Reading options conditions across the whole market
The problem
Screening for iron condors, strangles and 0-DTE plays by hand is hours of analysis for a handful of candidates, and only across the names you can watch.
The approach
Analyse volatility, Greeks and conditions continuously, surfacing candidate structures across the optionable universe — condors in ranges, strangles into expansion, 0-DTE on event days.
What changes
Let conditions surface the structure continuously, rather than once the spreadsheet is finished.
What this playbook covers
Building the screen itself, rather than choosing a structure for a name you already like. The order of the filters matters more than any individual one, and the step most often left out is the last: whether the candidate adds risk you already have.
Liquidity first, and it is not close
The instinct is to screen on volatility, because volatility is what the trade is about. That ordering is wrong, and it is expensive. A four-legged defined-risk structure crosses the spread four times going on and potentially four times coming off. On a chain with wide markets that cost can consume most of the credit before the position has done anything at all.
So liquidity is not a hygiene check applied at the end to the candidate you have already fallen for. It is the first filter, because it determines whether any edge you find is reachable. The most attractive-looking volatility in the market is often attractive precisely because nobody can trade it.
What to require before looking at volatility at all
- Chain volume and open interest: Not just in the name, but at the strikes and expiry you would actually use. A liquid front month and an empty back month is not a liquid chain for a calendar.
- Bid-ask as a share of credit: The number that matters is the spread relative to the premium you would collect, multiplied by the number of legs. Expressed that way, some apparently reasonable chains are obviously untradeable.
- Strike granularity: Wide strike spacing removes your ability to place short strikes where positioning suggests, which quietly turns a considered structure into whatever the chain permits.
- Exit liquidity, not just entry: You can usually get filled going on. The question is whether you can close at a fair price on a bad day, which is exactly when you will want to.
The funnel
Each stage should be cheaper to evaluate than the one after it, so the expensive analysis only ever runs on names that have already survived everything simple.
- Stage 1Tradeable chains only
Liquidity thresholds on volume, open interest, spread-to-credit and strike spacing. This removes most of the optionable universe immediately, and everything it removes was unreachable anyway.
- Stage 2Volatility that is unusual for this name
IV rank or percentile against the name’s own history, plus implied against realised. Absolute IV is close to meaningless across names — 40% is cheap in one and extreme in another.
Rank and percentile are not the same measure. Rank is where IV sits between its own high and low; percentile is the share of days it spent below. A single spike moves rank far more than percentile.
- Stage 3Term structure, and event exclusion
Inversion usually means an event is priced into the front month. Either you are deliberately trading that event, or you exclude the name — being long or short event premium by accident is how a screen produces its worst losses.
- Stage 4Positioning, to place the strikes
Open-interest walls, and whether price sits above or below the gamma flip. This is where a range-selling candidate can be eliminated entirely despite passing every earlier stage.
- Stage 5Portfolio fit
The stage most screens omit. A candidate that is excellent standing alone may add nothing but concentration to the book you already hold.
Stage five is where systematic screening usually fails
A screen run over the whole universe is very good at finding many candidates that all pass at once. That is the danger, not the benefit — because the reason they all pass at the same time is usually the same reason.
Twelve positions, one trade
Suppose the screen surfaces twelve premium-selling candidates on a quiet, range-bound afternoon. Each is individually well-formed: liquid chain, high IV rank, price above the gamma flip, short strikes outside the walls. Taking all twelve feels like diversification.
It is not. Every one of them is short volatility, and volatility is the most correlated thing in the market — in a broad expansion, they will all move against you together, on the same day, for the same reason. Twelve well-chosen short-vol positions are one large short-vol position wearing twelve tickers, and sizing them as independent is how accounts are seriously damaged in a single session.
The practical control is to size against total portfolio short-vega and against sector concentration, not against per-trade maximum loss alone. Per-trade risk limits are necessary and they are not sufficient, because they are blind to the thing that makes many small positions behave as one big one.
What a screen result should carry
Batch versus continuous
A daily batch run
- 1.Block out an hour, usually at the same time each day.
- 2.Pull chains for a familiar set of names.
- 3.Compute volatility measures for those that seem worth it.
- 4.Check positioning on the two or three that survive.
- 5.Act on what the screen said at the moment it was run.
Continuous screening
- 1.Liquidity, volatility and term structure evaluated across the optionable universe through the session.
- 2.Candidates surface when conditions change, not when you happen to look.
- 3.Positioning context attached rather than fetched separately.
- 4.Portfolio overlap evaluated against what you already hold.
The difference is not speed. A daily batch answers what was true at 4pm yesterday; conditions that appear at 11am and resolve by 2pm are invisible to it entirely.
- Did liquidity filter before volatility, or did I fall for the chain afterwards?
- Is IV rich against this name’s own history and against its realised movement, rather than merely high?
- Have I checked the term structure for an event I am not intending to trade?
- Do the short strikes sit outside the open-interest walls, on the supportive side of the gamma flip?
- How much short vega does the book already carry, and does this candidate add diversification or repetition?
- Would I accept the maximum loss on every correlated position at once, on the same day?
Where this goes wrong
- Screening on volatility before liquidity: The highest IV rank in the universe is frequently in a name whose chain cannot be traded at a fair price. Ordering the funnel wrongly wastes the analysis and occasionally produces a position that cannot be closed.
- Taking every candidate the screen produces: A screen that fires twelve times has usually found one condition, not twelve opportunities. Correlation is the hidden variable, and it is the one that turns a good month into a bad year.
- Confusing IV rank with IV percentile: They answer different questions and diverge sharply after a volatility spike. Using one while reasoning about the other silently biases the whole screen.
- Letting the screen decide rather than propose: Systematic screening is a proposal mechanism. Handing it the trade decision without judgement about regime and portfolio state removes the only part of the process that adapts.
What this does not do
Continuous screening changes which candidates you see. It does not change the odds on any of them, and it does not know your book, your risk tolerance or your view. Volatility regimes persist until they do not, and the transitions are where premium-selling structures take their largest losses — no screen anticipates that, because the whole point of a regime change is that the recent past stops being a guide. Options carry the risk of substantial loss, defined-risk structures included, and nothing on this page is a recommendation to trade.
Key Takeaways
- Filter liquidity before volatility — an edge you cannot trade at a fair price is not an edge.
- Order the funnel so cheap tests run first and expensive analysis only touches survivors.
- IV rank and IV percentile are different measures and diverge after a spike; pick one and reason in it consistently.
- Inverted term structure flags event premium — trade it deliberately or exclude the name.
- A screen firing many times has usually found one condition; correlated short-vol positions are a single position.
- Size against portfolio short-vega and concentration, not against per-trade maximum loss alone.