Datasets:
input stringlengths 2 140 | category class label 5
classes | stages listlengths 0 3 |
|---|---|---|
Standard: 90 mins for all moves with 30 sec increment from move 1 | 0std | [
[
0,
5400,
30
]
] |
Standard: 1 ½ hours with 30 second increments from move 1 | 0std | [
[
0,
5400,
30
]
] |
Rapid: 20 minutes on game | 1rpd | [
[
0,
1200,
0
]
] |
Standard: 90 minutes for the first 40 moves, and 30 minutes for the rest of the game. 30 second increment from | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 90' +30'' | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 minutes for first 40 moves followed 30 minutes for rest of the game | 0std | [
[
40,
5400,
0
],
[
0,
1800,
0
]
] |
Standard: 40 in 120 then all in 60 | 0std | [
[
40,
7200,
0
],
[
0,
3600,
0
]
] |
Standard: 120 minutos all game | 0std | [
[
0,
7200,
0
]
] |
Standard: 1h + 30'' | 0std | [
[
0,
3600,
30
]
] |
Standard: 90 min. + 30 sec./each move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 minutes + 30sec / move | 0std | [
[
0,
5400,
30
]
] |
Standard: 40 in 90, all in 30, + increment 30 sec /move | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 100' + 30 " bonus | 0std | [
[
0,
6000,
30
]
] |
Rapid: 25 minutes for game with 15 seconds increment from move 1 | 1rpd | [
[
0,
1500,
15
]
] |
Standard: 90-90min/40 moves +30-30 min/all + 30 sec./each move | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 1h30 + 30sec/mv | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 min. + 30 sec. | 0std | [
[
0,
5400,
30
]
] |
Standard: 60 Moves in 10 Min. / / | 2blz | [
[
60,
600,
0
]
] |
Standard: All Moves in 120 Min. Del. 5 Sec. / / | 0std | [
[
0,
7200,
-5
]
] |
Standard: 1h30/40cps - 30' ko + 30'' | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 100'/40 ms + 20' + 30 | 0std | [
[
40,
6000,
30
],
[
0,
1200,
30
]
] |
Blitz: 5 minutes for games with 2 seconds from move 1 | 2blz | [
[
0,
300,
2
]
] |
Standard: 40/90, 30 minutes + increment 30 sec from move 1 | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 90 m with 30 sec from the first move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 min 40 mov + 30 min + 30 sg / mov | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Rapid: 25 min + 5 sec/move | 1rpd | [
[
0,
1500,
5
]
] |
Blitz: 3mins/all game + 2sec per move | 2blz | [
[
0,
180,
2
]
] |
Standard: 90 min. plus 30 sec. per move | 0std | [
[
0,
5400,
30
]
] |
Standard: 40/90 min+30min+30 sec/move | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Rapid: All Moves in 35 Min. Incr. 5 Sec. / / | 1rpd | [
[
0,
2100,
5
]
] |
Standard: 1h + 30 min | 0std | [
[
0,
5400,
0
]
] |
Standard: 120/40 + 30 | 0std | [
[
40,
7200,
0
],
[
0,
1800,
0
]
] |
Standard: 1h30/40 cp + 30min +30sec/cp | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Blitz: 3 min+ 2 sec/move | 2blz | [
[
0,
180,
2
]
] |
Standard: 90-90 min/all + 30 sec/move | 0std | [
[
0,
5400,
30
]
] |
Standard: 60 Moves in 90 Min. Incr. 30 Sec. / / 10 Min. | 0std | [
[
60,
5400,
30
],
[
0,
600,
30
]
] |
Standard: 40/90 min+30min | 0std | [
[
40,
5400,
0
],
[
0,
