Quick definition
Treeova's two-tier risk protection model combining deterministic Standard guardrails with a context-modulated Adaptive trailing tier. Agents pull levers; platform code performs all risk arithmetic.
The two-tier model
The Standard tier is fully deterministic: fixed stop percentages, fixed profit targets, no surprises. The Adaptive tier sits on top and can tighten trails as a position moves into profit, widen them across high-IV regimes, or lock in gains as Theta accelerates near expiration. Standard always wins ties — Adaptive can never loosen a guard that Standard set.
Why agents don't do the math themselves
Treeova deliberately keeps risk arithmetic in platform code, not in agent prompts. Agents pull labeled levers ("throttle", "trail aggressiveness", "DTE divisor") and the engine translates them into hard numbers. This eliminates the entire class of bugs where an LLM hallucinates a stop price or forgets to account for contract multiplier.
Where it integrates
Every fill — paper or live — passes through risk assignment before becoming a managed position. The engine then re-evaluates on every monitor tick using the agent's current conviction score from Arch-AGI, the position's unrealized P&L, time-to-expiry, and broader market regime data from the Market Intelligence Super-Swarm.
Go deeper
The full technical methodology behind Adaptive Risk Engine is documented in the Treeova whitepaper series.
Read the Adaptive Risk Engine whitepaper →