That’s the number. Here’s what it means.

The setup was valid.

The entry was a Trend Continuation signal on NVDA in a Chop regime. This is one of the framework’s OOS-Watch edges. OOS stands for out-of-sample: it means the edge has been tested on historical data the framework wasn’t built on, a separate slice reserved specifically to check whether a pattern holds outside the period used to find it. OOS-Watch means the out-of-sample results are favorable, but the sample is small enough that the edge hasn’t earned full validation status yet. It’s a probationary tier. Higher confidence than a fresh idea, lower confidence than an edge with years of clean OOS data behind it.

The pre-entry score came in at 80. Regime classification was correct. The signal was real.

Under the rules in force on May 14, this was a legitimate entry. I took it because the framework said to take it.

The position stopped out Friday morning on a broad risk-off gap. Trump-Xi headlines overnight, VIX up 10%, the semiconductor cluster gapped down at open. The stop filled on first liquidity. Clean execution on an ugly morning.

What the debrief found.

The R42 debrief process runs the same way after a loss as after a win. You go back through the entry conditions and ask whether anything in the setup predicted the failure.

Two things came back.

VIX at entry was 17.92. The framework’s OOS analysis of this edge shows a recurring failure bucket: entries with VIX between 16 and 20 carry meaningfully worse outcomes than entries with VIX below 16. The position was entered in the middle of that failure bucket.

DTE to NVDA earnings was four trading sessions. The framework requires a minimum buffer between the position hold window and the company’s earnings date. Four sessions is inside that buffer. The position was structurally exposed to earnings-adjacent volatility from the moment it opened.

Neither of these is hindsight bias in the dishonest sense. The OOS failure signatures for this edge were documented before the trade was entered. What wasn’t done was converting those signatures into hard entry gates. The trade was rule-adherent under the framework as written on May 14. The debrief surfaced that the framework as written on May 14 had a gap.

What changed.

As of v3.7.7, two gates now apply to this edge:

VIX at entry must be below 16. T11’s entry VIX was 17.92. Under current rules, this trade does not get entered.

DTE to earnings must be six sessions or more. T11’s DTE to NVDA earnings was four. Under current rules, this trade does not get entered.

Both gates were added the same day the stop filled. Not as a reaction to losing money, but because the failure conditions matched a documented pattern, and a documented pattern with a clear gate solution gets gated.

The framework’s response to a loss isn’t to stop trading the edge. It’s to characterize what went wrong, check whether it matches a known failure mode, and add a rule if the data supports one. T11 supported two.

The honest accounting.

This was the framework’s first live trade on this OOS-Watch edge. The lower confidence tier exists for exactly this reason: thinner evidence means more variance, and more variance means early live trades carry more information than later ones will. T11 added one live data point to an edge with three prior OOS samples. The data point was a loss in a specific condition set, and that condition set now has a gate.

The loss cost $490 CAD. The two gates it produced will apply to every future entry on this edge. Whether that tradeoff is positive depends on how those gates perform over the next twenty trades. The framework’s bet is that they will. That bet is based on the OOS failure signatures, not on one bad morning.

Two trades, one week.

Tuesday: MRVL, 18 minutes, T3 exit, +$742 CAD. The exit architecture executed exactly as designed. (Full breakdown in the prior post)

Friday: NVDA, one session, stop at open, -$490 CAD. The framework learned something it didn’t know before and added two rules.

Net on the week: positive. Net on the framework: better than it was Monday.

That’s what a systematic process looks like from the inside. Not a clean win rate. A feedback loop.

If that framing interests you, the Friday weekly review covers how the week sits against the broader framework context. Free, in the archive.

Subscribe to follow the framework as it develops. Every trade documented, every rule change explained, wins and losses both.

Helio documents a retail options framework in public. Posts are educational case studies of the author’s own trades, published after the fact. Nothing here is investment advice or a recommendation to buy or sell any security. Full disclaimer at helioletter.com/disclaimer.

Here is what I was not doing during those 18 minutes: making decisions.

