Squares in Squares on Kaggle: Rules, Scoring and Our Run
Squares in Squares is a Kaggle Playground competition with a $50,000 prize pool that closes on 18 July 2027. For every n from 1 to 1,000 you place n unit squares, rotated if you like, inside the smallest square you can. Each puzzle scores L/√n − 1, where L is the side of your container, and the leaderboard is the sum. Lower is better.
Competition facts
- Host
- Kaggle (Playground)
- Prize pool
- $50,000
- Final deadline
- 18 Jul 2027, 23:59 UTC
- Task
- Pack n unit squares into the smallest square, for n = 1 … 1,000
- Score
- Sum of L/√n − 1 over all 1,000 puzzles · lower is better
- Submission
- A CSV of every square’s x, y and angle — not a code competition
- Validity
- No two squares may overlap; the check is strict floating point
- Medals
- None — Playground competitions award no Kaggle medals
- Teams
- 2,086 on the public leaderboard (15 Nov 2026)
- Leader
- 22.0400 public (15 Nov 2026)
Deadline, prize and team count from the Kaggle API; rules from the competition page. Check the official rules before you rely on any of it.
Our run with ResearchForge
- 401 / 2,086
- Public rank
- 22.1756
- Best public (lower wins)
- 19.2%
- Top of the field
- 46
- Experiments · 7 rejected
Medal standing · Playground · no medals
Squares in Squares is a Playground competition, so it awards no Kaggle medals; the $50,000 prize pool still applies. Top 10% today is 22.1528; our newest local packing measures 22.1580 and is not submitted yet.
Public score from 22.2502 to 22.1756 in eight days, over 46 rounds. Almost every gain came from certified merges of public packings; our own search ideas have barely moved it yet.
- 5 Nov—
Kaggle Overnight read the competition and 2 starter notebooks, and pulled 12 public packing outputs
- 5 Novrejected
Rounds 1–2: per-n merge of public packings — passed the quick screen, failed the full check
- 5 Nov22.2502
Round 3: margin-safe certified merge — kept and submitted
- 5 Novrejected
Round 4: free-angle polish of every non-grid puzzle — hit the 30-minute round limit
- 5 Novno gain
Round 5: the same polish with a deadline — finished, no measurable improvement
- 9 Noverror
A merge that passed locally came back from Kaggle’s scorer as an error — kept on file, not retried blind
- 11 Nov22.2004
Merging the newest public packings per n, rounds 20–22
- 12 Nov22.1926
A validated public merge plus the outputs it was built from
- 13 Nov22.1756
Merge of every notebook the watcher had recorded — best public score so far
- 14 Nov22.1580
Local only, not yet submitted: notebooks that claim a better score, merged per n
What the record showed
17 rounds improved the score; 16 were merges of public outputs, checked for overlap. Our own polishing, neighbour transfer and seed shapes ran 12 times for one gain of 0.0000015. The field moves daily, so a better score can still mean a lower rank.
Questions people ask
What is the Kaggle Squares in Squares competition?
A Kaggle Playground competition based on a classic open problem: for each n from 1 to 1,000, pack n unit squares into the smallest possible square. Squares may be rotated. The prize pool is $50,000 and the final deadline is 18 July 2027, 23:59 UTC.
How is Squares in Squares scored?
Each puzzle scores L/√n − 1, where L is the side of the smallest axis-aligned square that contains your packing of n squares. The leaderboard score is the sum over all 1,000 puzzles, and lower is better. A perfect square n, such as 16, packed as a grid scores 0.
Is Squares in Squares a code competition?
No. You upload a CSV with one row per square (id, n, square_id, x, y, angle), so there is no notebook runtime limit. The only hard rule on a packing is that no two squares overlap, and the check is strict, so packings that touch need a tiny safety margin.
Where do people start in Squares in Squares?
From published best-known packings, such as Erich Friedman’s survey and David Ellsworth’s table, and from public Kaggle notebooks that merge them. A certified merge of public packings scores about 22.25; the leaders at 22.05 have found better packings for some n.
How did ResearchForge do on Squares in Squares?
Our first submission on 5 November 2026 scored 22.2502. By 15 November the best public score was 22.1756, rank 401 of 2,086, after 46 recorded rounds. Almost every gain came from certified merges of public packings; our own polishing and seed ideas have barely moved it yet. The newest local packing measures 22.1580 and is not submitted yet.
Sources
- Kaggle competition page
- Erich Friedman, “Packing Unit Squares in Squares: A Survey and New Results”
- Square packing (Wikipedia)
- Our scores and ranks: public Kaggle leaderboard, 15 Nov 2026, and the ResearchForge project record.
Run it on this competition
Kaggle Overnight is open-source autoresearch for Kaggle: it keeps your score level with the best public notebooks, then keeps experimenting on top of them. Give it the competition URL tonight.