---
license: odc-by
pretty_name: Doosra cricket ball-by-ball
language:
- en
tags:
- cricket
- sports
- sports-analytics
- tabular
size_categories:
- 10M<n<100M
configs:
- config_name: matches
  data_files: data/matches.parquet
- config_name: deliveries
  data_files: data/deliveries.parquet
- config_name: deliveries_wickets
  data_files: data/deliveries_wickets.parquet
- config_name: players_matches
  data_files: data/players_matches.parquet
- config_name: people
  data_files: data/people.parquet
- config_name: batting_innings
  data_files: data/batting_innings.parquet
- config_name: bowling_innings
  data_files: data/bowling_innings.parquet
- config_name: ball_expectation
  data_files: data/ball_expectation.parquet
- config_name: missing_matches
  data_files: data/missing_matches.parquet
---

# Doosra: cricket, ball by ball

Every ball of 22,983 cricket matches (men's and women's internationals and the major
leagues and domestic competitions, 1,107 competitions), cleaned, validated and ready to query, with
per-innings batting and bowling tables derived from it. Built 2026-09-28; matches up to 2026-09-17.

This is the database behind [Doosra](https://github.com/sarthak213/Doosra), a cricket analytics workbench (player
hubs, comparisons, context-adjusted metrics such as true strike rate, a local AI copilot). It is rebuilt every week
from [Cricsheet](https://cricsheet.org/).

## Tables

| Table | Rows | What it is |
|---|---:|---|
| `matches` | 22,983 | One row per match: format, competition (event_name), teams, venue, date, season, toss, result. |
| `deliveries` | 11,615,100 | One row per ball: innings, over and ball, batter, bowler, runs (batter, extras by type), wicket. |
| `deliveries_wickets` | 354,976 | Every dismissal on a ball (a ball can have two, for example a run out off a no-ball). |
| `players_matches` | 507,012 | Who played for which team in each match. |
| `people` | 18,554 | The Cricsheet register: one row per person, with their identifiers on other sites. |
| `batting_innings` | 435,515 | Derived: one row per batter per innings (runs, balls, boundaries, dismissal, batting position, entry point, and expected runs/outs for the balls faced). |
| `bowling_innings` | 300,496 | Derived: one row per bowler per innings (legal balls, runs, wickets, dots, maidens, expected runs/wickets). |
| `ball_expectation` | 54,458 | Derived: expected batter runs and dismissal chance, and expected bowler runs and wicket chance, for every ball state (format group x gender x year x innings x over x wickets down). |
| `missing_matches` | 2,930 | Matches Cricsheet lists as missing or withheld from its data (a coverage gap, not an error). |

```python
from datasets import load_dataset
deliveries = load_dataset("Sarthak213/doosra-cricket", "deliveries", split="train")
```

Or query the Parquet files directly, with no download step:

```python
import duckdb
duckdb.sql("""
    SELECT d.batter, SUM(d.runs_batter) AS runs
    FROM 'hf://datasets/Sarthak213/doosra-cricket/data/deliveries.parquet' d
    JOIN 'hf://datasets/Sarthak213/doosra-cricket/data/matches.parquet' m USING (match_id)
    WHERE m.event_name = 'Indian Premier League' AND NOT d.is_super_over
    GROUP BY 1 ORDER BY runs DESC LIMIT 5
""").show()
```

## Using it correctly

Cricket data has traps that make hand-written queries silently wrong. The rules Doosra's own tools follow:

```text
Tables (DuckDB):
- matches(match_id, match_type, event_name, match_number, gender, team_type, overs_per_innings, date TEXT 'YYYY-MM-DD', venue, city, season TEXT, team1, team2, toss_winner, toss_decision, winner, win_by_runs, win_by_wickets, player_of_match)
- deliveries(match_id, innings_num, batting_team, over_num (0-indexed), ball_in_over, batter, bowler, non_striker, runs_batter, runs_extras, runs_total, extra_type, is_wicket, wicket_kind, player_dismissed)
- players_matches(match_id, team, player)
Rules: match_type 'T20' + team_type 'international' = T20Is ('IT20' = non-official internationals); 'ODM'/'MDM' = domestic one-day/multi-day. winner NULL = tie/draw/no result. gender is 'male'/'female' and team names are the same for both (always filter gender). Innings > 2 in limited-overs matches are super overs -- exclude them. A wide is not a ball faced; wides/no-balls are not legal balls; byes/leg-byes are not charged to the bowler; 'retired hurt' is not a dismissal; run outs are not bowler wickets. Player names use Cricsheet form ('V Kohli', 'RG Sharma'), and competitions can span several event_name values -- prefer the stats tools, which handle all of this.
```

