Get Completed Trades

Overview

The simulator.getCompletedTradesFromDB endpoint retrieves historical completed trades for a specific AI model from the database, along with calculated performance statistics. Unlike the in-memory simulator state, this endpoint queries persistent database records to provide historical trade analysis.

Request

Parameters

string

The AI model ID to filter trades by. If not provided or empty, returns an empty result set.Model ID corresponds to records in the Models table and links trades to specific AI agents.

number

default:"200"

Maximum number of trades to return. Must be a positive integer.

Default: 200
Maximum: 500 (hard cap enforced by the endpoint)
Used to prevent large result sets from impacting performance

Response

array

Array of completed trades, ordered by close date (most recent first).

string

Unique trade identifier in format {toolCallId}:{index}.

string

AI model ID that executed the trade.

string | null

Human-readable model name (e.g., "Apex", "Trendsurfer").

string

Trading pair symbol (e.g., "BTC-USD").

enum<string>

Trade direction: "LONG" or "SHORT".

number | null

Position notional value in USD: quantity * exitPrice.

number | null

Realized profit/loss from closing the position.

number | null

Leverage multiplier used when opening the position (if tracked).

number | null

AI model confidence level (0-1) when opening the position (if tracked).

string | null

ISO 8601 timestamp when the position was closed.

string

AI invocation ID that triggered the trade.

object

Aggregated performance statistics calculated from the returned trades.

number

Total number of completed trades.

number

Sum of all realized P&L values.

number | null

Average P&L per trade: totalRealized / tradeCount.null if no trades.

number | null

Mean leverage across all trades with non-null leverage values.null if no leverage data.

number | null

Median leverage across all trades with non-null leverage values.null if no leverage data.

number | null

Maximum leverage value observed across all trades.null if no leverage data.

array<number>

Array of all non-null leverage values (used for calculations).

number | null

Mean confidence level across all trades with non-null confidence values.null if no confidence data.

number | null

Median confidence level across all trades with non-null confidence values.null if no confidence data.

array<number>

Array of all non-null confidence values (used for calculations).

Code Example

import { useQuery } from "@tanstack/react-query";
import { orpc } from "@/server/orpc/client";

export function TradeHistory({ modelId }: { modelId: string }) {
  const { data, isLoading, error } = useQuery(
    orpc.simulator.getCompletedTradesFromDB.queryOptions({
      input: {
        modelId,
        limit: 100, // optional, defaults to 200
      },
    })
  );

  if (isLoading) return <div>Loading trades...</div>;
  if (error) return <div>Error: {error.message}</div>;
  if (!data) return null;

  const { trades, stats } = data;

  return (
    <div className="space-y-6">
      {/* Performance Summary */}
      <div className="grid grid-cols-4 gap-4">
        <div>
          <h3 className="text-sm text-gray-500">Total Trades</h3>
          <p className="text-2xl font-bold">{stats.tradeCount}</p>
        </div>
        <div>
          <h3 className="text-sm text-gray-500">Total P&L</h3>
          <p className={`text-2xl font-bold ${
            stats.totalRealized >= 0 ? "text-green-500" : "text-red-500"
          }`}>
            ${stats.totalRealized.toFixed(2)}
          </p>
        </div>
        <div>
          <h3 className="text-sm text-gray-500">Expectancy</h3>
          <p className="text-2xl">
            {stats.expectancy !== null 
              ? `$${stats.expectancy.toFixed(2)}` 
              : "N/A"}
          </p>
        </div>
        <div>
          <h3 className="text-sm text-gray-500">Avg Leverage</h3>
          <p className="text-2xl">
            {stats.averageLeverage !== null
              ? `${stats.averageLeverage.toFixed(2)}x`
              : "N/A"}
          </p>
        </div>
      </div>

      {/* Trade List */}
      <div>
        <h3 className="text-lg font-bold mb-4">Recent Trades</h3>
        <div className="space-y-2">
          {trades.map((trade) => (
            <div key={trade.id} className="border p-4 rounded">
              <div className="flex justify-between items-start">
                <div>
                  <span className="font-bold">{trade.symbol}</span>
                  <span className={`ml-2 px-2 py-1 rounded text-xs ${
                    trade.direction === "LONG" 
                      ? "bg-green-100 text-green-700" 
                      : "bg-red-100 text-red-700"
                  }`}>
                    {trade.direction}
                  </span>
                </div>
                <div className="text-right">
                  <p className={`font-bold ${
                    (trade.realizedPnl ?? 0) >= 0 
                      ? "text-green-500" 
                      : "text-red-500"
                  }`}>
                    {trade.realizedPnl !== null
                      ? `$${trade.realizedPnl.toFixed(2)}`
                      : "N/A"}
                  </p>
                  {trade.closedAt && (
                    <p className="text-xs text-gray-500">
                      {new Date(trade.closedAt).toLocaleString()}
                    </p>
                  )}
                </div>
              </div>
              <div className="mt-2 grid grid-cols-3 gap-2 text-sm text-gray-600">
                {trade.notional !== null && (
                  <div>Notional: ${trade.notional.toFixed(2)}</div>
                )}
                {trade.leverage !== null && (
                  <div>Leverage: {trade.leverage}x</div>
                )}
                {trade.confidence !== null && (
                  <div>Confidence: {(trade.confidence * 100).toFixed(0)}%</div>
                )}
              </div>
            </div>
          ))}
        </div>
      </div>
    </div>
  );
}

