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

# Strategy Events: Review AI Decisions and Track Performance

> Strategy events log each AI evaluation: action, confidence, reasoning, and credits used. Validate and refine strategies before scaling up.

Every time an active strategy evaluates a market, Everstrike records the result as an event. New events appear in real time on the AI dashboard and the AI activity page, and selecting an event opens its full details. Events capture the full picture of each AI decision: what action the AI chose, how confident it was, the reasoning behind that choice, whether an order was submitted, and how many credits the evaluation consumed. Use events to validate a new strategy before increasing its trade quantity or evaluation frequency. They are your primary tool for understanding how the AI interprets your prompt.

## Reviewing an Event

**1. Check the outcome**

Look at **Action**, **Amount**, and **Order Submitted** to confirm what the AI decided and whether it actually placed a trade. A `HOLD` action with **Order Submitted** set to `false` means the AI evaluated the market but did not meet the confidence threshold.

**2. Compare confidence and reasoning with your prompt**

Read the **Confidence** score and **Reasoning** field together. If the AI is consistently uncertain or its reasoning doesn't reflect your intent, revise the prompt to be more specific about entry and exit conditions.

**3. Track the strategy over time**

The **Statistics** tab on the AI dashboard sums a strategy's events: decisions, confidence, credits used and uptime. Use it to spot trends that single events don't show.

**4. Review credit consumption**

Check **Credits Used**, **Prompt Tokens**, and **Completion Tokens** to understand the cost of each evaluation. If costs are higher than expected, consider enabling Cost Savings Mode, switching to a more economical model, or limiting the strategy's input data to what the prompt uses.

## Event Properties

Each event exposes the following properties:

| Property | Description |
| - | - |
| **Creation Time** | Time when the event was first recorded. |
| **Pair** | Market evaluated by the strategy. |
| **Action** | AI decision: `BUY`, `SELL`, `HOLD`, or `CLOSE`. |
| **Amount** | USD value of any orders submitted for the event. |
| **Price** | Market mark price at the time of evaluation. |
| **Confidence** | AI confidence in the decision, from 0% to 100%. |
| **Reasoning** | The reasoning that led the AI to select its action. |
| **Strategy** | Strategy that generated the event. |
| **Order Submitted** | Whether the event resulted in a submitted order. |
| **Forced Evaluation** | Whether the event came from a manually forced evaluation. |
| **Model** | AI model used for the evaluation. |
| **Prompt Tokens** | Tokens consumed by the prompt sent to the model. |
| **Completion Tokens** | Tokens consumed by the model's response. |
| **Discretion** | Whether the AI applied discretion during this evaluation. |
| **Cost Savings Mode** | Whether Cost Savings Mode was active. Reasoning output is minimal in this mode. |
| **Thinking Time** | Time the AI took to reach its decision. |
| **Credits Used** | Total AI credits consumed by the event. |

## Event Frequency & Credit Use

Evaluation frequency determines how quickly events accumulate. Every event consumes AI credits, which are deducted automatically: from your free credit balance before your first deposit, and from your USD balance after it. The [AI credits page](https://app.everstrike.io/app/ai-credits) shows how many events each strategy generated and what they used. Pause a strategy at any time to stop event generation, trading, and credit use. You can resume it whenever you're ready.

The estimates below apply to each selected market individually. Frequencies of 15 seconds or less can only use GPT-4.1-nano, so those rows use its price; 30 seconds and slower use **GPT-6 Luna**, the default model.

| Evaluation Frequency | Model | Events per Market per Day | Est. Credits per Market per Hour |
| - | - | - | - |
| 1s | GPT-4.1-nano | 86,400 | \$0.72 |
| 5s | GPT-4.1-nano | 17,280 | \$0.144 |
| 15s | GPT-4.1-nano | 5,760 | \$0.048 |
| 30s | GPT-6 Luna | 2,880 | \$0.06 |
| 1m | GPT-6 Luna | 1,440 | \$0.03 |
| 3m | GPT-6 Luna | 480 | \$0.01 |
| 5m | GPT-6 Luna | 288 | \$0.006 |
| 15m | GPT-6 Luna | 96 | \$0.002 |
| 30m | GPT-6 Luna | 48 | \$0.001 |
| 1h | GPT-6 Luna | 24 | \$0.0005 |

Each check also costs the model's price once for the strategy itself, on top of the per-market figures above: a one-market strategy costs twice the figure shown, a two-market strategy three times its per-market figure.

## Event Costs by Model

The credit cost per event varies by the AI model your strategy uses. Higher-capability models with more reasoning depth cost more per evaluation.

| Model | Cost per Event |
| - | - |
| **GPT-6 Luna** (default) | \$0.0005 |
| GPT-5.6 Luna | \$0.0005 |
| GPT-5.4-mini | \$0.0005 |
| GPT-5.4-nano | \$0.001 |
| GPT-4o-mini | \$0.0002 |
| GPT-4.1 | \$0.004 |
| GPT-4.1-mini | \$0.0005 |
| GPT-4.1-nano | \$0.0002 |
| GPT-5-mini-minimal-reasoning | \$0.0005 |
| GPT-5-nano-minimal-reasoning | \$0.0001 |
| GPT-5-mini-low-reasoning | \$0.001 |
| GPT-5-nano-low-reasoning | \$0.0002 |
| GPT-5-mini-medium-reasoning | \$0.0025 |
| GPT-5-nano-medium-reasoning | \$0.0005 |
| GPT-5-mini-high-reasoning | \$0.005 |
| GPT-5-nano-high-reasoning | \$0.001 |

See [Models](https://docs.everstrike.io/ai/models.md) for each model's typical speed and when to use it.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.