QRATOR is a real-time AI intelligence system — powered by an AI cricket brain we call the Context Engine — that watches a live cricket match and generates the most valuable analytical story for every ball, grounded in a decade of ball-by-ball history, ranked, deduplicated, and delivered to the broadcast desk in under two seconds. It powered the entire IPL 2026 season.
| p_match | inns | bat | p_bat | team_bat | bowl | p_bowl | team_bowl | ball | ball_id | outcome | score | out | dismissal | p_out | over | noball | wide | byes | legbyes | cur_bat_runs | cur_bat_bf | cur_bowl_ovr | cur_bowl_wkts | cur_bowl_runs | inns_runs | inns_wkts | inns_balls | inns_runs_rem | inns_balls_rem | inns_rr | inns_rrr | target | max_balls | date | year | ground | country | winner | toss | competition | bat_hand | bowl_style | bowl_kind | batruns | ballfaced | bowlruns | bat_out | wagonX | wagonY | wagonZone | line | length | shot | control | predscore | wprob | nb | bruns | shot_type |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1527681 | 1 | Ryan Rickelton | 605661 | Mumbai Indians | Mukesh Kumar | 926851 | Delhi Capitals | 1 | 0.01 | four | 4 | False | — | — | 1 | 0 | 0 | 0 | 0 | 4 | 1 | 0.1 | 0 | 4 | 4 | 0 | 1 | — | 119 | 24.0 | — | — | 120 | 04-04-2026 15:30:00 | 2026 | Arun Jaitley Stadium, Delhi | India | Delhi Capitals | Delhi Capitals | IPL | LHB | RM | pace bowler | 4 | 1 | 4 | False | 7 | 243 | 6 | DOWN_LEG | GOOD_LENGTH | FLICK | 1.0 | 211.0 | 63.41 | 0 | 4 | FLICK |
Traditional scoreboards describe what happened. QRATOR contextualises it — every ball viewed through the lens of a decade of cricket history, so a commentator gets the fact that actually matters, at the exact moment it matters.
At the heart of QRATOR sits our biggest bet — the Context Engine. A cricketing brain that watches every ball, weighs it against a decade of history, and decides in under two seconds which story is worth telling. That judgement, made hundreds of times a match across an entire IPL season, is what turns raw data into broadcast intelligence.
If the generators are QRATOR's voice, the Context Engine is its mind. It continuously classifies every ball across twelve nested dimensions — phase, batter state, bowler state, tournament stakes, historical rarity, matchup type — and decides which stories should even be considered for firing, let alone published.
Over four months of development and three months of live-season operations, our team shipped 300+ system-level improvements based on analyst feedback after every match. QRATOR is now a living reference architecture for real-time sports intelligence.