How to filter options flow: the 7 settings that separate signal from noise

Why filtering is non-negotiable

On a typical trading day, the options tape generates hundreds of thousands of prints. A realistic breakdown:

• ~60–70% of all volume: retail orders of 1–10 contracts, primarily in high-IV names (TSLA, NVDA, SPY, QQQ)
• ~15–20%: market-maker hedging, rolling, and routine position management
• ~5–10%: institutional positioning (spreads, outright directional, portfolio hedges)
• ~2–5%: genuine informed directional bets with informational edge

The last category, 2–5%, is what most options flow tools are trying to surface. Without filtering, you're reading 95–98% noise to find the signal. Good filters don't just reduce quantity; they change the quality distribution of what you see.

Filter 1: Premium size (relative, not absolute)

The most common filter is a minimum premium threshold, only show prints above $X in total dollar premium. The problem with a fixed absolute threshold is that it ignores the context of the name being traded.

A $25,000 call sweep on a $400M market cap biotech, where the typical daily options premium is $15,000, is extraordinary. The same $25,000 print on NVDA, where $25 million in options premium trades daily, is invisible noise.

The correct approach: use relative thresholds tied to each ticker's 20-day average daily options premium, not a single absolute dollar figure. Most retail flow tools use absolute thresholds because they're simpler to implement, which means they systematically miss small-cap signals and surface large-cap noise.

Filter 2: Aggressor side (sweeps vs mid fills)

The aggressor side tells you who initiated the trade and how urgently.

• Sweep at ask (buy-side aggressive): The buyer paid the full ask price or swept across multiple exchanges to get filled, they wanted execution more than price improvement. This indicates urgency and directional conviction.
• Sweep at bid (sell-side aggressive): The seller hit the bid price. For put options, this means a buyer purchased puts aggressively. For calls, this suggests call selling (bearish or covered call writing).
• Mid-market fill: The trade filled between the bid and ask. Common for negotiated block trades, market-maker facilitated large orders, and closing positions. Less reliable as a directional signal.

The cleanest approach for a directional signal filter: show only calls at/above the ask for bullish signals, and puts at/above the ask for bearish signals. Mid fills can be added as secondary confirmation but should carry less weight.

Filter 3: Vol/OI ratio (new vs closing positions)

Volume/Open Interest ratio is one of the most underused filters in retail flow tools. It tells you whether today's options volume represents new positions being opened or existing positions being closed or rolled.

Practical filter setting: Vol/OI ≥ 5× surfaces new positioning. Below that, you're increasingly looking at routine position management. A call sweep at the ask with Vol/OI of 0.3× is almost certainly a position being closed, not opened, that's a potential bearish signal even though it looks like "call volume."

Filter 4: DTE range (expiration timing)

Days To Expiration tells you what timeframe the trader is expressing a view about.

Recommended DTE filter for directional signal flow: 14–60 DTE. This eliminates 0DTE retail noise at the bottom and LEAPS stock replacement at the top, leaving the expiration range where institutions express medium-term directional views with defined catalyst timing.

Filter 5: Sector / watchlist scope

An unfiltered flow feed shows activity across every sector and every name simultaneously. Unless you have a very specific reason to scan the full market, sector filtering dramatically improves the signal-to-noise ratio in your review session.

Two approaches:

• Sector-level filter: Show flow only for XLE + energy names, or QQQ + tech names, or XLV + healthcare names. Best when you have a sector thesis and want to see confirmation in the tape.
• Personal watchlist filter: Show flow only for the 20–50 names you actively follow. This sacrifices discovery but reduces noise drastically, you know the companies whose flow you're watching.
• Hybrid: Start with scored/ranked flow across the full market (EXTREME signals only), then drill into sector flow for names that appear. Discovery + focus.

Filter 6: Session timing window

Not all trading hours produce equivalent signal quality. The options tape has identifiable patterns in signal reliability across the trading day.

For a filtered high-signal-quality view: restrict to prints timestamped between 9:45–11:30 AM and 1:00–3:30 PM ET. This eliminates the open and close noise windows that account for a disproportionate share of false signals.

