How working professionals use AI and options to generate weekly income — while they sleep.
You don't need to understand the market. You need to understand the system.
Every week, professionals with no trading background and no interest in staring at charts are generating consistent income from the options market — not by predicting where prices will go, but by getting paid for patience, discipline, and time itself.
Here is the part almost nobody explains clearly: an option has an expiry date, and every single day that passes, a portion of its value evaporates — whether the market moves or not. This is called time decay, and it is the single most reliable force in all of finance. Prices are unpredictable. Time is not. It passes at exactly one second per second, for everyone, always. A strategy built around collecting that decay is a strategy built on the one variable in the market that never surprises you.
Selling a put spread is, in plain terms, agreeing to buy a stock index at a price you consider reasonable, in exchange for being paid today. If the market stays above that price, you keep the payment and the obligation quietly expires. If it falls, your loss is capped from the moment you enter the trade — not discovered afterward, but defined in advance, on paper, before a single dollar is at risk.
So why doesn't everyone do this? Three reasons, and all three are solvable.
First, most people were taught that "options" means "gambling" — leverage, YOLO calls, total loss overnight. That reputation comes almost entirely from buying options, which is a different activity with a different risk profile entirely. Selling defined-risk spreads is closer to underwriting insurance than placing a bet.
Second, the analysis required to do this well — reading volatility, macro conditions, technical structure, and market sentiment simultaneously, every single week — is more than a working professional can realistically do alone, on top of an actual job. This is not a discipline problem. It is a bandwidth problem.
Third, and most simply: nobody showed them a system they could trust enough to follow without becoming a full-time trader themselves.
This book exists to remove that third excuse. The next four parts show you exactly how.
Every trade this system places obeys three rules without exception. They are not suggestions. They are written directly into the code that generates every single recommendation, which means they cannot be skipped in a moment of excitement, and they cannot be forgotten under pressure — because there is no moment where a human has to remember to apply them. The system already has.
Rule One — Defined Risk. Before a trade is ever placed, the maximum possible loss is already known, in dollars, on that exact position. This is the single biggest difference between this approach and the "options = gambling" reputation. A bull put spread has a floor built into its structure — you sell one put and buy another, further out, purely to cap what can go wrong. You know your worst case before you know your best case.
Rule Two — High Probability. Every position is opened far enough from the current market price that, historically, the trade is far more likely to expire worthless — which, in this strategy, is the winning outcome — than to be tested. This is why the system targets a 5-delta strike: in plain terms, a position with roughly a 95% statistical chance of being a non-event by the time it expires.
Rule Three — Time Decay Does the Work. The position is never held out of hope, waiting for a big move in your favor. It is held because every day that passes without disaster is a day that quietly pays you. The plan already assumes boring weeks are the norm — because boring weeks are how this strategy actually makes money.
Together, these three rules mean that on any single trade, you already know your maximum loss, why the odds favor you, and why simply waiting is the strategy — not a failure to act.
The honest bottleneck in options income isn't the strategy — bull put spreads are decades old and well understood. The bottleneck is the weekly analysis: reading five different kinds of market information, every week, without fail, without fatigue, without letting one bad week make the process sloppy. That is not a job for a human alone. It is a job for a system, with a human making the final call.
This system runs five independent lines of analysis every week and blends them into a single score before a human ever sees a recommendation:
Almanac (15% weight) — seasonal and calendar-based market tendencies, the kind of pattern that shows up reliably over decades but that no single person can hold in their head week after week.
Macro (20% weight) — interest rates, inflation data, central bank posture, and the broader economic backdrop the market is trading against that week.
Technical (25% weight) — price structure, volatility levels, and where the market sits relative to its own recent behavior.
LLM Synthesis (20% weight) — multiple AI models independently reading news, sentiment, and unstructured information that doesn't fit neatly into a spreadsheet, then reconciling where they agree and where they don't.
Human Score (20% weight) — the final layer, where judgment earned from decades in technology, finance, and standards work gets the last word. The system proposes. The human decides. Every single week, without exception.
This exact five-agent framework was built and stress-tested first in a university classroom — fifty-six students across five teams, competing to predict weekly index movement using this same structure, months before it managed a single real dollar. By the time it touched a live account, it had already been through more scrutiny than most professional strategies ever get.
