Order direction

side × outcome × tick — four actions, one price field, no shorting.

Omnibook orders use side × outcome × tick-in-own-outcome-space. That is deliberate: one price field, four user actions, and no shorting concept.

Field	Values	Meaning
side	buy | sell	Are you acquiring or disposing?
outcome	yes | no	Which contract?
tick	1–99	Price in cents of that outcome

The matcher normalizes to YES-space internally. You always quote in the outcome you named.

The four actions

side	outcome	What happens
buy	yes	Pay tick¢ to go long YES
buy	no	Pay tick¢ to go long NO
sell	yes	Sell YES you already hold
sell	no	Sell NO you already hold

There is no “short YES” as a separate verb — selling requires inventory. reduce_only refuses any order that would increase exposure.

Tick does not flip with outcome

A buy-YES at tick 47 and a buy-NO at tick 47 are both “47¢ for that outcome.” They are not the same book level. A NO bid at 47 rests opposite a YES ask implied at 53 (100 − 47). See Orderbook responses.

How fills report direction

On your fills and the public tape, direction is the taker’s (or your own order’s) origin:

Field	Where	Meaning
side / outcome	Own fills	Your order’s side and outcome
taker_side / taker_outcome	Public trades	The aggressor’s side and outcome
tick_yes	Public trades	Always YES-space; NO price = 100 − tick_yes

Why not Kalshi’s bid/ask vocabulary?

Kalshi is migrating toward outcome_side + book_side (bid≡yes, ask≡no) to escape a legacy two-price shape. Omnibook never had that shape — side and outcome map 1:1 onto the core NEW_ORDER message. New integrations should use the table above; there is no deprecated action/is_yes layer to migrate off.