Filter 7: Flow score or confluence tier

A flow score combines the preceding filters into a single quality metric. Rather than manually checking each of the six filters above for every print, a properly built scoring system applies them and surfaces only the prints that pass multiple criteria simultaneously.

RadarPulse uses three tiers:

• EXTREME (85–100): Passes all major filters, high relative premium, sweep at ask, Vol/OI 10×+, DTE 14–60, quality session window. Multi-session follow-through confirmed. These are the highest-conviction institutional signals.
• ELEVATED (75–84): Passes most filters. May be missing one quality element (mid-market fill, slightly lower Vol/OI, or first-day with no follow-through yet). Still above retail noise floor.
• NOTABLE (60–74): Interesting but not confirmed institutional. Worth watching for follow-through but not acting on alone.

For a session focused on finding actionable signals: filter to EXTREME and ELEVATED only. This is typically 5–20 signals per session versus hundreds in an unfiltered feed.

Combining filters: preset configurations

Three practical preset configurations for different use cases:

Four common filtering mistakes

1. Using a single absolute premium threshold

A $25,000 minimum filters out nothing on NVDA (where this is background noise) while missing critical signals on small-cap biotechs (where this is 167% of daily average). Use relative thresholds tied to each ticker's typical volume.

2. Ignoring Vol/OI in favor of raw volume

High volume alone doesn't distinguish new positioning from position management. A 10,000 contract call volume with 100,000 OI (Vol/OI 0.1×) is positions being closed, potentially bearish. That same 10,000 contract volume with 500 OI (Vol/OI 20×) is aggressive new positioning. The direction of the interpretation is opposite with the same volume number.

3. Including 0DTE and weekly expiration prints

0DTE (same-day expiry) options are dominated by retail intraday speculation and gamma scalping. Including them dramatically inflates the false signal rate. Filter to 14+ DTE minimum for any signal you intend to act on.

4. Treating closing-hour flow as directional

The 3:45–4:00 PM window is the worst for signal quality. Index ETF hedging (as mutual funds and ETFs rebalance at the close), program trading, and MM delta rebalancing generate enormous options prints that look large and institutional but are purely mechanical. Excluding the final 15 minutes from your signal feed is one of the highest-impact single filter adjustments you can make.

Sector-specific filter calibration

One of the most persistent problems in options flow analysis is applying uniform filter settings across the entire market. What constitutes a meaningful print on NVDA is categorically different from what constitutes a meaningful print on a mid-cap biotech. The same absolute premium threshold that surfaces genuine institutional activity in one sector lets through background noise in another, while simultaneously filtering out real signals in a third.

Sector-specific calibration is the fix. Instead of one set of parameters applied globally, you adjust thresholds based on the liquidity profile, typical daily options premium, and behavioral patterns of each sector. The following calibration guide covers the major sectors and names where flow analysis is most commonly applied.

Technology mega-caps (NVDA, TSLA, AAPL, META, GOOGL, MSFT, AMZN)

These names run some of the deepest options markets in the world. Daily options premium in NVDA alone regularly exceeds $500 million. At that scale, a $100,000 sweep that would be extraordinary in a mid-cap name is completely invisible, it represents less than 0.02% of daily flow. The noise floor in mega-cap tech is enormously high, and your filters need to reflect that.

The minimum premium filter for mega-cap tech should sit at $500,000+, with the strong-signal threshold at $1 million or above. On Vol/OI, require at least 3× to distinguish new positioning, but 10× or higher is where you start seeing prints with no obvious explanation in existing OI. Aggressor side: sweeps only. Mid-market fills in mega-cap tech are almost always spreads, hedges, or market-maker facilitation. The sheer size of institutional desks means they can negotiate block executions at the mid without any urgency premium. A mid fill is not a directional signal in these names the way it might be in a thinly traded name.

DTE calibration also shifts in mega-cap tech: because these names have weekly expirations with high liquidity throughout the curve, the 14–60 DTE range remains appropriate for directional signals, but earnings-adjacent prints in the 7–14 DTE window should not be automatically dismissed, they can represent legitimate short-term positioning into a known catalyst.

