Data health

Every source, every game, every night.

A site that reports what happened is only as good as its record of it. This is that record, with its gaps: where each game’s draws come from, whether it is behind its own calendar, how much of it two sources agree on, and what last night’s integrity scan found. Generated 23 Sept 2026, 22:16 UTC. The same record by machine: feeds, CSV and the API.

Last night’s integrity scan

Seven checks, every night, for the ways the data has been found wrong before: one draw filed under two labels, a result posted again the next day, a rule era no lottery could have, a bonus ball on the wrong machine, a row of zeros, a number drawn twice. A fault here is a bug in our pipeline, never a pattern in the balls.

• OKduplicate slot twins0the same numbers promoted under two slot labels on one date; the resolver should have folded them
• Faultreposts2the same numbers on consecutive dates in one series; the resolver should have dropped the later one — us_ok_cash5:2
• OKera sanity0a derived format version that describes a game nobody runs — usually a feed packing add-on balls into the draw
• OKbonus pool flag0a same-drum bonus ball flagged as a separate machine; odds and star decks read it wrong
• OKplaceholders0all-zero results stored as draws
• OKrepeated balls0a number drawn twice in one draw — a packed payload or a parser fault
• Noteopen format violations228records the resolver could not promote because they break their era; each is a row a visitor cannot see — ca_bc_extra:90, us_md_bonusmatch5:31, us_2by2:24, us_ne_myday:12, us_nj_jerseycash5:12, za_lotto:8
• Notedeclared odds disagree with the rules80games whose published jackpot odds disagree with the rules on record; the operator states the odds on the ticket, so the format here is what is wrong — es_primitiva: says 1 in 139,838,160, rules give 1 in 13,983,816 (10.000x), ca_lottomax: says 1 in 33,446,140, rules give 1 in 133,784,560 (0.250x), us_ne_myday: says 1 in 36,525, rules give 1 in 7 (5217.857x), ca_qc_lottopoker: says 1 in 62,212,254, rules give 1 in 371,293 (167.556x), de_de_plus5: says 1 in 13,000,000, rules give 1 in 100,000 (130.000x), us_pr_pega2: says 1 in 10,000, rules give 1 in 100 (100.000x)
• Faultper-ball totals disagree with the feed4The feed publishes its own count of how often each ball has ever been drawn. It comes from the same draws we ingest, so it is useless as a statistic and useful as a check: where the two disagree, one record is missing a draw or holding one twice, and the ball named is where to look. Nothing here is a finding about a lottery — it is a finding about this database. — fr_loto: we hold 2981 draws, the feed 3960, gb_lotto: we hold 3238 draws, the feed 3244, us_powerball: we hold 3863 draws, the feed 3858, br_megasena: we hold 3057 draws, the feed 3060

Sources

Ranked by how much we trust them. A draw is “verified” only when at least two of these agree and one of them is official; a game with a single source is shown as such, draw by draw.

Results pushed to us

The feed posts a signed request within seconds of publishing a result, so a draw no longer waits for the next poll. A delivery still waiting is one the queue has not finished with, and the poller holds off on that draw until it has.

• Delivered in the last 7 days0
• Of those, handled0
• Still waiting0
• Most recent deliverynever

The prize record

A prize table arrives as prose — “Match 4 plus Bonus” — and means nothing to a machine until it is keyed to an outcome. Everything the site says about money rests on this number: implied sales, the return ledger, the crowd model, the add-ons.

• Prize rows held2,926,609
• Read as a tier98.6%
• With a prize fund80.5%
• Left unread40,810

Games reading below 99%, with the reason

• Lotto92.5%placeholder dash
• 2by294.3%more matches than the era can draw
• PowerBall95.9%placeholder dash
• Daily Lotto80.8%placeholder dash
• Weekday Windfall0.1%division table missing for this era
• Cash Pop89.1%placeholder dash
• Powerball61.0%more bonus than the era can draw
• Powerball0.1%division table missing for this era
• Texas Two Step94.1%bonus not in era
• Oz Lotto98.5%division table missing for this era
• Lotto Strike0.3%division table missing for this era
• Super 660.3%division table missing for this era

Behind the calendar

Games whose latest draw on record is older than the last draw their calendar says has happened. 11 of 280 games.

Retired games

Games that have ended. Their history stays here in full and stays searchable — what stops is the calendar, so they are never counted as behind it.

Open issues

Draws where sources disagree, and records the resolver refused to promote because they break the rules of their era. Each is a row a visitor cannot see until it is understood. 69 games.

…and 29 more, in the full table below.

Every game

No data at all

On the books, with no trustworthy source: the official feed is dead or no provider covers the game. We would rather show nothing than make it up.

• Vikinglotto EU
• Lotto FI