## Coverage

| match_type | gender | matches | first | last |
|---|---|---:|---|---|
| T20 | male | 10,678 | 2005-02-17 | 2026-09-17 |
| T20 | female | 3,596 | 2009-06-18 | 2026-09-17 |
| ODI | male | 2,576 | 2002-06-27 | 2026-09-15 |
| MDM | male | 2,217 | 2006-12-05 | 2026-09-15 |
| ODM | male | 1,710 | 2010-05-28 | 2026-09-11 |
| Test | male | 895 | 2001-12-19 | 2026-09-09 |
| ODI | female | 611 | 2007-01-22 | 2026-09-06 |
| ODM | female | 356 | 2016-07-15 | 2026-09-16 |
| IT20 | male | 240 | 2010-06-04 | 2024-05-26 |
| IT20 | female | 80 | 2013-07-23 | 2017-11-27 |
| Test | female | 24 | 2003-02-22 | 2026-07-10 |

Cricsheet doesn't have every match: `missing_matches` lists the ones it knows are missing or withheld. Some teams
are affected as a whole (for example, there are no Afghanistan matches), so check coverage before comparing teams.

## Schema

<details><summary><code>matches</code> (27 columns)</summary>

- `match_id` VARCHAR
- `match_type` VARCHAR
- `event_name` VARCHAR
- `match_number` VARCHAR
- `gender` VARCHAR
- `team_type` VARCHAR
- `overs_per_innings` INTEGER
- `balls_per_over` INTEGER
- `date` VARCHAR
- `venue` VARCHAR
- `city` VARCHAR
- `season` VARCHAR
- `team1` VARCHAR
- `team2` VARCHAR
- `toss_winner` VARCHAR
- `toss_decision` VARCHAR
- `winner` VARCHAR
- `win_by_runs` INTEGER
- `win_by_wickets` INTEGER
- `result` VARCHAR
- `eliminator` VARCHAR
- `method` VARCHAR
- `target_runs` INTEGER
- `target_overs` DOUBLE
- `player_of_match` VARCHAR
- `data_version` VARCHAR
- `revision` INTEGER

</details>

<details><summary><code>deliveries</code> (22 columns)</summary>

- `match_id` VARCHAR
- `innings_num` INTEGER
- `batting_team` VARCHAR
- `is_super_over` BOOLEAN
- `over_num` INTEGER
- `ball_in_over` INTEGER
- `batter` VARCHAR
- `bowler` VARCHAR
- `non_striker` VARCHAR
- `runs_batter` INTEGER
- `runs_extras` INTEGER
- `runs_total` INTEGER
- `non_boundary` BOOLEAN
- `extra_type` VARCHAR
- `extra_wides` INTEGER
- `extra_noballs` INTEGER
- `extra_byes` INTEGER
- `extra_legbyes` INTEGER
- `extra_penalty` INTEGER
- `is_wicket` BOOLEAN
- `wicket_kind` VARCHAR
- `player_dismissed` VARCHAR

</details>

<details><summary><code>deliveries_wickets</code> (9 columns)</summary>

- `match_id` VARCHAR
- `innings_num` INTEGER
- `over_num` INTEGER
- `ball_in_over` INTEGER
- `wicket_seq` INTEGER
- `kind` VARCHAR
- `player_out` VARCHAR
- `fielders` VARCHAR
- `fielder_ids` VARCHAR