The entry isn’t usually the problem in retail options trading. The hold is. Conviction takes over during the hold, and conviction doesn’t have a good track record on exits. The ways a winning trade becomes a disappointing one aren’t exotic or rare. Every experienced options trader has lived at least one version.

Decision point one: the position is up 30% in the first few minutes of the session. The rational thing is to take it. Something could reverse. The gain is real and the continuation is speculative. The trader who exits here isn’t wrong about the logic. They’re just capturing approximately a third of what the full thesis produces.

Decision point two: the position is up 60-70%. Two failure modes are running simultaneously. One voice says exit before it gives back. The other says this is moving, let it run past the original target. Both are plausible. Neither has a defined answer. The trader who acts on either one is making an in-the-moment decision about a situation they can’t fully read from the inside.

Decision point three: the position has hit the original target and is still moving. No exit order in place. Conviction says the momentum is real. What the trader doesn’t see is that they’ve exited the trade they entered and opened a different one. Not a defined-risk entry with a target. An open-ended speculation on continuation in a time-decaying instrument.

The framework avoided all three on Tuesday. Not because the system predicted the magnitude of the move. Because the exit was placed before the position opened.

The trade

MRVL May22 $180C. Entry at market open May 13, exit at 9:54 AM. Eighteen minutes.

The setup was PB (Pullback Bounce) in a Chop regime. No binary catalyst, no earnings inside the hold window, no sector event driving the move. Pure price structure.

The underlying had declined sharply in the prior session, found intraday support, and recovered to close above a pre-defined indicator threshold. The following morning’s gap in the direction of the bounce confirmation was the entry trigger. The scan flagged the setup the prior session as the bounce structure was establishing. Entry occurred at market open when the position confirmed rather than chasing into the gap.

Gates that passed at entry: regime classification (Chop), volatility environment within range, no earnings in the hold window, no open positions creating correlation conflicts. The setup pattern matched a backtested edge pairing for this ticker in this regime, which applied a scoring bonus. The R39 chase check cleared before entry.

Sizing: one contract, near-money call, 9 days to expiration at entry. The shorter expiry wasn’t a discretionary choice. Longer-dated near-money calls on MRVL at this price level exceeded the position cap threshold. The framework’s sizing architecture determined the instrument. The 9 DTE created structural urgency that a 30-DTE position wouldn’t have had, which is part of why the trade resolved in 18 minutes rather than three sessions.

What happened during the hold followed R23’s tier progression without deviation.

T1 fired within the first five to ten minutes. T1 in the framework is informational: the gain gets marked, the monitoring status updates, no exit action fires. The temptation at T1 is to take the win and close early. The rule says noted, keep going.

T2 fired shortly after. At T2, the live stop ratcheted to the entry premium. The position had a zero-loss floor from that point forward. A conviction trader here would have either already exited or made no stop adjustment at all. The framework required neither conviction nor discipline to execute the ratchet. The instruction was in the rules before the position opened.

T3 GTC was in place from entry. T1 fired and the gain was marked. T2 fired and the stop moved. When T3 filled at 9:54 AM, the position had managed itself from open to close. The fill landed inside the T3 zone, slightly below the strict T3 level. This is normal options execution. T3 exits land in the T3 zone, not on the exact number.

Total hold: 18 minutes. Clean T3 exit, no override, no intervention.

The counterfactual

A conviction-based trader watching this position would have faced three decision points, each with a plausible rational action that would have produced a worse outcome.

Decision point one, position up 30-40% within the first few minutes. Conviction says: the move is real, take the gain. The trader who exits here captures approximately a third of what the framework captured. The entry was right and the setup worked. The P&L shows a fraction of it.

Decision point two, position up 60-70%. Two failure modes run simultaneously. One voice says exit before it gives back. The other says this is a real move, hold past the original target. The trader who takes the first voice exits and leaves 30-40% on the table. The trader who takes the second is now holding without a defined exit, running an open-ended position on a time-sensitive instrument. Both outcomes are common. Neither required being wrong about anything.