Data Source

This endpoint queries the database, not the in-memory simulator state:

Tool Calls Table

Trades are reconstructed from tool call records:

CLOSE_POSITION tool calls contain closed position data
CREATE_POSITION tool calls contain leverage and confidence metadata
Trades are linked via invocationId to match opens with closes

Metadata Structure

Tool call metadata is JSON containing:

// CLOSE_POSITION metadata
{
  "closedPositions": [
    {
      "symbol": "BTC-USD",
      "side": "LONG",
      "quantity": 0.5,
      "exitPrice": 45000,
      "realizedPnl": 1250.50,
      "closedAt": "2025-03-15T10:30:00Z"
    }
  ]
}

// CREATE_POSITION metadata
{
  "decisions": [
    {
      "symbol": "BTC-USD",
      "leverage": 3,
      "confidence": 0.85
    }
  ]
}

Matching Logic

Query all CLOSE_POSITION calls for the model (ordered by date)
Extract closed position records from metadata
Query all CREATE_POSITION calls for the same invocations
Build an index of leverage/confidence by symbol from CREATE calls
Match each closed position with its creation metadata

Statistics Calculation

Expectancy

Average profit per trade:

expectancy = totalRealized / tradeCount

Positive expectancy indicates profitable trading on average.

Leverage Statistics

Average: Mean of all non-null leverage values
Median: Middle value of sorted leverage values
Max: Highest leverage used across all trades

Only trades with tracked leverage are included in calculations.

Confidence Statistics

Average: Mean of all non-null confidence values (0-1)
Median: Middle value of sorted confidence values

Only trades with tracked confidence are included in calculations.

Null Handling

Statistics gracefully handle missing data:

Fields with null values are excluded from calculations
If no valid data exists, statistics return null
Empty arrays are returned when no values are available

Use Cases

Performance Analytics Dashboard

Display historical performance metrics for AI trading models:

const { data } = useQuery(
  orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId: "model-apex" }
  })
);

// Show win rate, expectancy, total P&L

Model Comparison

Compare performance across different AI models:

const apexTrades = useQuery(
  orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId: "apex-model-id" }
  })
);

const sovereignTrades = useQuery(
  orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId: "sovereign-model-id" }
  })
);

// Compare stats.expectancy, stats.averageLeverage, etc.

Trade Audit Trail

Review individual trade execution history:

const { data } = useQuery(
  orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId: "model-id", limit: 500 }
  })
);

// Export trades for analysis
const csv = data.trades.map(t => 
  `${t.symbol},${t.direction},${t.realizedPnl},${t.closedAt}`
).join('\n');

Leverage Risk Analysis

Analyze leverage usage patterns:

const { data } = useQuery(
  orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId: "model-id" }
  })
);

const { averageLeverage, maxLeverage, leverageValues } = data.stats;

// Visualize leverage distribution

Performance Considerations

Query Optimization

Database query uses indexed columns (modelId, toolCallType, createdAt)
Limit parameter caps result set size (max 500)
Results ordered by createdAt DESC for recent trades first

Metadata Parsing

JSON metadata is parsed on-the-fly:

Expect ~1-5ms parsing time per trade
Large result sets (500 trades) may take 100-500ms total
Consider pagination for very large datasets

Caching Strategy

// Cache for 5 minutes (trades don't change frequently)
const { data } = useQuery({
  ...orpc.simulator.getCompletedTradesFromDB.queryOptions({
    input: { modelId }
  }),
  staleTime: 5 * 60 * 1000, // 5 minutes
});

Error Handling

Scenario	Behavior
No `modelId` provided	Returns empty `trades: []` and zero stats
Invalid `modelId`	Returns empty result set (no error)
`limit` exceeds 500	Capped at 500 automatically
Database connection error	Throws error (caught by React Query)
Malformed JSON metadata	Skips invalid records, logs error

This endpoint queries historical database records, not the current simulator state. Trades shown here may not reflect positions currently open in the simulator.

Overview

Trading

Models & Invocations

Simulator

Analytics

Variants

Get Completed Trades

Overview

Request

Parameters

Response

Code Example

Data Source

Tool Calls Table

Metadata Structure

Matching Logic

Statistics Calculation

Expectancy

Leverage Statistics

Confidence Statistics

Null Handling

Use Cases

Performance Analytics Dashboard

Model Comparison

Trade Audit Trail

Leverage Risk Analysis

Performance Considerations

Query Optimization

Metadata Parsing

Caching Strategy

Error Handling

​Overview

​Request

​Parameters

​Response

​Code Example

​Data Source

​Tool Calls Table

​Metadata Structure

​Matching Logic

​Statistics Calculation

​Expectancy

​Leverage Statistics

​Confidence Statistics

​Null Handling

​Use Cases

​Performance Analytics Dashboard

​Model Comparison

​Trade Audit Trail

​Leverage Risk Analysis

​Performance Considerations

​Query Optimization

​Metadata Parsing

​Caching Strategy

​Error Handling

Overview

Request

Parameters

Response

Code Example

Data Source

Tool Calls Table

Metadata Structure

Matching Logic

Statistics Calculation

Expectancy

Leverage Statistics

Confidence Statistics

Null Handling

Use Cases

Performance Analytics Dashboard

Model Comparison

Trade Audit Trail

Leverage Risk Analysis

Performance Considerations

Query Optimization

Metadata Parsing

Caching Strategy

Error Handling