1800,
0
]
] |
Standard: all moves in 90 min. + 30 | 0std | [
[
0,
5400,
30
]
] |
Rapid: 20mins/all game + 10secs per move | 1rpd | [
[
0,
1200,
10
]
] |
Standard: 90 min, 30 sec increment | 0std | [
[
0,
5400,
30
]
] |
Standard: 90min/40m; 30min/end + 30s/move | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 90 min/30 moves + 60 min. | 0std | [
[
30,
5400,
0
],
[
0,
3600,
0
]
] |
Standard: 90 Min+30 sec/move | 0std | [
[
0,
5400,
30
]
] |
Standard: G/90 + 60 sec inc | 0std | [
[
0,
5400,
60
]
] |
Rapid: 10m+5s | 2blz | [
[
0,
600,
5
]
] |
Standard: 90 + 30 seg de incremento | 0std | [
[
0,
5400,
30
]
] |
Standard: 60' + 15'' | 0std | [
[
0,
3600,
15
]
] |
Standard: 90min/40 moves, 15min/all + 30sec per move | 0std | [
[
40,
5400,
30
],
[
0,
900,
30
]
] |
Standard: 90 min for all moves plus 30 sec.move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90-90 min/40 moves + 30-30 min./all | 0std | [
[
40,
5400,
0
],
[
0,
1800,
0
]
] |
Standard: 90 min + 30 sg / mov | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 min/40 + 20 min ++ 30 sek | 0std | [
[
40,
5400,
30
],
[
0,
1200,
30
]
] |
Standard: 120 | 0std | [
[
0,
7200,
0
]
] |
Standard: 90 mins / 40 moves + 15 mins + 30 sec | 0std | [
[
40,
5400,
30
],
[
0,
900,
30
]
] |
Standard: 1h30+30s | 0std | [
[
0,
5400,
30
]
] |
Standard: 40/120, 25 + increment 5 sec final time control | 0std | [
[
40,
7200,
0
],
[
0,
1500,
5
]
] |
Standard: 90' /40 ms + 20' + 30" bonus | 0std | [
[
40,
5400,
30
],
[
0,
1200,
30
]
] |
Standard: 40/90, 15/30, 30, | 0std | [
[
40,
5400,
0
],
[
15,
1800,
0
],
[
0,
1800,
0
]
] |
Standard: 90min and 30sec per move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90mins/40moves, 30mins/all + 30 | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Blitz: 3 Min + 2 seg de incr | 2blz | [
[
0,
180,
2
]
] |
Standard: G90, 30 | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 min/40moves+15min, incr. 30 sec/move | 0std | [
[
40,
5400,
30
],
[
0,
900,
30
]
] |
Standard: Fisher, 90 minutes + 30 seconds a move | 0std | [
[
0,
5400,
30
]
] |
Rapid: 15 mIn + 10 seg de inc | 1rpd | [
[
0,
900,
10
]
] |
Standard: 65 minutes for game | 0std | [
[
0,
3900,
0
]
] |
Standard: 40/2, 1 hour | 0std | [
[
40,
7200,
0
],
[
0,
3600,
0
]
] |
Standard: 90-90 min./all moves | 0std | [
[
0,
5400,
0
]
] |
Standard: 90mins/40moves and 30mins/all +30 sec/move | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 90' +30"/move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90min + 15 min + 30 sec/coup | 0std | [
[
0,
6300,
30
]
] |
Blitz: 150 minutes for the whole game | 0std | [
[
0,
9000,
0
]
] |
Standard: 90min/40moves+30min + 30 sec.incr. | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 90 min finish + 30 seg de incr por jugada | 0std | [
[
0,
5400,
30
]
] |
Standard: 120 min. finish | 0std | [
[
0,
7200,
0
]
] |
Standard: 90 Min + 30 Seconds Per Move | 0std | [
[
0,
5400,
30
]
] |
Rapid: 15 min+ 6 sec. | 1rpd | [
[
0,
900,
6
]
] |
Standard: 90mins +30sce. per moves. | 0std | [
[
0,
5400,
30
]
] |
