Skip to content
Summer 2026 · 26% off every plan with SUMMER26 See pricing
Optionomics
What we publish

Trade ideas with the receipts attached.

Trade ideas, signals and short daily write-ups, all built off the same scored trades. Every idea is timestamped when it goes out and marked afterwards against what the stock actually did. The record is public, losses and all.

Included on Theta and Vega

Market note · 14:12 ET

Since 13:40, traders have been closing out their bets against the big index while still buying individual stocks. That split has resolved upward five of the seven times it has shown up this quarter.

Kind
Change of direction
How much it matters
Worth a look
Marked against
Tomorrow's close
2,062
Resolved ideas counted
77.5%
Aggregate lifetime win rate
5% / 30d
What counts as a win
30
Models tracked, every one published

Everything we run is in here, including the approaches that have not worked. See the full record · Educational data. Past performance does not guarantee future results — what worked before may not work next time.

Ideas
Trade ideas

An idea that is allowed to say no.

A set of scanners look for different things — urgent buying, very large trades, stocks running hot, stocks that have overshot, moves around earnings. Each one proposes a trade, and then has to survive the checks.

The call is the product.

Every idea comes with one of three answers — worth taking, wait, or skip — and the reasoning behind it. Some checks are deal-breakers rather than evidence: a price level already broken, too little time left, an earnings announcement landing right in the middle of it. Those end an idea outright instead of being averaged against the things that still look fine.

The rest get scored, and a check that comes back undecided lowers our confidence rather than being ignored. If too few checks have anything to say at all, the answer is wait. Three numbers nodding along is not a signal.

See how they turned out
Deal-breakers, then checks, then one number
Gates — any one of these ends it Price levels Time remaining Earnings window Checks that produced a reading Positioning 82 Volatility 71 Flow 64 Levels 48 Regime 44 Catalyst 22 Prior cases no reading and the rest of them Conviction 69 / 100
A check that says nothing lowers our confidence. It never raises it.
Where to get in, out and stop
Every idea names its prices and how long it is meant to play out over, so there is nothing left to interpret afterwards.
Tracked from the moment it publishes
We follow how it does from the second it goes out. Nothing is added to the record after the fact, and nothing is quietly removed from it.
A link you can share
Every finished idea has a public page showing what we said, how long it was judged over, and how far it actually got.
Tested before it ships
A new approach only goes live after it has been tested against real history using only what was knowable at the time, at prices you could have got.
Few
Signals

Almost all of it is nothing.

We look for shapes across the whole market — clusters of urgent buying, sudden reversals, one big trade confirming another — score each one into a single number, and then filter it again before it reaches you.

The work is throwing things away.

The hard part is not ranking things, it is refusing them. On an ordinary day the number of things that get through is small enough to read in a couple of minutes — which is the entire point, because a list of two hundred "signals" is just the raw market again with a nicer font.

What does survive arrives with a direction, a confidence score and the pattern that produced it, so you can disagree with it about something specific rather than in general.

What survives each stage
↓ below the floor ↓ no baseline behind it ↓ a gate said no Every print and every filing Scored vs. its own history Clears its floor tier, then baseline Published idea · signal · alert
The widths are illustrative. The shape is not — nearly everything we read gets discarded, and that discarding is the part you are paying for.
Models
Machine-learning forecasts

What the models expect next.

A separate set of models, trained on years of price and options history, that put a number on what the next few sessions might look like — and say how sure they are about it.

A forecast is a range, not a number Confidence attached
What happened What the models expect Possible Most likely Today
The cone widens on purpose. Anything drawn as a single confident line several sessions out is telling you something the model does not know.
Direction, with a size
Where the models think a stock goes over the coming sessions, as a percentage rather than a word, and the probability they put behind it.
How jumpy it is about to get
How much movement the models expect, next to what the options market is currently charging for movement. The two disagreeing is itself worth knowing.
Their own price zones
The models draw support and resistance zones of their own onto the chart, beside the ones worked out from where the options money is sitting.
When a day is genuinely odd
An anomaly reading that says how unusual today is for that stock — with a statistical measure of how surprised the model actually is, rather than a label somebody chose.
Confidence, always attached
No reading arrives bare. When a model is unsure, the number it hands back says so, and the screen shows that rather than rounding it away.

These models are new, and we would rather you treated them that way. Every reading is shown with its confidence and never on its own — a forecast is one voice beside the trades, the filings and the news, not a verdict. Past performance does not guarantee future results.

Notes
Daily write-ups

Catch up in thirty seconds.

Short notes written through the day about the things that actually changed — where the money turned, a very large trade, a shift in mood, a jump in nervousness, money moving between sectors.

Only when there is something to say
Each note is marked as routine, worth a look, or important — so a quiet day produces a short page instead of the same word count as a violent one.
Marked afterwards
Every note is checked against what the price actually did next. A writer that is consistently wrong is something we can see rather than something we hope about.
Earnings, read in advance
Filings and earnings announcements get their own summaries, so the event on your calendar has already been read by the time it arrives.
An assistant you can just ask
Ask about a stock in plain English and it answers from the same data everything else here reads — the contracts, the trades, the filings and the record.

Start on Theta.

Theta adds the ideas, the signals and the daily write-ups on top of everything Gamma already gives you.

Cancel anytime · Month to month · Switch tiers whenever