Decision point three, position at the original target and still moving. No GTC order in place. Conviction says the momentum is real. Sometimes this works. More often the position consolidates, theta accelerates, and the exit happens below T3 after the gain has been watched shrink for twenty minutes.

The framework avoided all three of these with the same mechanism: the T3 GTC limit placed before the position opened. The decision was made once, not three times during the hold.

The honest debrief

The R42 debrief on a winning trade runs the same pattern as on a losing one. Two categories matter: what the framework caught correctly, and what was luck rather than system.

What the framework caught: the regime classification was accurate (the Chop read held through the prior session’s sharp decline), setup recognition was clean, and the exit architecture executed without intervention.

What was lucky: the speed and magnitude. PB setups in this framework have a backtested expectancy that implies a gain when they work, but 18 minutes is not the average. The framework doesn’t model intraday velocity. The position could have taken three or four sessions to reach T3, with multiple consolidation periods and more opportunities for conviction to override the plan. The speed was fortunate. The T3 order being in place before the move was not.

This distinction matters because the lesson of the trade isn’t “enter correctly and the framework rewards you quickly.” The lesson is that the framework’s output was the same whether the hold took 18 minutes or four days. The T3 GTC doesn’t care how long it takes to fill.

The debrief also surfaced an unresolved tension in the rules. The framework includes a protocol (R51) for short-dated positions that fires when unrealized gain reaches a threshold, instructing the trader to take 50% off. At one contract, a partial exit isn’t possible. The rule fired the monitoring flag but had no executable action. The candidate resolution under consideration: at one contract, R51 firing means ratcheting the stop to the T1 level rather than taking a partial, which preserves the T3 target while protecting the majority of the T1 gain. This hasn’t been formally adopted yet. The trade surfaced the gap, and the fix is being worked out.

The transferable insight

Pre-commitment is the mechanism. Not willpower, not the discipline of watching a winning position and resisting the urge to close it.

The discipline required here was placing a limit order before the position opened and not canceling it. That act takes about 30 seconds and doesn’t feel like discipline while you’re doing it. It feels like setting up a trade.

The trade manages itself from there.

Two questions worth sitting with, even outside the specific framework.

When you have a winning options position, do you have a pre-defined exit with an order in the market, or do you have a number in your head that you’re watching? The gap between those two things is where most winning trades give back their gains.

When you opened your last winning options trade, did you place the exit order before the position started moving, or after it was already up? The answer tells you whether you’re managing exits or reacting to them.

The framework doesn’t predict which trades will win. It defines what “enough” means before the trade starts. An 18-minute hold and a clean T3 fill is what that looks like when the tape cooperates and the exit was already in the market.

The best trade you ever make won’t feel like discipline. It’ll feel like nothing.

This is the most expensive pattern in retail options trading, and the most common.

The problem is not that the trader lacks discipline. The problem is that the discipline architecture does not exist yet. Without a structural answer, the question of when to close a winner gets re-litigated in real time on every trade, against the social pressure to “let winners run” and the simple human bias to imagine a chart continuing in the direction it is currently moving. That fight has a known winner over enough trades, and it is not the trader.

R23 is the framework’s structural answer. It is the rule that decides what happens to a winning trade once the trade is in meaningful profit. It makes the decision in advance, encodes it as mandatory framework behavior, and surfaces it loudly enough that “forgetting” is not an available outcome.

R23, in plain language

R23 is the framework’s tiered profit-taking ladder for long options positions. It defines three calibrated premium-gain levels, each with a mandatory action attached. T1 is informational. T2 ratchets the stop-loss to entry premium. T3 forces a full exit, no override.

The rule’s force does not come from the specific tier percentages. It comes from making the tiers visible and the actions automatic. The calibrated levels and their dollar values are printed in the trade record from the moment the position opens. Each tier’s action is mandatory rather than discretionary. The daily scan output keeps live tier status visible next to every open position. The trader does not have to make the “should I sell?” decision in real time, because the decision was made at entry and the limit was already placed in the market.

