Risk Engine: Stop-Loss and Profit Lock Explained

How the Stakao risk engine sizes each position, sets an adaptive stop-loss that only moves up, locks in gains and keeps idle TAO on Root. Limits included.

Updated 2026-10-13

01

The AI suggests, the risk engine decides

Choosing subnets and deciding trades are two separate jobs, done by two separate parts of Stakao.

The AI agent, the quant score or the basket rule only suggests subnets. It cannot buy, sell or touch your keys.

A rules-based risk engine decides what to buy, how much and when to sell. The same rules apply to every wallet.

The part of Stakao that decides holds no keys.

02

Which strategies use the risk engine

The eight Premium strategies. Auto Validators and Smart DCA do not use it.

  • Premium: AI Agent, Quant Core, Top 10 Market Cap, Top 20 Market Cap, Compute, Inference / Training, Agents and Market Intelligence.
  • Free: Auto Validators and Smart DCA have no stop-loss and no profit lock. You choose the subnets and when to leave.
03

How big each position is

Each position is sized so that about 1% of your wallet is at risk.

Entries are spread over time, and the engine can decide not to buy at all.

Sizing each position by the risk it carries keeps one bad position from hurting the whole wallet.

04

How the stop-loss works

Every position gets a stop-loss that adapts to the subnet’s volatility and only ever moves up.

A stop-loss is a price level below your entry. If the subnet token falls to that level, the position is sold.

Because the level adapts to each subnet’s own volatility, a fast-moving subnet is not treated like a quiet one. Because it only ever moves up, the engine never gives back protection it has already earned.

05

How the profit lock works

As a position gains, its stop rises to lock in part of the gain.

The stop starts below your entry price. As the position gains, the stop climbs with it, so a sale that ends the position happens at a higher price and a part of the gain is kept.

06

Watched around the clock

Positions are watched around the clock, and a position is sold when its stop is hit.

You do not need to watch the market. Your dashboard shows each position with its entry, its stop and its profit protection, and any operation that is still in progress.

07

Idle TAO stays on Root

TAO that is not in a position stays on Root, where it keeps earning staking rewards.

Root is the part of Bittensor where TAO is staked directly, without being swapped into a subnet token. The Root and fees guide explains how its rewards work.

08

Where the protection ends

A stop-loss limits losses but cannot remove them: in a thin market, a sale can fill below the stop price.

  • Subnet tokens can move fast. When few buyers are around, a sale can fill below the stop price.
  • Staking and unstaking go through a pool, so a large trade moves the price a little. This is called price impact.
  • The risk engine limits the loss on each position. It does not predict the market, and it does not promise a return.
  • Staking and subnet tokens carry risk, and you can lose part of what you stake.