabajo con modelos predictivos cada día.
Y cada día me convenzo más de algo incómodo:
La predicción importa. Pero saber qué hacer cuando falla importa más.
El modelo cambia el mundo. Y el mundo cambia el modelo.
Más información en mi post de LinkedIn:
https://t.co/6O0jsPGqAh
In FX, most "alpha" is beta in disguise.
Separating them isn't a stats problem — it's a control problem.
Like sailing in open ocean: you don't control the wind. You learn to trim it.
Full post 👇
https://t.co/to0Cvqf6n2