Technology mid-caps (AMD, CRM, PLTR, SNOW, NET, DDOG, PANW)

Mid-cap tech occupies a middle zone: enough options liquidity to see genuine institutional flow, but not so deep that a $150,000 print disappears into the background. The calibration here is more nuanced than mega-caps.

Minimum premium: $150,000+. Strong-signal threshold: $400,000+. Vol/OI: 4× minimum, mid-cap tech has more concentrated OI (fewer strike-expiry combinations are liquid), so high Vol/OI ratios are somewhat more common and need a higher bar to be meaningful. The 4× threshold separates genuine new positioning from the routine strike concentration that happens in these names. Sweeps remain the primary aggressor filter, but large mid-market blocks above $500,000 are worth including, institutions sometimes prefer negotiated execution in mid-caps where sweeping would move the market.

Biotech small-caps

Biotech is the most challenging sector for threshold calibration. Daily options premium can be $10,000 on a typical day and $5 million on the day before a binary FDA readout. A single $25,000 print at baseline represents a 250% spike that any informed participant would note. The same $25,000 print on the day of the readout is irrelevant noise.

The correct approach for biotech: use a relative threshold of 20%+ of the trailing 20-day average daily options premium, with a hard floor of $25,000 to eliminate genuinely tiny prints. This relative framework automatically scales with the name's activity, the filter becomes more permissive in quiet periods (surfacing meaningful activity earlier) and more restrictive in binary event windows (preventing the event noise from drowning directional signals).

Vol/OI requirements for biotech should be higher, not lower: 10× minimum. Because biotech OI can be near zero for many strikes, even small volume can create high Vol/OI ratios. The 10× bar combined with a meaningful absolute premium floors the filter against false positives from tiny absolute prints.

Energy sector (XLE, XOP, and underlying names)

Energy flow has two distinct layers: the ETF layer (XLE, XOP, OIH) and the single-stock layer (CVX, XOM, SLB, HAL, MPC, etc.). These behave quite differently and require different filters.

At the ETF level, XLE and XOP carry significant institutional hedging flow related to portfolio exposure, a fund long energy stocks will often hedge via XLE puts. This means ETF flow carries more structural hedging noise than single-stock energy flow. For ETF signals, you need higher relative thresholds (75%+ of daily average) and sweeps-only. For single-stock energy, $75,000+ minimum premium is appropriate for large-cap names (CVX, XOM), but midstream and E&P names can signal meaningfully at $35,000+ given lower daily volumes.

Energy flow is also highly sensitive to macro timing. Flow preceding OPEC meetings, EIA inventory reports (Wednesday 10:30 AM ET), and API inventory releases (Tuesday 4:30 PM ET) carries more noise than flow in off-report windows. Apply tighter filters in these windows or interpret signals in that context.

Financials (XLF, JPM, BAC, GS, MS, regional banks)

Financial sector options flow is dominated by two distinct signal categories: rate-sensitive macro positioning (expressing views on Fed policy via bank stocks) and earnings pre-positioning (financials report in the first 2 weeks of each quarter). Calibrating for each matters.

For large-cap financials (JPM, BAC, GS, MS, BLK), $100,000+ minimum premium with sweeps-only is appropriate. For regional banks (FITB, RF, KEY, ZION, CFG), the threshold drops to $50,000+, daily options volumes are much lower and meaningful institutional prints occur at lower absolute levels. Vol/OI should be 5× for large-caps, 8× for regionals (where concentrated OI at major strikes is common and can depress the ratio).

Pre-FOMC flow in financials deserves special handling, covered in the dynamic filters section below. The key point for calibration: in the 48-hour window around FOMC meetings, financial sector options volume rises substantially and the signal-to-noise ratio drops. Tighten premium thresholds by 50% in FOMC windows.