Part Three mentioned that Technical analysis carries the single heaviest weight in the system — 25%, more than any other agent. That deserves an explanation, because for most working professionals, "technical analysis" conjures an image of someone staring at a chart covered in colored lines, convinced they can predict tomorrow's price. That image is mostly wrong, and it's worth replacing before going further.
Technical analysis, used properly here, is not fortune-telling. It is pattern recognition applied to a narrow, useful question: given how this market has been behaving recently, is now a sensible time to sell premium, and if so, how far away from the current price should the position sit? That's it. It is not trying to predict Thursday's headline. It is measuring the market's current temperament.
Trend versus range. Markets spend some periods moving persistently in one direction (trending) and other periods drifting sideways within a band (range-bound). A trending market that suddenly reverses is far more dangerous to a short put spread than a market that has been quietly going nowhere. The system checks which environment it's currently in before anything else.
Support and resistance. These are simply price levels where the market has repeatedly stalled, reversed, or struggled to pass in the recent past — the market's own memory of where buyers and sellers have fought before. A strike placed just below a well-established support level is safer than one placed in open air with no prior floor beneath it.
Volatility (the VIX). This is arguably the single most important number in the entire system, and it isn't a price at all — it's a measure of how much movement the market currently expects from itself. Higher volatility means richer premium for sellers, but it also means wider, less predictable swings. The system has a hard ceiling here: above a VIX of 25, no new trade is proposed that week, full stop. High fear is exactly when the temptation to chase larger premium is strongest — and exactly when the rules matter most.
Momentum indicators (like RSI). The Relative Strength Index, in plain terms, measures whether a market has moved unusually far, unusually fast, in one direction — a rough proxy for "stretched" versus "calm." A market that looks stretched is more prone to a sharp snap-back than one moving at an ordinary pace, which matters when deciding how much room a spread needs to breathe.
None of these signals, alone, decides anything. That is precisely why Technical is one of five agents rather than the whole system. It answers "is the current environment favorable, and where should this position sit" — not "will the market go up or down next week." Nobody reliably answers that second question, and this book will not pretend otherwise.
Of the five agents, Almanac carries the lightest weight — 15%, the smallest voice in the room. That's deliberate, and understanding why it's weighted so lightly is just as important as understanding what it measures.
The Almanac agent looks at something that has nothing to do with news, earnings, or economic data: the calendar itself. Markets, it turns out, have mild but measurable habits tied to the time of year, the day of the week, and even the position within the options expiration cycle — patterns compiled over decades and re-confirmed, imperfectly, year after year.
The "Santa Claus Rally." The last five trading days of December and the first two of January have, historically, produced positive returns far more often than an ordinary seven-day stretch chosen at random. Nobody has a fully satisfying explanation — thin holiday volume, portfolio window-dressing, year-end fund flows are all theories — but the pattern itself has held up across a remarkably long stretch of market history.
"Sell in May." The old market saying that the six months from November through April have historically outperformed the six months from May through October. It's real enough to have earned its own nickname, and weak enough that plenty of individual years break the pattern entirely.
Options expiration week. The week containing the monthly options expiration tends to see distinct volatility behavior — sometimes compressed, sometimes exaggerated — as large positions get closed, rolled, or exercised. For a strategy built entirely around options, this one matters more directly than the others.
Turn-of-month flows. The final and first few trading days of each month see disproportionate buying, largely driven by mechanical retirement contributions and fund rebalancing that happens on a schedule, regardless of what the news says that day.
Here is the honest caveat, and it's the reason this agent gets only 15% rather than a louder vote: every one of these patterns is a statistical tendency observed across many years, not a rule that holds every year. Treat the Almanac agent as background music, not a lead instrument — a mild tailwind or headwind that shifts the odds a little, layered underneath the macro, technical, and sentiment analysis that carry far more weight. On the rare week where the calendar and the other four agents actively disagree, the calendar loses every time. That hierarchy is exactly what keeps this signal useful instead of superstitious.
If Technical analysis reads the market's current mood and Almanac reads the calendar, Macro reads the weather system the whole market is flying through — interest rates, inflation, employment, and the posture of central banks. It carries 20% weight, tied for second-highest alongside LLM Synthesis, because it captures something the other agents structurally cannot: conditions that haven't shown up in price yet, but are about to.
