Tennis Scores: ATP, WTA and Challenger Results with Source Failover

Daily tennis scoreboards as clean dataset rows: one row per match, with set-by-set games, tiebreak points, status and winner. Built for sports analytics, dashboards, research projects and media apps that need a steady feed of public results.

The Actor reads one primary scoreboard and keeps going when it breaks. If the primary errors, is rate limited, times out, changes its format, or returns nothing for a date, the run switches to backup sources automatically and marks every row it filled that way.

What you get

• Tours: ATP, WTA and ATP Challenger. Singles and doubles.
• Match states: scheduled, live (with the current set), finished, retired, walkover, cancelled.
• Scores: games per set, tiebreak points, sets won, a score line such as 6-7 4-6, winner, match duration where published.
• Players: public name, nationality and ranking when the source shows them. No other personal data.
• Optional match stats (aces, double faults, break points, serve points) when the primary source publishes them.
• Provenance on every row: which source served it and whether a backup was used.

Coverage by source

Source	Tours	Doubles	Live	Tiebreaks	Round	Stats	Requests per date
365scores (primary)	ATP, WTA, Challenger	yes	yes	yes	yes	yes, optional	1 (+1 per match for stats)
livescore (backup 1)	ATP, WTA, Challenger	yes	yes	yes	no	no	1
espn (backup 2)	ATP, WTA	yes	yes	yes	yes	no	2
wta (backup 3)	WTA only	yes	partial	loser points only, winner side derived	yes	no	1 + one per WTA event that week

Challenger results come only from 365scores or livescore. If both are down, Challenger rows are missing for that date and the log says so. The WTA backup has no match stats, and stats in general exist only on 365scores rows.

How failover works

1. Sources are tried in the order of sources (default 365scores, livescore, espn, wta).
2. When the primary serves a date, no backup is contacted at all (unless crossCheck is on).
3. When it fails, backups fill the tours they cover, one after another, until every requested tour is covered or the list ends.
4. Rows from a backup carry fallback_used: true and a short fallback_reason, e.g. 365scores: HTTP 429.

Triggers: HTTP errors and 5xx after retries, HTTP 429, timeouts, invalid JSON, a changed payload shape (parse errors or more than 20% of rows failing validation), and an empty scoreboard or no rows for the requested tours.

Input

Field	Default	Notes
dates	today (UTC)	YYYY-MM-DD, at most 7 dates
circuits	ATP, WTA, Challenger	Doubles boards map onto the same tour
sources	365scores, livescore, espn, wta	Order is priority. The first source is primary. Later sources are backups
includeDoubles	true	Set false to keep singles only
statuses	scheduled, live, finished	Walkovers and retirements count as finished
includeStats	false	365scores only. One extra request per live or finished match that publishes stats
maxMatches	200	1–2000, across all dates
playerNames	empty	Case-insensitive substring on either player
requestDelaySecs	2	Minimum 1 second from one request to the next
maxRequestsPerRun	150	10–1000. Hard cap on HTTP requests in one run across all sources (stats requests included)
timezone	UTC	Sent as timezoneName on 365scores
crossCheck	false	Compare finished scores with the first backup that covers a selected tour
simulateFailure	empty	Testing only. Named sources fail before any request
healthCheck	false	Check sources and skip the dataset. No charge
healthCheckDate	yesterday UTC	YYYY-MM-DD used when healthCheck is true

Input example

{
  "dates": ["2026-10-21"],
  "circuits": ["ATP", "WTA", "Challenger"],
  "statuses": ["finished"],
  "includeDoubles": true,
  "maxMatches": 200
}

Output example

A real row from a cloud test run on 2026-10-22:

{
  "match_id": "4878233",
  "source": "365scores",
  "date_requested": "2026-10-21",
  "start_time_utc": "2026-10-21T03:03:33Z",
  "circuit": "Challenger",
  "discipline": "singles",
  "tournament": "Cali",
  "tournament_id": 6943,
  "round": "Round of 32",
  "status": "finished",
  "status_text": "Ended",
  "player1": {
    "name": "Federico Coria",
    "id": 47314,
    "country": "Argentina",
    "ranking": 746
  },
  "player2": {
    "name": "Nicolas Mejia",
    "id": 51765,
    "country": "Colombia",
    "ranking": 115
  },
  "sets_won_p1": 0,
  "sets_won_p2": 2,
  "set_scores": [
    {
      "set": 1,
      "p1": 6,
      "p2": 7,
      "tiebreak_p1": 4,
      "tiebreak_p2": 7,
      "duration": "01:11"
    },
    {
      "set": 2,
      "p1": 4,
      "p2": 6,
      "tiebreak_p1": null,
      "tiebreak_p2": null,
      "duration": "49'"
    }
  ],
  "score_text": "6-7 4-6",
  "winner": 2,
  "current_set": null,
  "match_duration": "02:00",
  "stats": null,
  "match_url": "https://www.365scores.com/en/tennis/match/cali/federico-coria-nicolas-mejia/4878233",
  "scraped_at": "2026-10-22T15:07:09Z",
  "fallback_used": false,
  "fallback_reason": null,
  "verified_by": null,
  "crosscheck": null
}

Extra fields:

• source: the source that served the row.
• fallback_used and fallback_reason: whether a backup supplied the row, and why the earlier sources were skipped.
• verified_by and crosscheck: filled only when crossCheck is on. crosscheck is match, mismatch or not_found; verified_by lists the backup that agreed.

Fields a source does not publish are null. match_id, source, date_requested and scraped_at are always set.

Health check

Set "healthCheck": true to test every source on a recent finished date (healthCheckDate, default yesterday UTC). The run pushes no rows and charges nothing. It saves a report under the key HEALTH_REPORT in the default key-value store, and its status line reads like health: 365scores PASS, livescore PASS, espn PASS, wta PASS. The run fails if any checked source fails.

The same check runs outside Apify:

python scripts/health_check.py --date 2026-10-21

Exit code 0 means all sources passed. For monitoring, start the Actor once a day from an Apify schedule with {"healthCheck": true}, or call the script from cron.

Pricing

Pay per event: $0.001 per match row. Nothing is charged for health checks, for dates where every source failed, or for empty results. A failed stats request still returns the match, with stats set to null.

Request etiquette

Each source has its own queue: one request at a time, at least 1–2 seconds apart plus a random pause, and backups stay idle while the primary works. HTTP 429 honors Retry-After (one retry when it is 30 seconds or less), server errors get two retries with growing pauses, and 403 is never retried. maxRequestsPerRun caps the total requests of a run.

Data sources

Scores come from public JSON scoreboards published by 365Scores, LiveScore, ESPN and the WTA. No login, key or browser is used. Each publisher has its own terms of use, and some of them restrict automated access or commercial reuse; you are responsible for how you use the data. This Actor is not affiliated with or endorsed by 365Scores, LiveScore, ESPN, the WTA or the ATP.

Also from Quill Data

• Google Hotels Prices — nightly and total-stay hotel rates from public Google Hotels search.
• Google Flights Cheapest Fare Monitor — cheapest public Google Flights fare for a route and date; failed or empty runs are free.

Limits

• Live scores are a snapshot at run time. Schedule the Actor for regular updates.
• Up to 7 dates per run, UTC calendar days.
• Round names are not available from the livescore backup.