Consumer discretionary (AMZN, HD, NKE, MCD, SBUX, TSLA)

Consumer discretionary has two distinct sub-regimes: the holiday quarter (October–December) and the rest of the year. During Q4, retail-related names see elevated retail investor options activity (appropriately enough), which elevates the noise floor and requires tighter filters. Outside Q4, the sector has more predictable flow patterns.

Standard calibration: $75,000+ minimum premium, 5× Vol/OI, sweeps-only. During the October through December holiday window: raise the minimum to $150,000+ for retail-facing names (AMZN, WMT, TGT, HD). The holiday options volume spike is driven by retail investors playing seasonal thesis trades in retail names, the noise is real and rises dramatically in October–November.

Dynamic filters: adjusting for market regime

Static filter thresholds assume a constant market environment. In practice, options market conditions shift substantially depending on volatility regime, earnings cycle, and macro event proximity. A filter configuration that surfaces 12 high-conviction signals on a calm Tuesday will surface 80 prints on a high-VIX event day, most of them noise. Market regime awareness is the difference between a filter that works on average and one that works on the specific day you're trading.

VIX-adaptive threshold adjustment

VIX is the most reliable single proxy for overall options market conditions. When VIX rises, overall options volume increases, more hedging, more speculation, more retail fear-buying. Your absolute premium thresholds stay constant while the surrounding noise level rises, degrading the signal-to-noise ratio of your filter output. The solution: scale thresholds up when VIX is elevated, down when it's suppressed.

A practical VIX-adaptive framework:

• VIX above 30 (high-volatility regime): Raise all premium thresholds by 50%. In a high-VIX environment, what was a noteworthy $100,000 print in a calmer market is now a routine transaction in a fear-driven tape. The institutional signals are still there, they're just larger, because institutions are trading larger size when market-moving events are in play. Only prints at 150% of your normal threshold have the same relative significance they would have on a calm day.
• VIX 20–30 (elevated, transitional): Raise thresholds by 20–30%. The market is stressed but not panicking. Adjust moderately.
• VIX 15–20 (normal range): Standard thresholds apply. This is the baseline calibration environment for most filter configurations.
• VIX below 15 (suppressed volatility): Lower thresholds by 20–30%. When VIX is in single digits or low teens, options premium is compressed across the board, institutional trades that would clear $500,000 in a normal VIX environment may only cost $300,000 in a low-VIX environment. Raising your filter to compensate for suppressed premium means you're systematically missing genuine institutional positioning. Lower your threshold to keep catching the same relative activity.

A simple rule: for every 5 points VIX moves above or below 18 (a rough baseline), adjust premium thresholds by ±10%. A VIX of 28 (10 points above baseline) raises thresholds by 20%. A VIX of 12 (6 points below baseline) lowers thresholds by 12%. This keeps your filter output volume roughly constant regardless of the VIX environment.

Earnings season calibration

The first two weeks of January, April, July, and October constitute earnings season for the majority of S&P 500 companies. During these windows, options volume spikes dramatically as institutions hedge earnings exposure, retailers speculate on binary outcomes, and market makers widen spreads to compensate for elevated gamma risk. The resulting noise is sector-specific and name-specific in ways that require active management.

The core rule: apply a strict earnings proximity filter during earnings season. Any name reporting within 10 calendar days should have its threshold raised by 100% (double the normal minimum) and should require a sweep-only aggressor filter. The elevated volume near earnings is dominated by hedge noise and event speculation, signals that pass your standard filter in this window have a dramatically higher false-positive rate.

One exception: names reporting more than 10 days out, in the same sector as names reporting this week, often show elevated pre-earnings flow as sector-rotation positioning. These prints can be genuine directional signals but require cross-sector context to interpret correctly. A tech infrastructure name seeing elevated call flow during hyperscaler earnings week may be a read-through trade, legitimate, but not the same as standalone positioning.

Post-earnings, the opposite dynamic applies: in the 3–5 trading days immediately after a company reports, options IV collapses (IV crush) and volume often surges as positions are closed or rolled. Prints in this window are heavily contaminated by position management. Wait at least 3 sessions post-earnings before treating flow in a name as a new directional signal rather than cleanup from the earnings event.

