Candidates that priced a real, liquid contract but did not clear the bars for a recommendation. Each one shows the measured probability, what it needs the stock to do, and what stopped it.
Industries scored on momentum across three timeframes at once, then checked for breadth — whether the whole theme is moving or just one company inside it. Themes that failed are shown with the reason.
Everything the scanner filtered out, and why. This is the most useful tab for judging whether to trust it — a tool that only shows you what it liked can't be audited.
The only question that really matters: of the trades predicted at 70%, how many actually won? If they win 45% of the time, the model is overconfident — and that shows up here as a number instead of a feeling.
This is not a tool that has been shown to make money. It is a tool that measures carefully and keeps reporting that the odds aren't there. What follows explains how it reaches that conclusion.
The one-sentence version
It looks for industries that are genuinely booming, finds the affordable stocks inside them, and then asks a hard question about each one: is there an options trade here that's actually likely to work? — rejecting almost everything.
Why options are harder than stocks
When you buy a stock, being right eventually is usually enough. You can hold. When you buy an option, you're making three predictions at once:
- Direction — the stock goes the way you think
- Size — it moves far enough
- Timing — it does both before the option expires
Get direction right and timing wrong and you can still lose everything you put in. Options also lose a little value every day just from time passing — that's theta, and it works against you the whole time you hold. This is why the tool rejects far more ideas than it accepts.
Step 1 — Find what's actually booming
About 40 industry themes are watched. Each is asked two separate questions:
Is it moving? Measured over roughly a week, a month, and three months at once. A theme that jumped 10% yesterday and did nothing for three months isn't booming, it's twitching.
Is the whole theme moving, or one company? Called breadth. If an industry ETF is up but only one company is actually rising, that's a single-company story wearing an industry costume. A majority of members must participate.
A theme going down with confirmation is just as useful as one going up — you can buy puts, which profit when a stock falls.
Step 2 — Find the names you can actually trade
A high-probability contract on a $20 stock can cost $250–600. On a $6 stock, a comparable contract might cost $40–90. Same quality of trade, wildly different price, because an option's cost scales with the share price. The tool also requires enough volume that the options have real buyers and sellers — including at least one strike liquid enough that you could sell it back. An option you can't exit is a trap.
The price range was widened from $1–$12 to $1–$60 in August 2026. The narrow band existed to fit a small per-trade budget, but it had an unintended effect: very cheap stocks have thin option chains, so the deep, safe strikes this tool wants often have no buyer at all. Over 55 scans that produced almost nothing. Widening the band brought in names whose options actually trade. The per-trade spending cap — currently $500 — now does the real filtering instead.
Step 2b — Decide it's the right moment
Finding a good company in a booming industry isn't enough; an option has a deadline, so when you buy matters as much as what.
The main rule looks for a stock that has stretched unusually far from its recent average — about a month's worth of trading — and bets on it snapping back. Stretched unusually low inside a rising theme is the classic setup.
Two things changed here in August 2026, both worth knowing.
First, the rule used to also demand that the stretch agree with the industry's direction, and would throw a candidate away otherwise. That turned out to discard 119 candidates that had otherwise qualified. The reason was structural: this tool deliberately hunts booming industries, so the stocks inside them mostly stretch upward — exactly the case the rule was throwing out. Industry direction still counts, but now as one piece of evidence that can be weighed, rather than an automatic veto.
Second, a candidate with no timing signal can now still qualify — but only on a stricter evidence bar: four independent sources agreeing and none disagreeing, versus three with dissent allowed on the normal path. These are labelled evidence-led. Be aware this variety follows a trend rather than fading one, and that approach has not been validated by the backtest — treat anything it produces with extra caution.
Step 3 — Estimate the chance of success
Each candidate gets a probability built from six separate kinds of evidence:
| Source | What it asks |
|---|---|
| Historical base rate | When this exact setup happened before, how often did it work? |
| The market's own odds | Option pricing contains a built-in probability estimate (delta). |
| News & current events | Is there a real catalyst, which direction — and is it already priced in? |
| Analyst & earnings | Are estimates being revised up or down? Beating or missing? |
| Insider activity | Are executives buying with their own money on the open market? |
| Theme strength | How strong and how broad is the industry trend? |
Independence is the point. Three technical indicators reading the same price chart is one source, not three.
