The asset selection problem
One of the biggest challenges in automated trading isn't execution. It's deciding what to trade in the first place.
Most discussions around trading automation focus on execution.
How to enter. How to exit. How to manage positions. How to optimize parameters.
Those questions are important. But they hide another problem that became increasingly obvious while building Oblivion.
What should the system trade in the first place?
The assumption I started with
Like many traders, I initially selected assets manually.
Choose a few cryptocurrencies. Configure the system. Let it run.
At first, this seemed perfectly reasonable. Then reality started creating problems.
Markets don't stay the same
The crypto market evolves constantly.
Assets gain attention. Assets lose relevance. New narratives emerge. Liquidity shifts. Volatility changes.
A cryptocurrency that looked attractive months ago may no longer be interesting today.
Yet many automated systems continue trading the same list indefinitely. The market changes. The configuration does not.
Automation creates a new challenge
Manual traders naturally adapt. They observe, they read and they react.
Automation removes much of that day-to-day involvement. That raises an important question.
If a human is no longer actively selecting assets every day, who is performing that task?
The answer cannot simply be "nobody."
Good execution doesn't fix bad selection
A system can execute flawlessly. Orders can be perfect. Risk management can be solid. Infrastructure can be reliable.
And the overall outcome can still disappoint if the underlying assets are poor candidates. Execution quality cannot fully compensate for weak asset selection.
Looking beyond individual assets
Over time, I became less interested in asking:
"Is this a good asset?"
And more interested in asking:
"How should the system identify opportunities across the entire market?"
The second question is much harder.
But it is also much more scalable. Instead of focusing on individual choices, the objective becomes designing a process.
A framework. A method. Something capable of adapting as market conditions evolve.
A problem worth solving
The more I worked on automation, the more I realized that asset selection deserved to be treated as a first-class problem.
Not an optional feature. Not a manual task left to the user. A genuine challenge requiring its own logic and methodology.
Eventually, this realization would lead to the development of entirely new mechanisms inside Oblivion. But before building solutions, it was necessary to fully understand the problem.
And that problem turned out to be much larger than I initially expected.