Trading Models API

This endpoint returns the currently dominant trading models for a symbol.

Use this endpoint when you want to rank, inspect, or integrate the strongest currently available model configurations for a symbol.

Endpoint

GET https://api.darwintiq.com/v1/models

Query Parameters

Parameter	Required	Type	Description
`symbol`	Yes	string	Market symbol, for example `EURUSD`, `USDJPY`, `XAUUSD`, `SP500`, or `DAX`.
`sort`	No	string	Sort order: `fitness` (default), `latest`, or `entryMove`.
`limit`	No	integer	Number of trading models to return. Default is `5`, maximum is `50`.
`entryType`	No	string	Restricts results to a specific entry logic type.
`entryTypes`	No	string	Compatibility alias for `entryType`.

How to Use It

A typical client flow is:

Request the endpoint with a symbol.
Sort the results according to your use case.
Read the top-ranked item or iterate over the returned list.
Use the nested model blocks and performance fields in your UI, automation, or analytics pipeline.

Example Request

curl -H "Authorization: Bearer [YOUR_TOKEN]" \
  "https://api.darwintiq.com/v1/models?symbol=EURUSD&sort=fitness"

Response Example

{
  "status": "success",
  "symbol": "EURUSD",
  "trading_models": [
    {
      "id": "abc123",
      "timeframe": "H1",
      "created": "2026-02-10 09:20:00",
      "tested": "2026-02-10 13:20:00",

      "fitness": 74.6,
      "stabilityScore": 72,
      "entryMovePips": -6.4,
      "entryPrice": 1.08742,
      "currentPrice": 1.08678,

      "entryLogic": {
        "type": "pullback_continuation",
        "class": "PullbackContinuation",
        "description": "Trend continuation after a controlled pullback.",
        "params": {}
      },
      "positionManager": {
        "type": "absolute",
        "class": "Absolute",
        "description": "Fixed stop and trailing stop.",
        "params": {},
        "initialStopPips": 20,
        "trailingStopPips": 10,
        "takeProfitPips": 40
      },
      "regimeFilter": {
        "type": "rsi_band",
        "class": "RsiBand",
        "description": "Mean-reversion gate using RSI threshold bands.",
        "params": {}
      },

      "lastEntry": {
        "type": "BUY",
        "time": "2026-02-10 10:00:00",
        "timeElapsed": "20 minutes ago"
      },

      "totalPips": 132,
      "pipsPerTrade": 3.1,
      "tradesWon": 14,
      "tradesLost": 7,
      "pipsWon": 196,
      "pipsLost": 64,
      "maxDrawdown": 12.4,
      "expectedValue": 0.62,
      "profitFactor": 1.74,
      "sharpe": 0.38,
      "sortino": 0.56,
      "calmar": 1.92,
      "avgDrawdown": 4.7,
      "exposure": 0.41,
      "returnStability": 0.78,
      "testPeriod": { "from": "2026-02-10 09:20:00", "to": "2026-02-10 13:20:00" },
      "analytics": {
        "stability": {
          "score": 72,
          "returnStability": 0.78,
          "walkForward": {
            "enabled": true,
            "sufficient_data": true,
            "fold_count": 3,
            "coverage": 1,
            "consistency": 0.91,
            "train_score": 0.74,
            "test_score": 0.7,
            "multiplier": 0.97,
            "reason": "ok",
            "folds": [
              { "index": 0, "effective_trades": 8, "score": 0.76, "profit_pips": 18.4 },
              { "index": 1, "effective_trades": 7, "score": 0.72, "profit_pips": 15.1 },
              { "index": 2, "effective_trades": 6, "score": 0.7, "profit_pips": 13.3 }
            ]
          }
        }
      }
    }
  ],
  "timeSemantics": {
    "version": 1,
    "fieldBases": {
      "trading_models[].lastEntry.time": "BROKER_NY_CLOSE",
      "trading_models[].lastEntry.timeElapsed": "RELATIVE",
      "trading_models[].testPeriod.from": "BROKER_NY_CLOSE",
      "trading_models[].testPeriod.to": "BROKER_NY_CLOSE",
      "trading_models[].created": "RELATIVE",
      "trading_models[].tested": "RELATIVE"
    }
  }
}