</details>

<details><summary><code>players_matches</code> (4 columns)</summary>

- `match_id` VARCHAR
- `team` VARCHAR
- `player` VARCHAR
- `player_id` VARCHAR

</details>

<details><summary><code>people</code> (22 columns)</summary>

- `identifier` VARCHAR
- `name` VARCHAR
- `unique_name` VARCHAR
- `key_bcci` VARCHAR
- `key_bcci_2` VARCHAR
- `key_bigbash` VARCHAR
- `key_cricbuzz` VARCHAR
- `key_cricheroes` VARCHAR
- `key_crichq` VARCHAR
- `key_cricinfo` VARCHAR
- `key_cricinfo_2` VARCHAR
- `key_cricinfo_3` VARCHAR
- `key_cricingif` VARCHAR
- `key_cricketarchive` VARCHAR
- `key_cricketarchive_2` VARCHAR
- `key_cricketworld` VARCHAR
- `key_nvplay` VARCHAR
- `key_nvplay_2` VARCHAR
- `key_opta` VARCHAR
- `key_opta_2` VARCHAR
- `key_pulse` VARCHAR
- `key_pulse_2` VARCHAR

</details>

<details><summary><code>batting_innings</code> (73 columns)</summary>

- `match_id` VARCHAR
- `innings_num` INTEGER
- `player` VARCHAR
- `team` VARCHAR
- `opposition` VARCHAR
- `position` BIGINT
- `entry_over` INTEGER
- `entry_score` HUGEINT
- `entry_wkts` HUGEINT
- `runs` HUGEINT
- `balls` HUGEINT
- `fours` HUGEINT
- `sixes` HUGEINT
- `dots` HUGEINT
- `out` BIGINT
- `dismissal` VARCHAR
- `exp_runs` DOUBLE
- `exp_outs` DOUBLE
- `runs_first5` HUGEINT
- `balls_first5` HUGEINT
- `exp_runs_first5` DOUBLE
- `n_dot` HUGEINT
- `x_dot` DOUBLE
- `n_four` HUGEINT
- `x_four` DOUBLE
- `n_six` HUGEINT
- `x_six` DOUBLE
- `n_inplay` HUGEINT
- `x_inplay` DOUBLE
- `n_bowled` HUGEINT
- `x_bowled` DOUBLE
- `n_lbw` HUGEINT
- `x_lbw` DOUBLE
- `n_ct_keeper` HUGEINT
- `x_ct_keeper` DOUBLE
- `n_ct_field` HUGEINT
- `x_ct_field` DOUBLE
- `n_ct_bowler` HUGEINT
- `x_ct_bowler` DOUBLE
- `n_stumped` HUGEINT
- `x_stumped` DOUBLE
- `n_run_out` HUGEINT
- `x_run_out` DOUBLE
- `inplay_runs` HUGEINT
- `x_inplay_runs` DOUBLE
- `fib_runs` DOUBLE
- `fib_outs` DOUBLE
- `player_innings_in_match` BIGINT
- `result` VARCHAR
- `date` VARCHAR
- `season` VARCHAR
- `yr` INTEGER
- `match_type` VARCHAR
- `team_type` VARCHAR
- `fmt` VARCHAR
- `fgroup` VARCHAR
- `event_name` VARCHAR
- `gender` VARCHAR
- `venue` VARCHAR
- `city` VARCHAR
- `team1` VARCHAR
- `team2` VARCHAR
- `winner` VARCHAR
- `mc_runs` DOUBLE
- `mc_outs` DOUBLE
- `era_runs` DOUBLE
- `era_outs` DOUBLE
- `k_runs` DOUBLE
- `k_outs` DOUBLE
- `k_dot` DOUBLE
- `k_boundary` DOUBLE
- `k_ct_field` DOUBLE
- `k_inplay_runs` DOUBLE

