> For the complete documentation index, see [llms.txt](https://references.everstrike.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://references.everstrike.io/ai/faq.md).

# FAQ

### Setup and risk

#### Which confidence threshold should I use?

Start at **90%** minimum confidence. Lower thresholds, such as 85% or 80%, can create more trades but may increase incorrect decisions.

See [Strategies](/ai/strategies.md#min-confidence) for common confidence settings.

#### Which order quantity should I use?

Start with a low order quantity, such as **$10**, or test the strategy on Everstrike Testnet. Validate events before increasing the quantity.

#### Which timeframe should I use?

Choose a timeframe that matches your trading horizon. Lower timeframes usually produce more frequent trades and shorter holding periods.

#### How can I manage risk with strategies?

Add a maximum position size to your prompt. The AI respects position limits that you provide.

For example:

> Buy the dip on BTC. Maximum position size: $3,000.

Also start with a low order quantity while you validate the strategy.

#### Can the AI close my position?

Yes. Describe the conditions for closing the position in your prompt.

For example:

> Buy the dip on BTC. Close the position when the dip is over.

You can also ask the AI to maintain a target position size.

### Strategy limits and behavior

#### How many strategies can I run simultaneously?

You can run up to **10 strategies** simultaneously.

#### How many markets can my strategies watch?

Each strategy can select up to **10 markets**. With 10 active strategies, AI can watch up to **100 markets**.

#### Does the strategy nickname affect trading?

No. The nickname is for display and for associating events with strategies.

#### When does a strategy start producing events?

Every model and evaluation frequency produces an event within 20 seconds of strategy creation. The event can lead to a trade if confidence exceeds your minimum threshold.

Track events in the frontend or through the API.

#### Why has my strategy stopped producing events?

If an active strategy misses its evaluation frequency, its model may not process the prompt quickly enough. Choose a more powerful model or simplify the prompt.

#### Can the AI maintain a grid of limit orders?

Yes. The AI can maintain up to **10 limit orders** per strategy, including for grid trading.

### Prompts and AI limits

#### What data can I reference in a prompt?

See [Supported Data](/ai/supported-data.md). The AI cannot access data outside that list and cannot perform visual chart analysis.

#### Which languages can I use in a prompt?

Use any commonly used natural language, including English, Spanish, and Norwegian. English usually produces the best results.

#### Can I receive reasoning in my native language?

Yes. Specify your preferred language in the prompt. The default reasoning language is English.

#### How can I minimize AI errors?

Use a minimum confidence of at least 90%, set temperature to 0, and disable discretion.

With these settings, the hallucination rate is significantly below 0.1%.

### Costs and credits

#### How much does a strategy cost?

The default GPT-4.1-mini model costs **$0.00015 per event**.

At a one-second evaluation frequency, one market creates 3,600 events per hour. That costs **$0.54 per hour**. Costs scale with every selected market.

See [strategy costs](/ai/strategies.md#strategy-costs) for model and frequency comparisons.

#### Do paused strategies consume credits?

No. Paused strategies do not generate events, trade, or consume credits.

#### How can I add more AI credits?

Fund your account with at least **$5**, or the equivalent in USDC or USDT. Strategy expenses are then deducted from your USD balance.

See [Credits](/ai/credits.md) for grants, replenishment, and funding details.

### Frequency

#### Is the AI suitable for high-frequency trading?

No. The fastest evaluation frequency is one second. This supports low- and mid-frequency strategies, not high-frequency trading.

Future releases may add shorter frequencies, such as 100 ms and 10 ms.