Response Field Reference

Field	Type	Description
`status`	string	`success` when the request succeeds.
`symbol`	string	Requested symbol (e.g., `EURUSD`).
`trading_models`	array	List of trading model objects.
`trading_models[].id`	string	Stable model identifier.
`trading_models[].timeframe`	string	Model timeframe.
`trading_models[].created`	string	Humanized relative creation time label (for example `2h ago`).
`trading_models[].tested`	string	Humanized relative backtest time label (for example `15m ago`).
`trading_models[].fitness`	number	Fitness score (higher = better).
`trading_models[].stabilityScore`	number \| null	Drift-based stability score (0-100).
`trading_models[].entryMovePips`	number \| null	Signed pip move since last entry.
`trading_models[].entryPrice`	number \| null	Entry price of last signal.
`trading_models[].currentPrice`	number \| null	Current price used for move calculation.
`trading_models[].entryLogic`	object	Structured entry logic block.
`trading_models[].positionManager`	object	Structured position manager block.
`trading_models[].regimeFilter`	object	Structured regime filter block. Returned models use explicit filter logic rather than an always-on no-op filter.
`trading_models[].lastEntry`	object	Last entry time and direction.
`trading_models[].lastEntry.time`	string	Last entry timestamp in broker chart time (`BROKER_NY_CLOSE`).
`trading_models[].lastEntry.timeElapsed`	string	Humanized relative time label (`RELATIVE`).
`trading_models[].totalPips`	number	Total net pips in test window.
`trading_models[].pipsPerTrade`	number	Average pips per trade.
`trading_models[].tradesWon`	number	Number of winning trades.
`trading_models[].tradesLost`	number	Number of losing trades.
`trading_models[].pipsWon`	number	Total pips from wins.
`trading_models[].pipsLost`	number	Total pips from losses.
`trading_models[].maxDrawdown`	number	Max drawdown (pips).
`trading_models[].expectedValue`	number	Expected value (pips).
`trading_models[].profitFactor`	number \| null	Gross wins divided by gross losses.
`trading_models[].sharpe`	number \| null	Return/volatility ratio (trade-return based).
`trading_models[].sortino`	number \| null	Return/downside-volatility ratio.
`trading_models[].calmar`	number \| null	Net return divided by max equity drawdown.
`trading_models[].avgDrawdown`	number	Average drawdown (pips).
`trading_models[].exposure`	number	Time in market ratio (0..1).
`trading_models[].returnStability`	number \| null	Smoothness of equity trend (R^2, 0..1).
`trading_models[].testPeriod`	object	Backtest window (`from`/`to`).
`trading_models[].testPeriod.from`	string	Backtest window start in broker chart time (`BROKER_NY_CLOSE`).
`trading_models[].testPeriod.to`	string	Backtest window end in broker chart time (`BROKER_NY_CLOSE`).
`trading_models[].analytics`	object	Personality + stability analytics.
`trading_models[].analytics.stability`	object \| null	Stability metrics (stdDev, drift, MI).
`trading_models[].analytics.stability.score`	number \| null	Stability score (0-100), derived from drift metrics.
`trading_models[].analytics.stability.stdDev`	number \| null	Standard deviation of trade outcomes (pips).
`trading_models[].analytics.stability.drift`	object \| null	Drift metrics across features.
`trading_models[].analytics.stability.drift.profit_pips`	object \| null	Drift on profit per trade.
`trading_models[].analytics.stability.drift.rr`	object \| null	Drift on reward-to-risk distribution.
`trading_models[].analytics.stability.drift.holding_time`	object \| null	Drift on holding time distribution.
`trading_models[].analytics.stability.drift.stop_loss_pips`	object \| null	Drift on stop-loss sizing.
`trading_models[].analytics.stability.drift.*.js`	number \| null	Jensen-Shannon divergence.
`trading_models[].analytics.stability.drift.*.wasserstein`	number \| null	Wasserstein distance.
`trading_models[].analytics.stability.drift.*.ks`	number \| null	KS statistic.
`trading_models[].analytics.stability.drift.*.psi`	number \| null	Population Stability Index.
`trading_models[].analytics.stability.mi`	object \| null	Mutual information metrics.
`trading_models[].analytics.stability.mi.winloss_profit_pips`	number \| null	MI between win/loss and profit distribution.
`trading_models[].analytics.stability.mi.winloss_holding_time`	number \| null	MI between win/loss and holding time.
`trading_models[].analytics.stability.mi.winloss_rr`	number \| null	MI between win/loss and reward-to-risk.
`trading_models[].analytics.stability.mi.winloss_stop_loss_pips`	number \| null	MI between win/loss and stop-loss sizing.
`trading_models[].analytics.stability.driftWindow`	object \| null	Drift window meta.
`trading_models[].analytics.stability.driftWindow.split`	string \| null	Split strategy (for example `first_half_vs_second_half`).
`trading_models[].analytics.stability.driftWindow.start`	string \| null	Backtest start time.
`trading_models[].analytics.stability.driftWindow.mid`	string \| null	Midpoint time used for split.
`trading_models[].analytics.stability.driftWindow.end`	string \| null	Backtest end time.
`trading_models[].analytics.stability.driftWindow.min_trades_per_half`	number \| null	Minimum trades per half for drift calculation.
`trading_models[].analytics.stability.driftWindow.insufficient_trades`	boolean \| null	True if drift could not be computed.
`trading_models[].analytics.stability.driftWindow.pre_trades`	number \| null	Trades in first half.
`trading_models[].analytics.stability.driftWindow.post_trades`	number \| null	Trades in second half.
`trading_models[].analytics.stability.walkForward`	object \| null	Walk-forward validation payload for out-of-sample fold consistency.
`trading_models[].analytics.stability.walkForward.fold_count`	number	Number of configured folds in the current walk-forward run.
`trading_models[].analytics.stability.walkForward.coverage`	number \| null	Share of folds that produced usable scores (`0..1`).
`trading_models[].analytics.stability.walkForward.consistency`	number \| null	Similarity of fold scores (`0..1`), higher is steadier.
`trading_models[].analytics.stability.walkForward.train_score`	number \| null	Average score of the earlier reference folds.
`trading_models[].analytics.stability.walkForward.test_score`	number \| null	Score of the latest evaluated fold.
`trading_models[].analytics.stability.walkForward.multiplier`	number \| null	Fitness multiplier derived from walk-forward validation.
`trading_models[].analytics.stability.walkForward.reason`	string \| null	Status/reason such as `ok` or `insufficient_trades`.
`trading_models[].analytics.stability.walkForward.folds`	array	Per-fold diagnostics.
`trading_models[].analytics.stability.walkForward.folds[].index`	number	Zero-based fold index.
`trading_models[].analytics.stability.walkForward.folds[].effective_trades`	number \| null	Number of winning + losing trades in that fold.
`trading_models[].analytics.stability.walkForward.folds[].profit_pips`	number \| null	Net pips generated inside that fold.
`trading_models[].analytics.stability.walkForward.folds[].score`	number \| null	Fold quality score used for train/latest comparison.
`timeSemantics`	object \| null	Optional timestamp semantics metadata for interpreting timestamp fields.
`timeSemantics.version`	number	Metadata version (currently `1`).
`timeSemantics.fieldBases`	object	Map of field paths to time basis identifiers.
`timeSemantics.notes`	array \| null	Optional human-readable notes about timestamp semantics.