Pre-FOMC and macro event windows

Federal Reserve meetings, major CPI/PPI releases, jobs reports, and geopolitical events create predictable pattern distortions in options flow. Knowing how each event affects specific sectors is essential for correct filter calibration.

In the 48 hours before an FOMC meeting:

• Financial sector: Relax filters slightly (lower premium threshold by 10–15%). Financial sector options volume rises as institutions position on rate expectations, more of this elevated flow is genuine directional positioning rather than hedging noise. The increased volume is informative, not purely noise.
• Defensive sectors (Utilities, Consumer Staples, REITs): Tighten filters significantly (raise threshold by 40–50%). Rate-sensitive defensives see heavy hedge flow before FOMC as institutions protect existing portfolio positions. The vast majority of this flow is portfolio insurance, not new directional positioning.
• Technology: Standard filters apply, but be alert to long-duration tech (high multiple, rate-sensitive growth names like PLTR, SNOW, DDOG) where institutional hedging before rate decisions can generate large put prints that look directional but are portfolio-level duration hedges.

On CPI/PPI release days (typically the second week of each month), the opening 30 minutes sees extreme options prints across rate-sensitive sectors as the market reprices. Extend your session timing filter: start at 10:15 AM instead of 9:45 AM on CPI/PPI days. The additional 30 minutes of post-release price discovery eliminates the mechanical first-reaction prints that carry minimal forward information.

Quadruple witching and OpEx windows

The third Friday of each month (monthly OpEx) and quarterly quadruple witching (March, June, September, December third Fridays) generate the highest single-day options volume of any session. This is almost entirely mechanical, options expiring, positions rolling, dealers delta-hedging pinned names. Options flow analysis on OpEx day is nearly impossible for directional signals. The fix is simple: avoid looking for new signals on OpEx Friday. Focus instead on the Monday and Tuesday before OpEx, when institutional players are rolling positions and their flow is more often directional (expressing new positioning) than defensive (managing expiration mechanics).

The multi-session filter: the most underused signal amplifier

Every filter covered so far operates within a single session. You apply premium thresholds, aggressor side, Vol/OI, DTE, and timing to the prints visible on screen. But a single session is a snapshot. Institutional positioning at scale, the kind that produces the high-conviction moves that options flow analysis is designed to surface, is rarely completed in a single day. Large positions are accumulated over multiple sessions to minimize market impact and price improvement cost. The traders executing these positions are often splitting orders across days precisely because they do not want to appear on the tape as a single obvious print.

The multi-session filter is the tool for catching this pattern. It answers the question: have we seen this name, strike, and expiration cluster attracting elevated buy-side aggressor flow across multiple recent sessions?

What multi-session accumulation looks like

The pattern is distinct. On Day 1, a name passes all 7 single-session filters: meaningful sweep at the ask, relative premium above threshold, Vol/OI 8×, 30 DTE, mid-session timing, EXTREME score. Note it. On Day 2, the same name, often the same strike, sometimes one or two strikes away, shows another round of elevated call sweeps at the ask. Open interest at the target strike, which you can check the following morning in the OI snapshot, has increased. The position was not closed overnight; it was added to. On Day 3, OI at the strike increases again. By the third session with rising OI at the same strike cluster, you have multi-session confirmation of accumulation.

Three consecutive sessions with Vol/OI above 3× on the same strike cluster, combined with increasing OI each morning, is the operational definition of accumulation. This pattern has a substantially higher signal-to-noise ratio than any single-session filter configuration, because it's very difficult for noise to produce this pattern. Random retail activity, one-day hedges, and mechanical flows do not build increasing OI at a concentrated strike cluster over multiple sessions. Deliberate directional positioning does.

The OI confirmation signal

Open interest is reported with a one-day lag, the OI you see when markets open on Tuesday reflects positions held at the close on Monday. This makes it a confirmation tool rather than a real-time signal. The workflow:

• Day 1 close: note any EXTREME/ELEVATED prints that pass all 7 filters. Record the ticker, strike, and expiration date.
• Day 2 open: check OI at those strikes. If OI increased overnight, the Day 1 positions were held, the institution did not unwind at the close.
• Day 2 session: watch whether the same strike cluster attracts additional buy-side flow.
• Day 3 open: check OI again. A second consecutive increase = accumulation confirmed.

