Current world records

World record

Pulls every round flagged regional_single_record = 'WR' or regional_average_record = 'WR', then per (event, single/average) keeps the minimum value — i.e. the WR that still stands today.

One row per currently-standing record, with full attempt breakdown for singles. Historic WRs that have since been broken are not shown.

By the numbers

17 × 2

Rows (full)

Single + average per event, no ties

Filter

regional_single/average_record = WR

asc

Tie order

Earlier start_date wins ties

attempts

Extra detail

All 5 attempts shown

Data source

Reads results filtered by either WR flag; joins persons (sub_id = 1) and competitions for name / comp / date. The single's attempts come from a GROUP_CONCAT subquery on result_attempts, then the client splits and formats each attempt (string split avoids the MBLD space-encoding quirk).

sql

SELECT r.event_id, r.best AS single, r.average,
       r.regional_single_record, r.regional_average_record,
       (SELECT GROUP_CONCAT(ra.value ORDER BY ra.attempt_number)
          FROM result_attempts ra WHERE ra.result_id = r.id) AS attempts,
       p.name, c.cell_name, c.start_date
FROM results r
JOIN persons p ON p.wca_id = r.person_id AND p.sub_id = 1
JOIN competitions c ON c.id = r.competition_id
WHERE r.regional_single_record = 'WR'
   OR r.regional_average_record = 'WR'
ORDER BY r.event_id, c.start_date;

Algorithm / pipeline

Pull every WR-flagged result

A single results row may have a WR single, a WR average, or both. The SQL OR captures all three; downstream transform splits single vs average and processes each independently.

Group by event, keep the min

For each event: single uses r.best, average uses r.average. Within that event's WR-flagged rows, the minimum value wins — that's the standing record.

Tie-break by earlier date

Ties (common in 333fm / 333bf singles) are sorted ascending by start_date — the earliest holder lists first; every tied row is still emitted, so the table shows all co-holders.

Format single's attempts

Singles get the per-attempt breakdown: split the comma-joined GROUP_CONCAT then run each through SolveTime.clockFormat(). Average rows reuse the same attempt list as the single — showing it on the single row only avoids duplication.

Emit (event, single/average, time, person, date, comp)

One row per standing WR; date as YYYY-MM-DD; person / comp rendered as markdown links to the WCA site.

Caveats & edges

WR flags come baked into the WCA dump — no re-computation here, but newly-broken records only appear after the next dump refresh.
Single and average are independent — one results row can be a WR single but not a WR average (or vice-versa).
Tied singles (notably old 333fm = 22 moves, blazing 333bf singles) are all listed; no dedup at the UI layer.
No history view here — for time-series, see the Single / Average panels under wr_metric.

Current world records

By the numbers

Data source

Algorithm / pipeline

Caveats & edges

Related stats & links