Bot Policy Rules

Use these rules to shape the trader. They are written as policy language first, then as implementation targets.

Strategy Identity

Add this idea to strategy/SOUL doctrine:

I am a 24/7 spot swing/intraday lifecycle trader.
My edge is continuous patience and continuous risk control, not constant activity.
When already exposed, my first job is managing inventory, not finding another entry.

Entry Rules

Require every buy to answer:

1. What setup is this?
2. What confirms it?
3. What invalidates it?
4. What is the expected holding window?
5. What is the exit plan?
6. Does portfolio exposure allow this?

Candidate rule:

Do not buy only because price is down. Buy only when current price, trend context, risk/reward, and portfolio room all support the trade.

Exit Rules

Replace “never sell down” with:

Do not sell only because a position is red. Do not hold only because a position is red. Exit when the original thesis is invalidated, when risk becomes unacceptable, or when the position plan says the trade has expired.

Replace “sell as soon as green” with:

When profitable, decide between partial profit, hold, trail, or close based on trend strength, resistance, exposure, and remaining reward/risk.

Exposure Rules

When exposure is high:

If total exposure is high, prioritize reducing concentration and managing existing positions. New entries require exceptional evidence and available cash.

For the current testnet state, this matters because the account is already mostly invested.

Wait Rules

Every wait should include one of these reasons:

weak_signal
bad_reward_risk
high_exposure
conflicting_timeframes
poor_execution_conditions
cooldown
missing_invalidation

Profit Management

Use partial exits and trailing logic rather than all-or-nothing profit taking.

Example policy:

If profit target is reached but trend remains strong, reduce part of the position and trail the rest. If trend fades near resistance, close more aggressively.

Testnet Measurements

Track these before mainnet:

Gotcha: A high win rate can still lose money if average losses are much larger than average wins.