Three must agree. This isn't arbitrary — it comes from measuring what actually happened. When only two sources agreed, results were worse than when just one did. Three or more was where agreement started to genuinely help. And when sources conflict, the probability goes down; disagreement is information, not noise.
Today’s Trades, open positions, and the Watchlist
These two tabs mean very different things, and confusing them is the easiest way to lose money here.
| Today's Trades | Still open | Watchlist | |
|---|---|---|---|
| What it means | Passed every check, in today's scan | Passed every check on an earlier day, and hasn't closed yet | Priced a real contract but failed at least one |
| When it was chosen | The most recent scan | The date printed on the card | The most recent scan |
| Judged by today's rules | Yes | No — by the rules in force on its entry date | Yes |
| Timing signal | Required | Was required, then | Not required |
| Chance of working | 60% or better | 60% or better, as measured then | Any — including 1% |
| Makes money on average | Required | Was required, then | Not required |
| Within the spending cap | Required | Was required, then | Not required |
| Read it as | A recommendation | A trade already underway — check its exit plan, not its entry | A window into what the tool looked at, and why it said no |
"Still open" is history, not a suggestion. Once a trade is recommended it stays on screen until it resolves, because it has money at stake and its result is what eventually tells you whether the tool's probabilities meant anything. But it was chosen under the settings in force on its entry date, and those can change. A symbol sitting in “Still open” can legitimately appear on today’s Rejected tab at the same time — that is the tool being consistent, not contradicting itself. What matters for an open trade is its exit plan: the sell price, the cut price, and the close-by date already printed on the card.
This distinction was added on 1 September 2026 after a real failure. One trade had been recommended in late August under a probability floor that was temporarily lowered and then put back the same day. Because the page showed every open trade under “Today’s Trades”, it kept appearing as a fresh pick for three days — while the scanner, correctly, was rejecting that same symbol every hour for failing the restored floor. The scanner was right and the page was wrong. Splitting the two is the fix.
The Watchlist is not a "second-best" list. It is not ranked for you to pick from — it is everything that got far enough to be measured. A contract with a 51% chance and one with a 1% chance appear on the same tab. Every card states plainly which check it failed.
It exists because "nothing today" is honest but tells you nothing. Showing the near-misses, with their real numbers attached, lets you see what the tool is actually doing. Buying something off the Watchlist means taking a trade this tool explicitly turned down.
Reading a trade card
Every recommendation shows the same set of numbers. Here's what each one means and, more usefully, what it should make you do.
| Field | What it means |
|---|---|
| Cost / contract | What you pay, for one contract covering 100 shares. Quoted at the ask — the price you actually buy at, not the friendlier midpoint. |
| Max loss | The same number as the cost, and that's the point: with a bought option you can lose all of it. There is no stop-loss that saves you and no "wait for it to come back" — it expires. |
| Expected value | Average profit or loss if you made this same trade many times, after the spread, time decay and commission. Must be positive or the trade is rejected, however likely it looks. |
| Breakeven move | How far the stock must move just to get your money back. A +4.2% here means the stock rising 3% still loses you money. This is the number people most often overlook. |
| Delta | Roughly the market's own estimate of the chance this finishes profitable. 0.75 means about a 75% chance. It also tells you how much the option moves per $1 of stock movement — at 0.75, about 75 cents. Higher delta costs more but is likelier to pay. |
| Days to expiry | Your deadline. Being right after this date is worth nothing, which is why contracts are chosen with roughly triple the expected holding time as breathing room. |
| Plan to hold | How long the thesis is expected to take, with a target exit date. The expected value above assumes you exit around here. Holding to expiry instead is a different trade with different odds — the contract is deliberately chosen with about three times this much life left, so there's room to be right slowly, not so you use it all. |
| Implied vol | How much movement the option's price assumes. High implied volatility means an expensive option — you're paying for drama the market expects. |
| Realized vol | How much the stock has actually moved over the last 30 days. The reality check against the number above. |
| IV / RV | The most important number here, and the least obvious. Implied divided by realized. Implied volatility alone tells you nothing: 65% is cheap on a stock that really moves 70%, and expensive on one that only moves 40%. Below 1.0 you're getting the option for less than the stock's actual movement justifies. Above 1.3 you're paying for hype, and the trade is rejected outright. |
| Bid/ask spread | The gap between what buyers offer and sellers want — a cost you pay twice, entering and exiting. On low-priced stocks this can be 10–20%, which is a real headwind before the stock moves at all. Always use a limit order, never a market order. |
| Open interest | How many contracts exist at this strike. Low numbers mean few people to sell back to — an option you can't exit is a trap, however good the thesis was. |
The exit plan, and the mistake that costs the most
Every card carries an exit plan: a price to sell at, a price to cut at, and a date to be out by — whichever comes first. That isn't decoration. With options, when you leave matters about as much as when you arrive.
