Rules for building systems that can survive being live.
Hydra is built on the idea that automated trading should be judged not only by signal quality, but by how well the surrounding system constrains failure, preserves operator visibility, and resists escalation under stress.
Core position
Automation should not be treated as inherently trustworthy. Strategy logic can be intelligent and still behave badly in live conditions. Infrastructure exists to constrain that.
Hydra separates alpha, execution, and governance because those are different responsibilities. A system that mixes them into one opaque unit becomes harder to reason about, harder to supervise, and easier to break.
The purpose of doctrine is to keep the stack anchored when implementation details evolve. Tools change. Environments change. Principles should not drift every week.
01
Survivability > profitability
A system that cannot survive live conditions does not deserve scale.
02
Fail closed
When state becomes ambiguous, the safe default is restraint, not improvisation.
03
Failures allowed, escalation not
Losses and mistakes happen. Cascading operational breakdown should not.
04
Governance above conviction
Strong belief in a strategy is not a substitute for hard controls.