How the Tradytics AI Fund picks its trades
By Faizan Ahmad. Updated .
Every trade in the Tradytics AI Fund goes the same way. The AI research desk writes an idea with a variant view (the market's number, the desk's number, and why) and a dated catalyst that will show who is right. The AI portfolio manager then compares the move it expects with the move options already price, picks a structure that wins if the view is right without selling a naked option, sizes it by how much of the fund it can lose, and writes down when it will take profit, cut, or give up on the catalyst. An independent reviewer can challenge it first, and after the catalyst a nightly review grades what happened. Every step is published at /fund.
What are the steps from idea to position?
| Step | Who | What gets decided |
|---|---|---|
| 1. Variant view | Research desk | What the market expects, what the desk expects instead, and why the price misses it |
| 2. Dated catalyst | Research desk | The event that reveals it, its date (confirmed or estimated) and what exactly it will show |
| 3. Moves vs the options | Portfolio manager | The bull, base and bear moves against the move options price to the catalyst date |
| 4. Structure | Portfolio manager | Calls, puts, debit or credit spreads, calendars, butterflies or condors, chosen against implied volatility, skew and term structure |
| 5. Fill check | Portfolio manager | Each leg's bid and ask; wide markets get nearer strikes, a simpler structure or a pass |
| 6. Size | Portfolio manager | The share of the fund it can lose, and the book's exposure before and after |
| 7. Review | Independent reviewer | Proceed, modify or pass, with evidence; the manager answers on the record |
| 8. Exit plan | Portfolio manager | What it takes profit on, what it cuts on, and the date the catalyst must have happened by |
| 9. Grade | Nightly review | Whether the catalyst worked, failed, was mixed or didn't happen, and the lesson |
What is a variant view?
A variant view is a specific disagreement with the market. Not "this is a good company" but "consensus expects this number, we expect that one, and here is why the price doesn't reflect it yet." The desk writes it in at most 80 words, and every idea answers six questions with numbers:
- What precise outcome does the desk expect, with units and a range?
- What does the market expect, as of what date, and how do we know (consensus, guidance, price behavior)?
- Why is the desk right, and why isn't it already in the price?
- What event reveals it, when, and what exactly will we learn?
- How does the stock move in the bull, base and bear case, and how does that compare with the move options already price?
- What proves the idea wrong, and what is the strongest case against it?
Consensus numbers are checked against two sources, because a single data provider's estimates are often stale, especially for mid-caps. The original filing sections are read, not search snippets, and every source is kept with its URL and date.
What counts as a dated catalyst?
The specific information that will reveal the view, on a known or estimated date. An earnings date by itself isn't a catalyst; the number it reveals (a margin, a guidance range, a subscriber count) can be. The desk draws from the whole range of events:
| Kind | Examples |
|---|---|
| Macro and policy | Rates, inflation and jobs prints, central banks, tariffs and trade, government spending, regulation |
| Industry | Pricing cycles, supply and capacity, commodity inputs, channel and inventory data, a peer's results that change the read for the group |
| Company | Product launches and ramps, investor days, approvals and rulings, large contracts, M&A, spin-offs, buybacks, management changes, index changes |
| Secular | A business mix shift or a long trend the market underrates |
No more than about a third of the book should hinge on an earnings print, and the desk spreads catalyst dates so one day or one macro release doesn't decide the whole book.
How does the fund choose the option structure?
By comparing its own expected move with the move options price to the catalyst date (the expected move). The gap between the two decides whether the fund should own options or sell them.
| If the desk expects | The edge is | Structures it leans to |
|---|---|---|
| A bigger move than options price | Owning optionality | Long calls or puts, bought straddles or strangles |
| A smaller move than options price | Selling rich premium, with defined risk | Credit spreads, iron condors and iron butterflies |
| The right direction, with the move fairly priced | Direction, not volatility | Debit spreads, calendars, longer expiries, so a drop in implied volatility after the event can't erase a correct call |
| A precise target | Where the stock lands | Butterflies |
It also reads skew and the term structure before choosing strikes and expiries, and it models profit and loss at the date it expects to exit, including delays and an IV crush, not only the payoff at expiry. Two rules hold for every structure: it never sells a naked option (every short call or put is covered by a long option of the same type, or a short call by long stock), and no one structure makes up more than about a third of positions.