Interest rates and Fed posture. The US Federal Reserve's stance — raising rates, holding, or cutting — sets the gravitational pull on nearly every other asset. A market pricing in rate cuts behaves very differently from one bracing for hikes, even at an identical index level. The system pays particular attention to FOMC meeting weeks, when a single afternoon announcement can move markets more than the previous month combined.
Inflation data (CPI and PPI releases). These scheduled monthly reports are known, dated events — and known events still produce outsized moves when the number surprises expectations. A hotter-than-expected inflation print can trigger a sharp risk-off reaction within minutes of release, regardless of how calm the technical picture looked the day before.
Employment reports. The monthly jobs report functions as a proxy for economic health broadly — strong enough to fuel confidence, weak enough to raise recession concern, and either extreme can move markets independent of anything visible on a price chart.
Why this matters for a weekly options strategy specifically: a defined-risk spread is only as safe as the assumption that this particular week won't see an extreme, low-probability move. Macro conditions are what shift the odds of that extreme move happening in the first place. A technically calm chart sitting directly in front of a major Fed decision is not actually calm — it's quiet in the way a room goes quiet right before something happens.
Macro doesn't try to predict the market's direction any more than Technical does. It answers a narrower, more useful question: is there a known event on the calendar this week capable of overriding everything else the other signals are seeing? When the answer is yes, that alone can be reason enough to sit out entirely, no matter how attractive the premium looks.
Four of the five agents work with structured numbers — prices, dates, economic releases, volatility readings. The fifth works with something far messier: language. Headlines, analyst notes, earnings call transcripts, and the kind of market chatter that never reduces cleanly to a spreadsheet cell. This is LLM Synthesis, and it carries 20% weight — the same as Macro, and for a related reason: it catches things the numbers-only agents structurally cannot see yet.
Multiple large language models — not just one — independently read the same body of text every week and form their own read on market sentiment. Using more than one model is deliberate, not redundant. Any single AI model can misread a headline, latch onto an outlier opinion, or simply get something wrong — the same way any single human analyst can. When several independent models are given the same information and asked the same question, agreement between them is a far stronger signal than any one of them alone, and disagreement is itself useful information: it flags a week where the situation is genuinely ambiguous, not just where one model happened to be confident.
Consider a concrete case: a central bank official makes an unscripted, ambiguous comment during a Sunday evening interview — too late for that week's scheduled economic data, too soft to move a technical chart on its own, but exactly the kind of thing that shifts Monday morning sentiment before the opening bell. Structured, numbers-only agents have nothing to say about this yet. LLM Synthesis can read the actual transcript, weigh the tone against how similar comments have moved markets before, and flag the shift before the price itself has caught up.
This capability comes with a real limitation, which is exactly why it doesn't carry more weight: language models can also overreact to a single loud headline, misjudge tone, or occasionally produce a confident-sounding read that simply isn't grounded in anything real — a known failure mode sometimes called hallucination. That's the case for keeping this signal at 20%, blended with four other independent perspectives, and always subject to the Human Score's final veto. A machine reading the news is a genuinely new capability. It is not, on its own, a reason to remove human judgment from the loop.
Every book in this category eventually has to answer the question it would rather avoid: what happens when the trade doesn't work? This book answers it now, in its own chapter, rather than burying it in a footnote — because a strategy is only as trustworthy as its worst week, not its best one.
No approach to markets wins every single time, and any book that implies otherwise is not being honest with you. What matters is not whether a loss happens — eventually, one will — but whether the loss was bounded, expected, and survivable in advance. That is the entire purpose of Rule One from Part Two: defined risk, decided before the trade exists, not discovered afterward.
The 200% stop loss, in plain terms. Say a spread is sold for a $500 credit. If the cost to close that position later rises to $1,000 — double what was collected — the position is closed automatically, by rule, without debate or hesitation. The loss on that single trade is capped at a known, pre-defined multiple of the credit received. It is not allowed to grow simply because closing it feels like admitting defeat.
The Thursday forced close. Regardless of profit or loss, any position still open by Thursday's close is closed before the weekend. Weekly options can move sharply on thin, low-liquidity Friday trading and gap unpredictably over a weekend with no way to react in real time. The system simply refuses to hold that particular risk, on principle, every single week.