Flat or declining OI after an EXTREME print tells a different story: positions were closed, rolled, or the print was a closing trade. The signal quality of the original print degrades substantially if OI doesn't build.

Why most retail flow analysis misses this

The daily top-flow list format, "here are today's 10 biggest prints", is optimized for freshness, not for cross-session pattern recognition. It shows you what happened today; it does not show you whether today's print is the first in an accumulation series or the third consecutive session of building OI at the same strike. To catch multi-session patterns, you need either a persistent watchlist that carries forward flagged names, or a tool that explicitly tracks OI changes at flagged strikes across sessions.

This is one of the structural edges in institutional analysis versus retail flow scanning. An institutional desk watching a name they're positioning in will see their own prints and know the context. A retail flow analyst sees only isolated daily snapshots with no cross-session memory. Building that memory, even manually, dramatically improves the accuracy of flow-based signals.

How RadarPulse surfaces multi-session patterns

RadarPulse tracks cross-session flow at the strike level. When a name clears the EXTREME or ELEVATED threshold on two or more consecutive sessions at the same expiration cluster, it receives a multi-session follow-through flag. This flag is visible alongside the single-session score and factors into the overall confluence tier. Names with multi-session follow-through represent the highest-conviction subset of the already-filtered institutional signal feed.

The practical effect: instead of scanning 5–20 EXTREME signals per session and deciding which to research further, the multi-session layer narrows that to 2–5 names with confirmed accumulation patterns. That is the most actionable subset of the flow tape on any given day.

The 8 categories of false positives and how to filter them

Even with all 7 single-session filters applied correctly, some prints that pass will not be genuine directional signals. These are structural false positives, transactions that look like informed directional bets but are actually something else entirely. Understanding exactly what they are and how to identify them is the difference between accurate flow interpretation and pattern-matching noise.

Building your daily filter workflow

Knowing which filters to apply is necessary but not sufficient. The sequencing of filters matters enormously for efficiency. Applied in the right order, each step eliminates a large fraction of what remains, so later steps operate on a much smaller and cleaner dataset. Applied in the wrong order, you spend time on detailed analysis of prints that an early filter would have eliminated in seconds.

The following 10-step workflow sequences all filters, including the multi-session and false-positive checks, into a practical daily process. The timing estimates assume you are using a tool that applies the first seven filters automatically, and that the multi-session and false-positive steps require manual or semi-manual review.

