README.md

metadata

license: cc-by-4.0
language:
  - en
pretty_name: SmartStake MLB Lineup Reactions (2026)
tags:
  - sports-betting
  - sports-analytics
  - mlb
  - baseball
  - odds
  - player-props
  - lineups
size_categories:
  - 10M<n<100M
task_categories:
  - time-series-forecasting
configs:
  - config_name: default
    data_files:
      - split: train
        path: mon=*/*.parquet

SmartStake MLB Lineup Reactions (2026)

Full-resolution sportsbook odds movements around every Underdog MLB lineup post of the 2026 season, for the six player-prop markets a batting-order change moves. Every row is one book's price for one selection at one moment, in a window around a lineup drop. This is the raw material behind the study "MLB Lineup Drops", a companion to SmartStake MLB Player Prop Odds and Results.

Coverage

• Events: 2,456 Underdog MLB lineup posts with odds, 25 March through 17 July
  2026. Each post is one team's lineup; the All-Star Game is excluded.
• Window: −5 to +20 minutes around each post (offset_sec), so both the pre-drop baseline and the full reaction are present.
• Markets: total bases, hits, RBIs, runs, hits + runs + RBIs, strikeouts.
• Books: 73 sportsbooks, exchanges, prediction markets, and DFS pick'em products.
• Rows: ~85 million (one book's price for one selection at one moment).
• Coverage gap: the odds archive was down on 26 May and from 20 to 29 June 2026, so lineup posts on those days (about 11% of the season's posts) have no odds and are absent.

Schema

column	type	description
event_id	string	One lineup post: team~tweet_ts. Group on this.
team	string	The team whose lineup posted.
tweet_ts	timestamp	When Underdog posted the lineup (UTC).
game_id	string	Stable per-game identity.
player	string	Canonical player name.
player_id	int64	Stable player id (consistent across books).
market	string	bases, hits, rbis, runs, hits + runs + rbis, or strikeouts.
line	double	The over/under number for this selection.
side	string	over or under.
book	string	Sportsbook / exchange / DFS product.
engine	string	Pricing engine; collapses shared-brain skins (see notes).
ts	timestamp	When the quote was live (UTC), full resolution.
offset_sec	int	ts − tweet_ts, in seconds. Negative is before the post.
odds	double	Decimal odds. Null for DFS pick'em products (see notes).
is_suspended	bool	true when odds ≤ 1.02: the market was pulled, not a real price.
is_alt	bool	true for an alternate (non-main) line.

Notes and caveats

• Suspensions are kept but flagged. A book (Pinnacle especially) often pulls a market before reposting a new number; those ticks carry odds ≤ 1.02 and is_suspended = true. Filter them out before treating odds as a price.
• Skins are collapsed by engine. Books that share one pricing brain get one engine value so a single book is not counted several times: betmgm covers betmgm, bwin, and Sports Interaction; pinnacle covers pinnacle and its sister site. Every other book is its own engine.
• The window is the whole game. A post names one team, but the window carries both teams' player props for that game (a lineup change moves the team's batters and the opposing pitcher's strikeouts). Filter by player/player_id for the players you care about.
• DFS pick'em products are included, with a null price. PrizePicks, Underdog, Sleeper, DraftKings Pick6, Betr, SplashSports, OwnersBox, Dabble, and FanDuel Picks set a line without a two-sided price, so their odds is null. Their line and how it moves are still the signal, so the rows are kept.

Loading

import pandas as pd
df = pd.read_parquet("hf://datasets/SmartStake/mlb-lineup-reactions/mon=2026-05")

-- DuckDB, straight from the hub
SELECT book, count(*) FROM 'hf://datasets/SmartStake/mlb-lineup-reactions/**/*.parquet' GROUP BY 1;

Reproduce a finding: which books reprice first

For total bases, the median time each book takes to make its first real price move after a lineup post (lower is faster), straight from the hub with DuckDB:

import duckdb
duckdb.sql("INSTALL httpfs; LOAD httpfs;")
print(duckdb.sql("""
WITH src AS (
  SELECT * FROM 'hf://datasets/SmartStake/mlb-lineup-reactions/**/*.parquet'
  WHERE market='bases' AND side='over' AND NOT is_suspended AND NOT is_alt),
seq AS (                                   -- consecutive quotes per selection per book
  SELECT event_id, engine, player, line, offset_sec, odds,
         lag(odds) OVER (PARTITION BY event_id, engine, player, line
                         ORDER BY offset_sec) AS prev
  FROM src),
firstmove AS (                             -- first >=1pt implied-prob move after the post
  SELECT event_id, engine, min(offset_sec) AS move_off
  FROM seq
  WHERE offset_sec >= 0 AND prev IS NOT NULL
    AND abs(1/odds - 1/prev) >= 0.01
  GROUP BY 1, 2)
SELECT engine AS book, count(*) AS reactions, round(median(move_off)) AS median_seconds
FROM firstmove GROUP BY 1 HAVING reactions > 300 ORDER BY median_seconds
""").df())

The full study centres each book against the field per event to strip coverage and polling-cadence bias; this pooled version is a directional first look. Writeup: https://smartstake.app/learn/mlb-lineup-drops

Provenance, license, and responsible use

Odds were collected from public sportsbook and exchange feeds. Released under CC BY 4.0 for research and educational use. This dataset is informational and historical: past prices do not predict future outcomes, and betting carries risk of loss. Not affiliated with any sportsbook. 21+.

Citation

SmartStake (2026). SmartStake MLB Lineup Reactions (2026). Hugging Face.