Core Ranking Fields

fitness: overall ranking score for recent performance and consistency.
stabilityScore: drift-based stability estimate. Higher values indicate more stable behavior.
entryMovePips: signed pip move since the last entry, measured at request time.
entryPrice and currentPrice: raw price context used for the move calculation.
Friendly symbol aliases are normalized before the upstream request, for example DAX -> GDAXI and SPX500 -> SP500.

Timestamp Semantics

BROKER_NY_CLOSE means Darwinex/MT4 chart time aligned to New York close, typically UTC+2 in winter and UTC+3 in summer.
RELATIVE means a display label such as 5 minutes ago, not an absolute timestamp.
If timeSemantics is present, prefer it over hard-coded assumptions in client code.

Stability Metrics

stdDev: dispersion of trade outcomes in the backtest window.
Drift metrics (js, wasserstein, ks, psi): distribution shift between the first and second half of the test window.
Mutual information (mi): how informative certain distributions are about win/loss outcomes.
profitFactor, sharpe, sortino, calmar: risk-adjusted performance metrics.
avgDrawdown, exposure, returnStability: drawdown pressure, time-in-market, and equity smoothness context.
analytics.stability.walkForward: recent out-of-sample validation. Compare train_score vs. test_score and inspect multiplier for fitness drag or confirmation.

Errors

Status	Meaning
`400`	Invalid symbol or query parameter.
`401`	Missing or invalid bearer token.
`403`	Valid token, but no access to the requested symbol or subscription scope.
`429`	Rate limit exceeded.
`500`	Internal or upstream server error.

Repository

Public example code is available here:

https://github.com/darwintIQ/API