The 10-step daily workflow

1. Step 1: Sector and watchlist scope (2–3 minutes)
   Start every session by defining your universe. If you have a sector thesis, set the sector filter before the market opens. If you're following a personal watchlist, activate it. If you want broad market discovery, set the filter to EXTREME scores only across the full tape. This single step typically eliminates 80–90% of all prints from consideration before you've even looked at a single signal. Do not skip it or try to review everything, the full unfiltered tape is unworkable.
2. Step 2: Premium threshold (instantaneous)
   Applied automatically if your tool supports relative thresholds. If you're using an absolute threshold, set it at the appropriate level for the names in your watchlist (use the sector calibration table above). This step cuts another 70–85% of what remains after Step 1. The combination of watchlist scope and premium threshold should reduce a day's tape from 50,000+ prints to a few dozen candidates.
3. Step 3: Aggressor side check (instantaneous)
   Filter to sweeps and blocks at/above the ask only. Any bid-side or mid-fill transaction that passed the premium threshold is either a seller (covered call write, put write) or a negotiated close (ambiguous direction at best). Remove them. This step typically removes 40–60% of what's left after Steps 1–2, and is almost free from a time perspective if your tool supports it as a filter.
4. Step 4: Vol/OI verification (instantaneous or 30 seconds per print)
   Check Vol/OI for each remaining print. Apply the appropriate sector minimum (4× for mega-cap tech, 5× for most names, 8–10× for biotech and regionals). Any print with Vol/OI below your threshold is primarily position management, closing, rolling, or hedging existing positions. These are not new directional signals. Removing them typically cuts another 30–50% of the remaining list.
5. Step 5: DTE range validation (instantaneous)
   Remove anything below 14 DTE (0DTE and weekly speculation) and anything above 90 DTE (LEAPS stock replacement). For a tighter filter, use 14–60 DTE. This step removes both the retail noise at the short end and the portfolio management flow at the long end. After Steps 1–5, you should be looking at a list of 5–25 prints on a typical day.
6. Step 6: Session timing check (instantaneous)
   Verify that each remaining print occurred in a high-quality session window: 9:45–11:30 AM or 1:00–3:30 PM ET. Remove any print from the first 15 minutes of the session, the lunch lull (if you prefer higher signal quality), or the final 15–30 minutes before the close. On macro event days (CPI, FOMC), adjust the session start to 10:15 AM. This step typically removes 15–25% of remaining prints.
7. Step 7: Flow score application (instantaneous)
   If your tool provides a flow score or confluence tier, filter to EXTREME and ELEVATED only. This acts as a catch-all for any multi-factor quality check your tool applies that isn't explicitly covered by the manual steps above. If you've applied Steps 1–6 manually, the flow score serves as confirmation that you haven't missed a disqualifying factor. After all 7 steps, you should have 2–15 prints worth reviewing in depth.
8. Step 8: False positive screening (1–3 minutes per print)
   For each print that survived Steps 1–7, run the false-positive checklist quickly:
   , Is the company reporting earnings within 10 days? (If yes, treat any puts with elevated skepticism; note that large calls may be pre-earnings bullish positioning, but threshold accordingly.)
   , Is there a matching put print (call volume) or call print (put volume) at the same expiration within the same session? (If yes, possible straddle/strangle, verify timing and premium balance.)
   , Does the fill appear to be at the bid (call writes, put writes)? (If your aggressor filter is tight, this shouldn't appear; if you included mid fills, double-check.)
   , Is it VIX settlement Wednesday morning? (If yes, disregard early SPX prints.)
   , Does the date coincide with index reconstitution or OpEx? (If yes, confirm single-name catalyst before treating as directional.)
   This screen takes 60–90 seconds per print once you have the checklist internalized. A print that passes all five false-positive checks has a substantially higher probability of being genuine institutional positioning.
9. Step 9: Multi-session confirmation (2–4 minutes)
   Check your prior-session notes or your flow tool's historical data. Has this name, or a closely related strike cluster in this name, appeared on your filtered list in any of the past 3 sessions? If yes, check overnight OI: did OI at the target strike increase? A print that represents the second or third consecutive session of elevated buy-side flow at the same strike cluster, with increasing OI confirming position accumulation, is your highest-priority signal. A first-session print with no prior context is an unconfirmed signal, worth watching but not acting on alone unless the other filters are extremely strong.
10. Step 10: Fundamental and technical context (3–7 minutes per name)
    The final step, and the one most flow analysts either skip or do first (both mistakes). For each print that survived all 9 prior steps, verify that a plausible thesis exists to explain the positioning. Does the technical chart setup support the direction implied by the options flow? Is there an upcoming catalyst (earnings, FDA date, regulatory decision, macro report) in the expiration window? Is there recent news or analyst activity that could explain interest in this name? A print with a supporting fundamental story and a clean chart setup is a tradeable signal. A print that passed all 9 filter steps but has no identifiable thesis should remain on your watchlist for follow-through, not in your active signal list. Options flow is a leading indicator, but leading indicators still need a mechanism to make them actionable.

Daily time allocation

The full 10-step workflow should take 15–30 minutes on a typical day, less if your tool automates Steps 1–7, more on high-volume or event days when more prints survive the automated filters and require manual review. A practical allocation:

Executed consistently, this workflow transforms options flow from a firehose of noise into a structured research process that yields 2–5 high-conviction signals per day, a number small enough to research properly, large enough to find actionable setups across a variety of market conditions.