R23 lives in the exit-ladder layer of the framework. It runs in parallel with the time-and-thesis layer (R5 hard exit before earnings, R56 catalyst-failure exit, R43 thesis-break exit) and the downside-stop layer (R61). When a winning trade is in profit territory, R23 is the active governor. When something faster fires, R23 yields. R23 says nothing about entry, sizing, or losing trades. It is purely about how to terminate a winning position. That narrow scope is deliberate.

Here is how each tier actually works, and what that looked like on a real trade.

Read more

This is the post about that trade.

On Monday Apr 27, Helio Scan flagged a CRWD May 15 $480 call. The setup was trend continuation: the underlying had pulled back into a defined re-entry zone after a prior leg up, the software tape was constructive, and the directional thesis was that a dovish-lean Fed at the Wednesday FOMC would extend the rate-sensitive growth complex. Trend continuation, in framework terms, means the system is reading a pause in an established uptrend as a re-entry rather than a top. It is one of three primary setup classifications the scan uses, the other two being earnings-driven IV expansion and event-driven catalyst plays.

The framework’s entry gates all opened. (The scan grades trades pass/fail at multiple checkpoints; this trade cleared every one.) Macro and structural gates were open. Volatility sat below threshold. The geopolitical binary was dormant. No anti-chase trigger fired on the entry session. The chart-screenshot hierarchy confirmed the setup against a fresh price read rather than a stale watchlist quote. Every objective check the framework runs at entry passed.

Sizing was below cap. Single contract, May 15 expiry, $480 strike. The minimum-premium floor was respected.

Sessions one and two played out within expected ranges. The premium drifted modestly without committing to direction. Nothing fired. Nothing nearly fired. The catalyst sat inside the hold window. The thesis was waiting on the event.

Wednesday was the FOMC. The Fed held rates as widely expected. The dovish lean that the directional thesis required was already in the curve, so the dovish hold was a sell-the-news event rather than a tailwind. The underlying drifted within a one-percent band of pre-event spot through the four hours after the release. Volume did not confirm a move in thesis direction. The premium decayed materially into the close. Meta reported after the bell, the capex guide was a shock, and Thursday’s open inherited a tech tape that didn’t fit the entry thesis.

The position closed on a stop fill Thursday afternoon. The GTC fired clean, no slippage worth flagging. A clean stopped-out loss within the personal hard-stop band that the framework uses as a backstop when no faster exit rule fires.

The part that matters: every rule that applied to this trade was followed. The entry was scored objectively. The stop was placed correctly. The stop was honored without hesitation. The debrief was filed inside the system’s 24-hour window. There was no discipline failure. There was no narrative rescue, no holding-and-hoping past the level. The trade lost money cleanly and on rule.

So what was the problem?

The framework had a gap.

The exit ladder, before this trade, had four primary rungs. A price-driven thesis-break exit, which fires when the underlying moves against the position by a defined amount in a defined window. A calendar-driven pre-earnings exit, which closes any open position before the company reports. A dead-trade flag, which surfaces positions drifting toward expiration without meaningful progress to breakeven. An outsized-early-move re-evaluation, which forces a decision when a position runs hard in either direction inside the first few sessions. Below all of these sits the personal hard stop, the GTC backstop that catches anything the faster rules miss.

What the framework did not have was a catalyst-failure exit. A named binary catalyst inside the holding window passed without confirming the directional thesis, and there was no rule that said “exit now.” The trade was carried from Wednesday’s non-reaction into Thursday’s drift into the PM stop, on a thesis that was structurally invalidated by mid-morning Thursday at the latest. The hard-stop backstop did its job. It just did the job too slowly for this specific failure mode.

