API reference

The backtest of a screener filter set

GET/track-recordKey required

Returns the record of the past trades of one family (or all five) whose facts at entry matched every filter: the TradingPal screener's Track record panel for the same filters, computed from the same stored past-trade ledger. filters takes the same predicate list as /scan. Filters that describe today rather than the trade date cannot cut past trades and come back under filters.left_out. edge=on counts only the trades the Edge Filter kept (tiers S to C). Under minimum_trades matched trades, track_record is null and note says why. unfiltered is the same family's record without the filters, for comparison. Needs a key on an account with Premium+.

Authentication

Needs a key: send Authorization: Bearer tp_live_… (see Authentication). Counted against the key's limits of 60 requests a minute and 10,000 a day.

Parameters

Name	In	Type	Required	Description
`family`	query	string (enum)	no	Keep one family: rising_wedge, falling_wedge, bullish_pennant, bearish_pennant, triangle. Omit for all five. Anything else answers 400 bad_request. One of `rising_wedge`, `falling_wedge`, `bullish_pennant`, `bearish_pennant`, `triangle`.
`filters`	query	string	no	The screener's own filters, URL-encoded: a JSON array of predicates, ANDed together; a kind may repeat (for example two moving-average lengths). Numeric kinds take op (<, <=, >, >=, =) and value, or op between with value and upper (inclusive), choice and text kinds take value, boolean kinds take value true or false, and the moving-average, performance and high/low kinds also take period. For example filters=[{"kind":"price","op":">","value":10},{"kind":"rsi","op":"<","value":40},{"kind":"asset_type","value":"stock"}]. Every predicate must match; a setup with no evidence for a kind never matches. The kinds, with what each takes and means, are the x-filters list on this operation (the reference page shows it as a table). A malformed list answers 400 bad_request with the reason.
`edge`	query	string (enum)	no	on: only the past trades the Edge Filter kept (tiers S to C); off: every trade. One of `off`, `on`.
`X-TradingPal-Client`	header	string	no	Optional name of the agent or script making the call. It is recorded in the account's usage log so the user can see which client made which calls; the User-Agent is recorded when it is absent.

Request

curl "https://api.tradingpal.io/api/v1/track-record?family=value&filters=value&edge=value" \
  -H "Authorization: Bearer $TRADINGPAL_API_KEY"

Response

JSON. The fields this route adds are below; every response also carries the envelope fields (api_version, as_of_session, data_version, source, disclaimer).

Field	Type	Description
`family`	string · nullable	The family asked for (rising_wedge, falling_wedge, bullish_pennant, bearish_pennant, triangle); null for all five.
`label`	string	The family's display name, or All patterns.
`edge`	object	The Edge Filter side the record was cut from.
`edge.requested`	string (enum)	off or on, as asked. One of `off`, `on`.
`edge.applied`	boolean	False when on was asked of a ledger that seals no Edge Filter tiers; the record is then every trade's.
`filters`	object	How the filters were used.
`filters.applied`	array of object	The predicates that cut the past trades, canonical.
`filters.left_out`	array of object	Predicates the record ignores, each with its reason: they describe today (market cap, security type, exchange, sector, country, phase, distance, the model's estimates) or the ledger holds no values for them.
`matched_trades`	integer	Past trades that matched every applied filter.
`too_few`	boolean	True under minimum_trades matched trades; track_record is then null.
`minimum_trades`	integer	The fewest matched trades a record is given for.
`track_record`	Track record	The record of the matched trades (scope filtered). Null when too_few.
`unfiltered`	Track record	The same family's record without the filters, on the same Edge Filter side, for comparison (scope unfiltered).
`note`	string · nullable	Why no win rate is given, when too_few.

Example response

Response

{
  "api_version": "v1",
  "data_version": "patterns-D-2026-11-12",
  "family": "falling_wedge",
  "label": "Falling wedge",
  "edge": {
    "requested": "off",
    "applied": true
  },
  "filters": {
    "applied": [
      {
        "kind": "min_rr",
        "op": ">=",
        "value": 2
      }
    ],
    "left_out": [
      {
        "kind": "market_cap",
        "op": ">",
        "value": 1000000000,
        "reason": "describes today, not the trade date, so it cannot cut past trades"
      }
    ]
  },
  "matched_trades": 1287,
  "too_few": false,
  "minimum_trades": 30,
  "track_record": {
    "scope": "filtered",
    "family": "falling_wedge",
    "symbol": null,
    "win_rate_pct": 54,
    "sample_size": 1287,
    "wins": 695,
    "losses": 592,
    "avg_win_pct": 8.1,
    "avg_loss_pct": -4.2,
    "median_win_pct": 6.3,
    "avg_trade_pct": 2.4,
    "avg_r": 0.41,
    "expected_r": 0.41,
    "profit_factor": 1.61,
    "avg_hold_bars": 11.2,
    "backtest_range": [
      "2006-09-25",
      "2026-09-25"
    ],
    "measured": "Every detector-confirmed breakout … open trades are excluded.",
    "summary": "Falling wedge: 1,287 resolved trades from 2006-09-25 to 2026-09-25; 54% reached the target, average win +8.1%, average loss -4.2%, median win +6.3%."
  },
  "unfiltered": {
    "scope": "unfiltered",
    "family": "falling_wedge",
    "symbol": null,
    "win_rate_pct": 54,
    "sample_size": 1287,
    "wins": 695,
    "losses": 592,
    "avg_win_pct": 8.1,
    "avg_loss_pct": -4.2,
    "median_win_pct": 6.3,
    "avg_trade_pct": 2.4,
    "avg_r": 0.41,
    "expected_r": 0.41,
    "profit_factor": 1.61,
    "avg_hold_bars": 11.2,
    "backtest_range": [
      "2006-09-25",
      "2026-09-25"
    ],
    "measured": "Every detector-confirmed breakout … open trades are excluded.",
    "summary": "Falling wedge: 1,287 resolved trades from 2006-09-25 to 2026-09-25; 54% reached the target, average win +8.1%, average loss -4.2%, median win +6.3%."
  },
  "note": null
}

Errors

Errors are JSON with the Error shape; the fix field says what to do.

Status	Code	Meaning	Fix
401	`missing_api_key`	The request had no `Authorization: Bearer tp_live_…` header and the route needs one.	Send the header on every keyed call. Only the demo, families and universe routes answer without it.
401	`invalid_api_key`	The key is not one we know, or it has been revoked.	Create a new key on the API Keys page and replace it wherever the old one was stored.
403	`forbidden`	The account behind the key is not on Premium+. API and MCP access requires Premium+ subscription at $30/month; Premium alone does not unlock the API.	Get Premium+ ($30/month, all of the benefits in Premium plus API and MCP access) from the Quickstart, where a first subscription is $0 for 1 month, or switch an existing Premium subscription to it there, then retry with the same key.
429	`rate_limited`	The key passed 60 requests in a minute or 10,000 in a day, or a keyless caller passed 30 a minute from one IP.	Wait `details.retry_after_seconds`, then retry. Cache the nightly answers; they do not change until the next session.
400	`bad_request`	A query parameter is malformed: `limit` or `offset` is not an integer, `outcome` is not `all`, `resolved` or `open`, a scan filter (`family`, `direction`, `status`, `max_distance_pct`) is not one of its allowed values, or `filters` is not a valid predicate list (the message names the predicate).	The `message` names the parameter. Correct it and retry.