Standard: 40/90' + 30' + 30 | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: 2 hours/ 40 moves, plus 1 hour | 0std | [
[
40,
7200,
0
],
[
0,
3600,
0
]
] |
Blitz: All Moves in 2 Min. / / | 3blt | [
[
0,
120,
0
]
] |
Standard: 40/90min. + 15 min. + 30 sec. for each move | 0std | [
[
40,
5400,
30
],
[
0,
900,
30
]
] |
Standard: 90 mts+30 sec increment per move | 0std | [
[
0,
5400,
30
]
] |
Standard: 90'+30'' per move | 0std | [
[
0,
5400,
30
]
] |
Standard: 40/90; SD/30 + 30 Seconds Throughout | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Standard: una hora y 30 minutos mas 30 segundos de adición | 0std | [
[
0,
5400,
30
]
] |
Standard: 90/40 + 15 + 30' | 0std | [
[
40,
5400,
0
],
[
0,
900,
30
]
] |
Standard: 90 Min. plus 90 Sec. | 0std | [
[
0,
5490,
0
]
] |
Standard: Game 90 + 30 second increment | 0std | [
[
0,
5400,
30
]
] |
Standard: 90 min for 40 moves + 15 min. 30 sec inc from move 1. | 0std | [
[
40,
5400,
30
],
[
0,
900,
30
]
] |
Rapid: 10' + 5'' | 2blz | [
[
0,
600,
5
]
] |
Standard: 90 minutes for all moves, plus 30 sec per move, beginning from 1st move | 0std | [
[
0,
5400,
30
]
] |
Blitz: 3 minutes plus 2 seconds per move | 2blz | [
[
0,
180,
2
]
] |
Standard: 90 mins plus 30 sec increment from move 1 | 0std | [
[
0,
5400,
30
]
] |
Standard: all moves in 2h | 0std | [
[
0,
7200,
0
]
] |
Standard: 90/40, 30 + 30" | 0std | [
[
40,
5400,
30
],
[
0,
1800,
30
]
] |
Blitz: 3 Min. Plus 2 Sec. | 2blz | [
[
0,
180,
2
]
] |
Standard: 90 minutes + 30 sec. incr. mper move. | 0std | [
[
0,
5400,
30
]
] |
Standard: 120-120 min. for all /QPF | 0std | [
[
0,
7200,
0
]
] |
Standard: 75 mins for all moves plus 30 second Fischer increments | 0std | [
[
0,
4500,
30
]
] |
End of preview. Expand in Data Studio
Chess Time-Control String Parsing
Real-world chess time-control strings, in two forms:
.txtfiles — the source of truth. Every unique time-control string, one per line, with a frequency count. These are the raw, real strings (messy, multilingual, sometimes junk) as scraped/collected. No interpretation..jsonlfiles — a tagged, partially-correct derived artifact. Each unique string with an auto-derived(category, stages)parse. The tags are heuristics, not verified ground truth (see Caveats).
Files
| File | Rows | Kind | Source |
|---|---|---|---|
2025-08-10_fide-time-control-patterns-with-counts.txt |
2,193 | raw truth | FIDE tournament pages |
2025-08-10_fide-tournament-time-controls.jsonl |
2,063 | tagged | derived from the above |
2000-2025_lumbra-pgn-timecontrol-formats-with-counts.txt |
416 | raw truth | Lumbras GigaBase OTB PGN [TimeControl] tags |
2000-2025_lumbra-gigabase-time-controls.jsonl |
416 | tagged | derived from the above |
Both .txt files use the same format: {count:>6} {string} (6-wide right-aligned count, single space, raw string), CRLF line endings, no header. Filenames: YYYY-MM-DD_ = scrape/extract date; YYYY-YYYY_ = data coverage range. Each .jsonl has a matching .metadata.json.
The .txt row counts (2,193 / 416) exceed the .jsonl row counts because the .jsonl drops blanks, duplicates across the two forms, and non-parseable rows.
