betting_combat.rounds.market.trades

The raw trade layer under Sets 3 and 4: every Kalshi trade in the fight markets, on the dataset’s keys and the fight clock.

Ported from the research (ufc/dataset/trades.py: build).

trades_main   the two winner markets (win_a, win_b)
trades_side   every side market (distance, method of victory, method of finish, rounds,
              victory round)

Columns (both): fight_id, event_id, event_date, market (A-framed name), ticker, ts (UTC), yes_price, count (contracts), dollars (count x yes_price), taker (‘yes’: the buyer lifted an offer; ‘no’: the seller hit a bid), block, the clock at the trade: t_min, round, phase (settled / pre / post_bell / round / break), and trade_id: the trade’s number within its fight, over the fight’s tapes read in market-list order, each tape in its own order (a fight’s numbers do not depend on which other fights are built). Sorted by (fight_id, ts, trade_id), so trades in one millisecond keep the tape’s order whatever the sort; text columns are categorical. :func:time_order puts any tape in (ts, trade_id) order.

Functions

time_order

time_order(tape: pd.DataFrame) -> pd.DataFrame

The tape in time order, trades at the same ts in trade_id order (unique in a fight: no tie is left to the sort); a tape without trade_id keeps its own order among ties (stable sort).

trade_tables

trade_tables(fights: pd.DataFrame, markets: pd.DataFrame, tapes: Mapping[str, list[list[Any]]]) -> tuple[pd.DataFrame, pd.DataFrame]

(trades_main, trades_side) from the markets’ trade tapes.

Args: fights: :func:~.listing.live_fights rows. markets: :func:~.listing.market_list rows. tapes: ticker -> the tape’s trades ([ts_ms, yes_price, count, taker_yes, block], the trades list of mm/<ticker>.json.gz); a ticker without a tape is absent.

with_pulled

with_pulled(markets: pd.DataFrame, tapes: Mapping[str, Any]) -> pd.DataFrame

The market list with pulled: whether the market’s tape exists (live_markets).