Unusual Options Flow

WHAT IT READS. Recent trades on one crypto options venue for a symbol, tallying call against put notional across them, computing a notional put/call ratio, and isolating prints of $500,000 notional or more.

THE PREMISE, stated more carefully than the marketing version. Large trades MAY come from institutions or informed traders - and may equally be inexpensive low-delta positions, hedges, inventory transfers or legs of a structure. Size alone does not establish intent, and notional is not premium: a large-notional far-out-of-the-money trade can cost very little. There is genuine academic work here: Easley, O'Hara and Srinivas (1998) model informed traders CHOOSING BETWEEN the stock and option markets, and find that option volume can carry information about subsequent stock price changes under some conditions.

THE FOUR REASONS THE INFERENCE IS WEAKER THAN IT LOOKS, and a student should be able to produce this list unprompted.

• THE TAKER DIRECTION IS PUBLISHED HERE, and it still does not carry intent. Deribit's public trade records do include the taker's direction, so this venue is better than the general case. What it does not reveal is whether the trade OPENS or CLOSES a position, who the counterparty is, or whether the position is directional, volatility-driven or a hedge. A large call buy is as consistent with opening a bullish view as with closing a short call.
• MULTI-LEG STRUCTURES MAY APPEAR AS SEPARATE LEGS. Spreads, collars, risk reversals and calendars execute in several parts, and counted individually each leg looks directional - a risk reversal shows a large call print and a large put print that together express one position. This analysis counts each large trade on its own and does not relink legs, so a multi-leg position can read as separate directional trades, which misstates intent.
• MUCH LARGE FLOW IS A HEDGE against spot or against another option, and expresses no directional view on its own.
• THE COUNTERPARTY IS OFTEN A MARKET MAKER, whose prints look identical and carry no view - but not always: two customers or proprietary traders can trade with each other, and the print does not say which happened.

SO A NOTIONAL PUT/CALL RATIO SUMMARISES ACTIVITY, NOT INTENT. Note also that notional weights by contract count rather than by risk. Far-out-of-the-money options have small delta and low premium, so a LARGE NUMBER of them produces large underlying notional while carrying very little actual exposure - which is how a notional ratio comes to be dominated by cheap lottery-ticket strikes. Weighting by premium paid, or by delta, would answer different and arguably better questions.

THE THRESHOLD IS THE ANALYSIS. Which prints count as institutional is set by the $500,000 cut-off, and 'unusual' is set at three or more such trades. Both are conventions, and moving either changes every result - so compare readings for the same asset over time rather than across assets with different typical trade sizes.