Shortest time to all singles + averages
PersonA harsher version of "all singles": also fill every event-with-average's average. 3x3 Multi-Blind has no average, so the average target is current event count − 1.
Averages are harder than singles — BLD events need ≥ 2 successes (success / 3 attempts) to register an average; MBLD opts out of the average dimension entirely. Day counts here run 1.5-3× the singles board.
By the numbers
singles N + avg (N − 1)
Target
MBLD has no average
best > 0 / average > 0
Validity
Earliest success per type
取最大日期
Closing date
Latest across both sets
Data source
Two subqueries intersected: ① all-singles candidates (COUNT DISTINCT event_id = N); ② all-averages candidates (COUNT DISTINCT event_id = N − 1 since MBLD has no avg). Both pass → outer pulls all results into TS.
sql
SELECT event_id, person_link, start_date, best, average
FROM (SELECT person_id FROM results JOIN events ON id=event_id
WHERE rank<900 AND best>0
GROUP BY person_id
HAVING COUNT(DISTINCT event_id)=
(SELECT COUNT(*) FROM events WHERE rank<900)) singles
JOIN (SELECT person_id FROM results JOIN events ON id=event_id
WHERE rank<900 AND average>0
GROUP BY person_id
HAVING COUNT(DISTINCT event_id)=
(SELECT COUNT(*) FROM events WHERE rank<900) - 1) avgs USING(person_id)
JOIN results r USING(person_id)
JOIN persons person ON wca_id=person_id AND sub_id=1
JOIN competitions ON id=competition_id
ORDER BY start_dateAlgorithm / pipeline
1
Intersect two subqueries
Must satisfy both "all singles" and "all averages" subqueries;
= N − 1 excludes MBLD which has no avg by definition.2
Collect best + average dates
TS iterates personRows twice: once collecting earliest
best > 0 per event, once for average > 0. Concatenate the two date lists.3
Max date − first date
max of firstSuccesses = "last piece" date; subtract first comp date.
4
Sort asc
Days asc, no LIMIT. Counts typically far above the singles board because the average set includes BLD events.
Caveats & edges
- BLD events (
333bf/444bf/555bf) get only 3 attempts each, need ≥ 2 successes to log an average — this gate stalls many cubers for months. - MBLD opts out entirely (
N − 1target); 3x3 FMC averages ("mo3") are part of the set. firstSuccessesholds per-event single AND average earliest dates — they often differ (single first, average later).