> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trinigence.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How Clarifications Work

> When ATI asks follow-up questions, what it’s trying to resolve, and how to answer quickly.

## Why clarifications exist

A backtestable strategy requires specifics.

If a trading idea is missing something important, ATI will ask clarifying questions to ensure:

* the strategy is deterministic
* it matches your intent
* it can be backtested correctly

<Info>
  Clarifications are a safety feature, not friction.
</Info>

***

## Common reasons ATI asks questions

### Missing required structure

Examples:

* market or timeframe missing
* direction not specified (long/short)
* no exit rules provided

### Ambiguous language

Examples:

* “enter on strong momentum” (what defines strong?)
* “use a tight stop” (how tight?)
* “trade when trend is bullish” (which trend definition?)

### Conflicting instructions

Examples:

* “trade only weekdays” + “enter during weekend pump”
* “one position at a time” + “scale in multiple entries”

***

## How to answer clarifications

Best answers are:

* short
* explicit
* numeric where possible
* consistent with the original idea

Examples:

```text theme={null}
Use BTC/USDT, timeframe 1h.
Trade both long and short.
Exit long: TP 6%, SL 1%.
Exit short: TP 5%, SL 1%.
```

***

## Typical clarification formats

ATI may ask:

* to choose between options (A or B)
* to provide missing values (TP %, SL %, times)
* to confirm an assumption (e.g., candle-close evaluation)

<Steps>
  <Step title="Answer only what’s asked">
    Don’t rewrite the whole strategy unless you want to.
  </Step>

  <Step title="Use explicit numbers">
    Replace “tight” with “0.8%”, “short” with “15m”, etc.
  </Step>

  <Step title="Keep intent consistent">
    Your clarification should reinforce the original logic.
  </Step>
</Steps>

***

## If you don’t know the answer

If you’re unsure, you can say:

* “use standard defaults”
* “pick a conservative value”
* “optimize later”

ATI will either:

* apply safe defaults, or
* suggest a next step

<Card title="When not to optimize" icon="ban" href="/iteration-optimization/when-not-to-optimize" horizontal>
  Don’t over-tune early.
</Card>

<Note>
  Clarifications help ATI match your intent with deterministic rules.
</Note>