That is not a corner case here — it's the expected path. This tool deliberately buys contracts around 0.75 delta, meaning roughly three in four are expected to finish in the money. So forgetting to close isn't unlucky; it's the default outcome.
What being exercised actually does:
- A call buys 100 shares at the strike. On a $6 strike that's $600; on a $10 strike, $1,000. For a small account that can be most of the balance, spent on a stock you never meant to own.
- A put sells 100 shares you don't have — which opens a short position. Basic options approval doesn't permit that, so it becomes a forced liquidation rather than a trade.
The fix is trivial: close the position. There isn't even a cost argument for holding on, since buy-to-close orders of $0.65 or less are free at Fidelity. Every card shows what an auto-exercise would cost you, next to the date you should be out by.
Step 4 — Three automatic disqualifiers
- Earnings inside the holding window. Option prices inflate before earnings and collapse right after — IV crush. You can predict the direction correctly and still lose.
- The option is too expensive relative to how much the stock moves. When an industry gets hot, option prices inflate faster than the stock. This is the quiet way a booming sector takes your edge back.
- Probability below a coin flip. A cheap long-shot looks tempting because the payoff is huge. It's still a losing trade if it rarely pays.
Step 5 — Would it make money even if it works?
A trade that wins 85% of the time but gains a little and loses everything the other 15% is still bad. So every survivor must clear an expected value test that accounts for the cost, the bid/ask spread, time decay, and the broker commission. If the math isn't positive, it's rejected even at high probability.
Step 6 — Prove the predictions were real
Every recommendation is recorded with its predicted probability and the real option price at that moment, then tracked against actual prices. That's what the Calibration tab measures. Until enough trades have been graded, every probability shown is an estimate, not a proven rate.
This is why an open trade is never removed from the page. Grading it is the only way a predicted probability ever becomes a measured one, so a trade stays under Still open until it hits its target, its stop, or its close-by date — and then moves into the record behind the Calibration tab.
What this deliberately does NOT do
- It doesn't place trades. A human reviews and enters manually.
- It doesn't trade spreads or sell options. Those need higher broker approval and more capital.
- It doesn't chase lottery tickets. Cheap far-out-of-the-money options offer huge payoffs and almost never pay. Excluded by design.
- It doesn't promise winners. Even a 75% probability loses one time in four.
Honest limitations
- Market data is delayed and can be stale on thinly-traded options. Re-check every recommendation on the broker platform before entering.
- Options can lose 100%. Unlike a stock, there's no "wait for it to come back" — the contract expires.
- The AI reading news can be confidently wrong, and news is often already reflected in the price by the time it's published. That's why it's one bounded source of six and can never qualify a trade by itself.
- Historical base rates come from stock outcomes, and a 62% stock win rate does not translate to a 62% option win rate — time decay and the spread take a cut.
- Past performance doesn't predict future results.
- A small account limits what's possible. The most attractive trades are frequently unaffordable, and the tool says so rather than substituting a worse one.