How is each trade sized?
From the risk, not the premium. The fund decides how much of its value it is willing to lose if the position goes to its max loss or its planned exit, then sets the quantity to match.
| Position | Risk at max loss or planned exit |
|---|---|
| Starter | About 0.5% of the fund |
| Standard | About 1% of the fund |
| High conviction, clean catalyst, good option pricing | Up to 2 to 3% of the fund |
Before opening, the manager checks the whole book before and after the trade: net and gross delta, sector, industry and factor exposure, theta, vega, and how the book does on a fast 5 to 10% market move. A trade that pushes the book somewhere the manager wouldn't choose gets resized, paired with a hedge in liquid ETF options, or dropped. It also checks each leg's market: if crossing the spread would cost more than about a tenth of the position's value, it picks nearer strikes or a simpler structure, or passes.
What does the independent reviewer do?
It forms its own view from the original evidence, then recommends proceed, modify or pass. Each finding names its evidence, how it changes the decision, and what would fix it: a fact error, a forecast risk or an implementation problem. The review is advice, never a veto. The portfolio manager owns the decision, may disagree with reasons, and records its response, which appears under "Review and response" on the trade.
How does the fund decide when to exit?
It writes the exit plan when it opens the trade, in business and option terms, then follows it:
- Take profit when the thesis has been priced in.
- Cut when the idea is wrong, not when it is merely early.
- A deadline: the date by which the catalyst must have happened.
- No hiding: rolling a position is a new decision with a real cost, never a way to cover a failed forecast.
Closing uses the opposite legs at the bid and ask. After the catalyst, the nightly review grades it once (worked, failed, mixed or didn't happen) on what happened to the thesis, not on whether the trade made money, and the desk leans on the catalyst types that have worked and asks more of those that haven't.
Where do you see all this on Tradytics?
Open /fund while signed in. Each idea in the feed opens to where the market is wrong, its catalysts and time window, what proves it wrong, revisions, desk notes and sources. Pick a position to see its legs with their fill, bid, ask and midpoint, its cost and max loss as a share of the fund, the thesis, the plan, the review and response, and every fill with its quote time. The fund is open to every signed-in member, Free plan included.
Common questions
Why does the AI Fund trade options instead of stock?
Options let the fund match the structure to the forecast: own a move when options underprice it, sell premium with defined risk when they overprice it, or use spreads when only the direction is the edge. They also cap the loss on each position at a known amount.
Does the AI Fund use stop losses?
Each trade has a written plan: a level or condition to cut at when the idea is proven wrong, and a date by which the catalyst must arrive. Most positions are defined-risk structures, so the max loss is known before the trade opens.
Does the AI Fund give a confidence score for each trade?
No. Its mandate forbids confidence ratings and odds of being right, because they can't be measured. Conviction shows in the evidence, the size of the position, the structure and the conditions that would prove it wrong.
How long does the fund hold a position?
Ideas work over a few weeks to six months. Each trade's plan names the catalyst date and the exit conditions, so the hold depends on when the catalyst lands and how the price reacts.
Can I copy the fund's trades?
Every leg, price and plan is published, but the fund doesn't give advice and its letters never tell readers to trade. If you act on anything, check the live chain yourself; the options strategy builder shows any structure's payoff, breakevens and odds.
Sources
- Cboe, Why Trade Options Around Earnings Season (October 19, 2026): the market prices in an expected move around earnings; straddles need a move bigger than their cost.
- FINRA: Options: FINRA's investor guide to calls, puts, premium and the risks of options.