This trade was, in retrospect, a hybrid setup being scored against a pure rubric. It was classified as trend continuation at entry, but a named binary catalyst sat inside the planned hold by design. That is operationally distinct from both a pure trend-continuation trade (where the catalyst, if any, sits outside the hold window) and a pure event-driven trade (where the position is sized and timed around the binary itself). The framework had been collapsing all three structures into the existing classifications without flagging the catalyst-modifier as its own risk vector. The loss surfaced the gap.

The rule that came out of the debrief is R56, the catalyst-failure exit. The structure: when a named binary catalyst passes inside the hold window without the underlying moving in thesis direction beyond a defined spot-reaction threshold, with volume confirmation against a defined multiple of average, inside a defined post-event reaction window, the position exits at market regardless of P&L.

The threshold values, the volume multiple, and the reaction-window duration are calibrated and live in the system. They are not in this post. The structure of the rule is the educational content. The parameter values are the execution edge.

Applied retroactively to this trade, R56 would have triggered exit on the Apr 29 close, the same session as the catalyst event itself, a full session before the Apr 30 stop fill. The new rule shortens the exit window by one session in this specific failure mode.

A companion change went into the daily scan. The watchlist and catalyst-calendar sections now flag “binary inside hold” as a risk modifier on every session of every active trade, not only at entry.

This is what “framework” actually means, and it is the distinction the manifesto last week was pointing at. A signal is a one-shot recommendation. A framework is a system that updates from its own evidence. The CRWD trade is the origin entry in the rule-change log for R56. Future trades that meet the rule’s conditions will exit faster than this one did, because this one didn’t.

A note on the week’s tape, since it explains why a rule like this needed to exist before it did. The week of Apr 27 to May 1 was a high-density catalyst window. Mega-cap earnings stacked alongside the FOMC inside three sessions, with a softening jobs picture at the end. Volatility was contained sub-threshold through Monday and Tuesday, then ticked up Wednesday afternoon as the FOMC dovish hold proved a sell-the-news event and Meta’s capex guide produced a single-stock reset that bled into the tech tape. A widely circulated weekend report on AI-economics softness reframed the sector through the week. The macro and volatility gates stayed nominally open most sessions, but catalysts inside hold windows became the dominant structural feature of every active and proposed position. The framework was reading “open gates” but executing into a tape where binary events were stacked tight enough that almost every trade was an event-driven trade in disguise.

When binaries cluster, “trend continuation” becomes “trend continuation through a binary,” and the framework that worked in calmer environments needs an additional rule to handle the failure mode where the binary fires and the trend doesn’t continue.

R56, the catalyst-failure exit, is now in the framework. The paid tier opens Sunday.

It’s a signals service that won’t show you its math, just the calls. It’s a hype account posting cropped P&L screenshots from the green days and quietly archiving the red ones. Or it’s a paid Discord room where the value proposition is being in the room, watching someone else’s screen, hoping the proximity rubs off.

I’m not writing this to tell you those services are run by bad people. Most of them aren’t. The model is broken regardless of who’s running it.

The model is broken because it sells the wrong thing. It sells the answer when the answer is the easy part. The hard part is the question, and the discipline, and the rules that govern what you do when the trade you took stops behaving the way the signal said it would.

Helio is the opposite of that model.

The thesis

Most retail options content collapses two decisions into one. “Buy this” is implicitly “until I tell you to sell.” That’s a category error.

Entry and exit are independent problems. They use different inputs. They demand different discipline. The signal that gets you into a trade is almost never the signal that should get you out.

The conflation is what blows retail traders up. They take someone else’s entry, paid for or scraped from a screenshot, and improvise their own exit. They watch a 30% gain become a 15% gain become a flat trade become a 20% loss, because nobody told them what the exit rule was and they were too anchored to the entry to figure one out in real time.

You can see this happen on any options-trading subreddit, any week of the year. Someone posts a screenshot at the top of a position bragging. They post another one a week later asking the room what to do. By the time they ask, the answer doesn’t matter. The trade was lost the moment they took someone else’s entry without owning the exit.

Helio’s framework refuses to make that mistake. Every published trade has both rules visible. Both were decided before the trade. Not after.