Row format (.jsonl)
{"input": "90min/40moves+30min/end+30sec increment per move starting from move 1",
"category": "std",
"stages": [[40, 5400, 30], [0, 1800, 30]]}
category:std=Standard,rpd=Rapid,blz=Blitz,blt=Bullet,var=variable per-round.stages[i]=[moves, time_seconds, increment_seconds].moves=0= rest of game (nullin the raw.txt-adjacent form, encoded as0here because HuggingFace's loader cannot hold nullable ints in nested lists). Negative increment = delay.- Conversion: minutes×60, hours×3600.
Schema
input: raw time-control string (may include non-English words, unicode, USCFG/90, or multi-stage schedules).category: FIDE speed class by base time (minutes per player):std≥ 60rpd> 10 and < 60blz> 3 and ≤ 10blt< 3 (not FIDE-rated)varvariable per-round
Labels use base time only; FIDE rating eligibility uses base + 60×increment (Handbook B02, eff. 2024-03-01). A label may differ from a textual prefix (e.g. "Blitz: 150 minutes" →
std).stages: one or more time-control stages.varrows encode per-round schedules as multiple stage tuples (e.g. "R1-4: 15+5, R5-8: 90+30").blt(bullet) appears in thepgnset but notfide(FIDE does not rate bullet games).
Caveats (read before trusting the tags)
- Tags are auto-derived, not human-verified. The
category/stagesare the output of a heuristic parser, not gold labels. Spot audits found errors (e.g. some 60-min bases taggedrpd, boundary mis-sorts, spurious[null,0,0]rest stages). Use the.txtfiles as the authoritative source; treat the.jsonltags as a starting point. - Some
.jsonlrows are junk — org/federation names, player names,n/a,---— that slipped through scraping. They are valid scraped content but not time controls; they were kept rather than silently dropped. - FIDE vs PGN differ in shape — FIDE is messy natural language (2,193 unique from 167,677 pages), PGN is mostly canonical
[TimeControl]values (416 unique).
Provenance
- Sources. FIDE ratings-site tournament pages (
ratings.fide.com/tournament_information.phtml?event=<id>); Lumbras GigaBase OTB PGN corpus (LumbrasGigaBase_OTB_{2000-2004,2005-2009,2010-2014,2015-2019,2020-2024,2025}.pgn),[TimeControl]header tags. FIDE Handbook B02 (effective 2024-03-01) for speed-class thresholds. - Collection. FIDE: 167,677 tournament pages scraped on 2025-08-10 → 2,193 unique time-control strings (source of truth). PGN:
[TimeControl]tags extracted on 2025-08-07 across the 2000–2025 GigaBase files → 443 raw format tokens → 416 unique strings. Data coverage 2000–2025 (2025 partial). - Methodology.
.txtfiles are the raw, uninterpreted source of truth (string<TAB>countfor PGN;count <string>for FIDE)..jsonlrows are auto-derived by a heuristic parser that maps each string to(category, stages); blanks, duplicates, and non-parseable rows were filtered, so.jsonlcounts (2,063 / 416) are ≤.txtcounts (2,193 / 416). Tags are not human-verified ground truth (see Caveats). - Citations. FIDE Handbook, Title 02 (Rating Regulations), eff. 2024-03-01: https://handbook.fide.com/chapter/B02RBRegulations2024 . Lumbras GigaBase: https://www.lumbra.gigabase.org/ . Dataset released under CC0-1.0.
Evaluation
If you treat the .jsonl as a labeling task: exact-match on stages (ordered tuple equality) plus category accuracy. Split as you see fit — no pre-splits included. For a parsing benchmark, better to generate your own targets from the .txt strings.
Prior art
Existing tools (python-chess, @jackstenglein/pgn-parser, Lichess) parse only standardized PGN [TimeControl] strings, not natural-language or multi-stage descriptions.
License
CC0 1.0 — public domain, no restrictions.
- Downloads last month
- 49