# Adaptive Risk Optimization (Mempool- & Volatility-Aware LTVs)
Helios’s risk engine is **fully algorithmic**: every Bitcoin block (\~10 min) it re-solves a convex optimization problem that chooses current **max-LTV** and **liquidation-bonus (LP)** values. No human flips a switch, no opaque AI—just a provably optimal LP model anchored to two on-chain data feeds:
| Input feed | Sample frequency | Why it matters |
| -------------------------------------------------------------------------------------------------------------------------------------- | ---------------- | ------------------------------------------------------------ |
| **Mempool congestion (M)** – median sats/[vB](https://learnmeabitcoin.com/technical/transaction/size/#vbytes) for next-block inclusion | Every block | Longer settlement time → greater liquidation delay risk |
| **Realized BTC volatility (σ)** – 1 h and 24 h windows | Every block | Larger price swings → larger potential collateral draw-downs |
**Formal objective**
Minimize **Expected Bad Debt (EBD)** subject to the hard constraint
for every active loan, where is collateral value after a worst-case liquidation delay. This caps the protocol-wide probability of under-collateralisation at **< 0.1 %**, equivalent to a 99.9 % one-block VaR.
If rising M or σ would push EBD above that ceiling, the solver **lowers max-LTV** and/or **raises LP** until the inequality is satisfied.
**Control loop & publication cadence**
1. **Oracle pull** – Fetch M, σ, order-book depth.
2. **Optimize (off-chain)** – Linear program returns LTV\*, LP\*.
3. **On-chain commit** – The algorithm writes the new limits into the contract’s state, and the updated rules apply automatically from the next block.
4. **Guard-band** – LTV cannot move >2 percentage points per block in calm markets, guaranteeing borrowers a minimum reaction window.
**Why this matters**
* **Predictable safety bar** – The 0.1 % insolvency target is auditable and stress-tested; every new block re-validates it.
* **Market-aligned incentives** – Higher LP during congestion guarantees liquidator profitability despite higher fees, ensuring prompt liquidations.
* **Capital efficiency** – In low-risk regimes lenders enjoy up to 80 % utilization while the solvency bound is still met.
* **Instant throttling** – During 2020-style mempool jams the engine would have cut LTV to 50 % within two blocks, preventing MakerDAO-type shortfalls.
Helios thus couples **Bitcoin settlement finality** with a risk framework that is as transparent as a margin rulebook and as fast as the chain itself.
All these adjustments happen algorithmically and frequently (potentially every block or every few blocks) as conditions evolve. This is a major innovation – rather than waiting for a governance vote after a market crash to adjust risk parameters (which is what happened in some DeFi incidents), Helios acts *preemptively and continuously*. It’s as if Helios has an embedded risk officer tweaking settings in real-time to keep the system safe yet efficient. This dynamic approach aims to maximize capital efficiency **without sacrificing solvency**.
From a due diligence perspective, this adaptive model is valuable because it reduces the reliance on manual intervention and can be stress-tested. It’s transparent (the rules and inputs are known) and can respond faster than human governance to emerging risks. For institutional partners, the fact that Helios is self-regulating to maintain a healthy risk profile provides confidence that extreme scenarios are being handled in a principled way.
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://helios-finance.gitbook.io/helios-finance/risk-framework/quickstart-1.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.