What the framework actually is

Helio Scan v3.1 is a written, versioned framework for trading pre-earnings momentum on liquid US equities. Pre-earnings momentum, specifically, because that’s the regime where the framework has been tested, refined, and stress-tested against actual trades. It does not trade on FOMC days. It does not trade meme-driven retail squeezes. It does not chase. Most months, it makes two to four trades. Most weeks, it makes zero.

The entry filter is a small set of conditions that all have to be true at once. Volatility, momentum, structure, and a check on the implied-volatility surface to make sure the option is priced in a way that gives the trade a defensible expected value. If even one condition fails, the trade doesn’t happen. The framework would rather miss a winner than force a marginal entry.

The exit filter is independent. It governs the trade from entry through close, with predefined rules for adverse moves, time decay (theta works against long premium, fast), and the boundary condition of earnings itself. The framework exits before earnings, every time. It is not in the business of holding through a binary event.

Both filters are written down. Both are versioned. Both can be argued with. That’s the point.

The framework is conservative in scope and aggressive in discipline. It does one thing. It does it the same way every time. When the conditions don’t line up, it doesn’t trade. Sitting on hands is not weakness. Sitting on hands is the rule working.

What I won’t pretend

Helio is calibrated to my account, my risk tolerance, my time of day, my tax jurisdiction. I’m a Canadian retail trader. None of those things describe you.

The framework is shareable as methodology. It is not shareable as instruction. A reader with a different account size, a different volatility tolerance, a different schedule, or a different tax setup would build different rules. The shape of the framework would translate. The specifics wouldn’t.

This is what most paid newsletters won’t tell you. The rules that work for the writer almost certainly don’t transfer cleanly to the reader’s account. Anyone who tells you they do is selling you confidence, not edge.

This is why every trade Helio publishes is post-hoc. By the time you read about it, the position is already entered and exited. You can’t act on it. That’s not an oversight. That’s the design.

Why I’m not running a signals service

A signals service requires the seller to be right more often than wrong, in real time, on someone else’s clock. Nobody sustains that, and the people who claim to are either lying about the win rate or lucky inside a regime that hasn’t broken yet.

The signals seller is also in the wrong job structurally. Their incentive is to keep selling signals even when the regime changes underneath them, because subscribers pay for signals and silence doesn’t pay. My incentive is to publish the framework even when it’s losing, because the framework being honest is the product. If the framework breaks, you’ll read about it here first. That post will not be fun to write. I’ll write it anyway.

Personalized advice has a different problem. It would require knowing your account, your risk tolerance, your time horizon, your tax situation, your psychology under drawdown, and probably more. You can’t deliver that in a 200-word email to a stranger. Anyone who does is wrong by construction.

Hype is the opposite of methodology. Methodology shows the losses. Hype hides them. You can run one or the other. You can’t run both.

Helio shows the framework. The framework gets it wrong sometimes. When it does, the post will say so.

Subscribe now

What you’ll actually receive

Every Friday, free: a weekly review. What the framework caught, what it missed, what was traded, what was passed on. Always free. No paywall. No withholding the good stuff for paid readers.

Within 48 hours of any trade close, paid: a debrief. Anonymized, methodology-focused. The signal that triggered entry. The rule that governed exit. The gap, if any, between what the rule said and what actually happened.

Last Friday of each month, paid: an edge validation report. Numbers, not narrative. Whether the framework actually worked that month, by the math.

Quarterly, paid: a system architecture deep-dive. How the rules interact. Why the stops are where they are. What’s been changed in the framework, and why.

Paid subscribers also get the full archive. Every debrief, every validation report, every architecture piece, from launch onward. The work compounds.

The free tier is always free. The paid tier opens May 10. Until then, everything is open.

The deal

The free Friday review is everything you need to evaluate whether the framework is worth more attention. A few weeks of those, and you’ll know.

If the free tier doesn’t earn the paid tier, the paid tier doesn’t deserve your money. That’s the deal.

The free tier is the proof. The paid tier is the depth.

Helio