What a losing week actually looks like, operationally: the same alert arrives on your phone as any other week. The dashboard shows the position approaching its stop level rather than its profit target. The system closes it — automatically, or with your one-click confirmation, depending on how the workflow is configured — at the pre-agreed loss, not a cent beyond what was already known when the trade was placed. Nothing about that week required a difficult judgment call in the moment, because the difficult decision was already made in advance, while calm, before there was any money on the line.
This is precisely why Rule Two — high probability, roughly a 95% historical chance of a non-event on any single trade — matters so much in combination with Rule One. A strategy that wins most weeks and caps its rare losses at a known multiple can absorb an occasional bad week and still come out ahead over a full year. A strategy that wins most weeks but allows one loss to spiral unchecked cannot. The goal was never zero losses. The goal is losses that stay exactly the size you decided, in advance, that you could afford.
Before trying any of this, most people have the same handful of worries — reasonable ones, usually inherited from a bad experience with a different kind of options trading entirely. They deserve straight answers.
"Will I get assigned shares I don't want?" No — and this is one of the most reassuring, least-discussed facts about trading SPX index options specifically. SPX options are European-style and cash-settled, meaning they can only be exercised at expiration, never early, and settlement is in cash rather than shares. The classic nightmare of single-stock options — waking up unexpectedly owning, or short, one hundred shares you never intended to hold — structurally cannot happen here.
"How much capital does this actually tie up?" Far less than people assume. Margin on a defined-risk spread is calculated on the width of the spread minus the credit already received — not on the full notional value of the index. This is one of the specific advantages of trading a spread rather than a single naked option: the capital efficiency is built into the structure itself.
"What if I'm asleep, traveling, or simply forget to click approve?" Nothing happens, on purpose. The system is built to default to inaction, not action. No approval means no trade that week. There is no version of this workflow where a position gets opened silently on your behalf without a deliberate, reviewed decision.
"What about taxes?" This varies meaningfully by country and by personal circumstances, and this book is not the place to get specific tax guidance — treat any number in this book as pre-tax, and speak to a qualified accountant familiar with derivatives income in your own jurisdiction before treating this as a net income figure.
"Isn't this still just a form of gambling?" Gambling, properly defined, is accepting negative expected value in exchange for the thrill of a payout. This strategy is closer to underwriting: getting paid today for accepting a defined, bounded risk that, on average, resolves in your favor far more often than against you — the same basic logic an insurance company runs on, applied to a different kind of risk.
Every Monday, at 9:35am New York time, the system wakes up and runs its full analysis before you've likely finished breakfast — or, from Singapore, well after most people are asleep.
Step 1 — The Recommendation. The five-agent pipeline runs, produces a composite score, and — if conditions clear the minimum bar for a trade — proposes a specific SPX bull put spread: the strikes, the width, the target profit, and the point at which the position would be closed for a loss rather than allowed to run.
Step 2 — The Alert. A message lands on your phone the moment the recommendation is ready. No dashboard-checking required to know something needs your attention.
Step 3 — The Review. You open the dashboard, see exactly why the system is recommending this trade — which of the five signals are driving it, and which are cautious — and you decide, with full visibility, not blind trust.
Step 4 — The Approval. One click. The order goes to the broker exactly as reviewed. Nothing hidden, nothing automatic without your sign-off.
This book opened with a single claim: you don't need to understand the market, you need to understand the system. Everything between that first page and this one has been an attempt to earn that claim rather than simply assert it — walking through each of the five signals honestly, including their limits, and being equally honest about what happens on the weeks it doesn't go your way.
There is a quieter thread running underneath all of it, worth naming directly: this system did not begin as a personal trading tool. It began in a classroom, built and stress-tested first by fifty-six students who had never placed a real trade, competing across five teams to see whose analysis could hold up against a real market. By the time it touched a live account, it had already survived more scrutiny, from more independent perspectives, than most professional strategies ever receive. The students who built the intelligence layer are, in a very real sense, this book's first co-authors.
Parts Ten and Eleven of this book are still being written — deliberately, in public, alongside the live track record itself. Every trade that closes from here forward is a page. The dashboard is open to anyone who wants to watch it happen in real time, not just read about it after the fact. That transparency is the actual point: not a promise that this always works, but a record honest enough that you can judge for yourself whether it does.