</details>

<details><summary><code>bowling_innings</code> (70 columns)</summary>

- `match_id` VARCHAR
- `innings_num` INTEGER
- `player` VARCHAR
- `team` VARCHAR
- `opposition` VARCHAR
- `balls` HUGEINT
- `deliveries` BIGINT
- `runs` HUGEINT
- `wickets` HUGEINT
- `dots` HUGEINT
- `fours` HUGEINT
- `sixes` HUGEINT
- `exp_runs` DOUBLE
- `exp_wkts` DOUBLE
- `n_dot` HUGEINT
- `x_dot` DOUBLE
- `n_four` HUGEINT
- `x_four` DOUBLE
- `n_six` HUGEINT
- `x_six` DOUBLE
- `n_wide` HUGEINT
- `x_wide` DOUBLE
- `n_noball` HUGEINT
- `x_noball` DOUBLE
- `n_inplay` HUGEINT
- `x_inplay` DOUBLE
- `n_bowled` HUGEINT
- `x_bowled` DOUBLE
- `n_lbw` HUGEINT
- `x_lbw` DOUBLE
- `n_ct_keeper` HUGEINT
- `x_ct_keeper` DOUBLE
- `n_ct_field` HUGEINT
- `x_ct_field` DOUBLE
- `n_ct_bowler` HUGEINT
- `x_ct_bowler` DOUBLE
- `n_stumped` HUGEINT
- `x_stumped` DOUBLE
- `n_hit_wicket` HUGEINT
- `x_hit_wicket` DOUBLE
- `n_run_out` HUGEINT
- `x_run_out` DOUBLE
- `inplay_runs` HUGEINT
- `x_inplay_runs` DOUBLE
- `fib_runs` DOUBLE
- `fib_wkts` DOUBLE
- `maidens` BIGINT
- `mc_runs` DOUBLE
- `mc_wkts` DOUBLE
- `result` VARCHAR
- `date` VARCHAR
- `season` VARCHAR
- `yr` INTEGER
- `match_type` VARCHAR
- `team_type` VARCHAR
- `fmt` VARCHAR
- `fgroup` VARCHAR
- `event_name` VARCHAR
- `gender` VARCHAR
- `venue` VARCHAR
- `city` VARCHAR
- `team1` VARCHAR
- `team2` VARCHAR
- `winner` VARCHAR
- `k_runs` DOUBLE
- `k_wickets` DOUBLE
- `k_dot` DOUBLE
- `k_boundary` DOUBLE
- `k_ct_field` DOUBLE
- `k_inplay_runs` DOUBLE

</details>

<details><summary><code>ball_expectation</code> (38 columns)</summary>

- `fgroup` VARCHAR
- `gender` VARCHAR
- `yr` INTEGER
- `innings_num` INTEGER
- `state_over` INTEGER
- `wkts` HUGEINT
- `n_faced` HUGEINT
- `n_deliveries` BIGINT
- `exp_bat_runs` DOUBLE
- `exp_bat_out` DOUBLE
- `exp_bowl_runs` DOUBLE
- `exp_bowl_wkt` DOUBLE
- `exp_b_dot` DOUBLE
- `exp_b_four` DOUBLE
- `exp_b_six` DOUBLE
- `exp_b_wide` DOUBLE
- `exp_b_noball` DOUBLE
- `exp_b_inplay` DOUBLE
- `exp_b_bowled` DOUBLE
- `exp_b_lbw` DOUBLE
- `exp_b_ct_keeper` DOUBLE
- `exp_b_ct_field` DOUBLE
- `exp_b_ct_bowler` DOUBLE
- `exp_b_stumped` DOUBLE
- `exp_b_hit_wicket` DOUBLE
- `exp_b_run_out` DOUBLE
- `exp_t_dot` DOUBLE
- `exp_t_four` DOUBLE
- `exp_t_six` DOUBLE
- `exp_t_inplay` DOUBLE
- `exp_t_bowled` DOUBLE
- `exp_t_lbw` DOUBLE
- `exp_t_ct_keeper` DOUBLE
- `exp_t_ct_field` DOUBLE
- `exp_t_ct_bowler` DOUBLE
- `exp_t_stumped` DOUBLE
- `exp_t_run_out` DOUBLE
- `exp_inplay_runs` DOUBLE

</details>

<details><summary><code>missing_matches</code> (6 columns)</summary>

- `kind` VARCHAR
- `name` VARCHAR
- `gender` VARCHAR
- `date` VARCHAR
- `team1` VARCHAR
- `team2` VARCHAR

</details>

## Licence and attribution

The data is from [Cricsheet](https://cricsheet.org/) and is made available under the
[Open Data Commons Attribution License (ODC-By 1.0)](https://opendatacommons.org/licenses/by/1-0/). If you use it,
credit Cricsheet. The derived tables and this packaging are released under the same licence by the Doosra project.