The portfolio is finished before you sit down with a coffee.
Every trading day, before the US open, four proprietary models read every market we cover and build a Systematic Portfolio from what they find.
- It holds ten names or fewer, with an exact weight on each and cash for the rest.
- It aims to beat the S&P 500 with lower volatility and shallower drawdowns.
- It never changes its mind because of how the morning feels.
You read the same page I do, at the same time.
I am Yimin Xu, and I built this in London after years pricing two-way risk on a bank rates desk.
Everything published is listed there, including the free pieces.
The hard part is rarely a shortage of information. It is a process that still holds up in a bad month.
At some point, a position quietly turns into a judgement on you.
Once a view is said out loud, part of you starts defending it. Losing positions get held too long because closing them concedes something. Winners get trimmed early because banking them settles the question.
None of that is a character flaw, and over a career it costs more than any single bad call.
Reading more does not fix it. A process that answers the same way on a bad afternoon as a good one does, and leaves a dated record either way.
Written for people who run capital to a professional standard.
The hard part was never the loss. It was being asked why, and hearing myself reach for something that sounded like reasoning and was a feeling.
I priced two-way risk on a G10 rates desk, where that answer does not survive.
I built the thing I could not buy. You get the machinery itself, the record it has already built, and the rules it follows even on the mornings I would rather it did not.
The research operation a fund would build in-house, run every trading day and published where anyone can check it.
Where I came from, and why I built it →Three steps, and the first costs nothing.
There is no call to book, no discovery form and no enterprise quote. The record and the construction notes are open so this can be judged on the work.
One service at two depths, and a companion portfolio for the parts of the cycle equities find hard.
Start with how deep to go on Macro & Megacaps, then decide whether Commodities belongs beside it.
Macro & Megacaps: Signals
The models, the portfolio, every morning.
Four models read all 73 names before the US open, and build the portfolio from what they agree on.
- Daily signals on 73 names, before the US open
- The portfolio: tickers and target weights, weekly
- Ten years of testing on data the models never saw
Macro & Megacaps: Signals & Research
A bad month is therefore never a surprise.
The portfolio will hold something through a bad stretch. Whether you stay with it depends on whether you saw it coming.
- Everything in Signals, in full
- The macro note, twice a week: the Fed path, the curve, then the indices, the Mag-7, metals and crypto
- Deep dives on the names the portfolio holds: the business, the numbers, the valuation and what would drop it
Commodities: Signals
A companion portfolio for the parts of the cycle equities find hard.
24 names across metals, energy, agriculture and uranium, plus the producers. It runs the same method on a separate universe.
- Daily signals on 24 names, before the US open
- The commodities portfolio, weekly
- Ten years of testing on data the models never saw
- A third off beside either Macro & Megacaps tier
Every subscription includes its portfolio at no extra cost, and your price is locked for as long as you stay. Prices exclude VAT.
How my work helps, in the order you will notice it
Changing nothing is itself a position.
Very little on the left is carelessness. It is what happens when one person does the work of a research team.
- A process that is whatever there was time for.
- Positions exited in week three of a drawdown, for reasons no note records.
- A mandate narrowed to what one person can cover.
- A view to defend, and a portfolio shaped around defending it.
- Nothing to learn from, because nothing was written down.
- One structured read a day, so weeks compare.
- Risk reduced by rule, not by nerve on a bad afternoon.
- A drawdown that reads as the process working.
- A view that belongs to the models, so changing it costs a line.
- A dated record that answers the question before it is asked.
The difference is not intelligence or effort. It is whether the same work happens on the mornings nobody would notice.
Independent means something specific here.
Most market research sits next to trading, banking or advertising revenue, and those incentives shape what gets said. No trading desk, no banking clients, no sponsors, no affiliate links. I sell subscriptions and nothing else.
Everything published is locked the moment it goes out and cannot be edited afterwards. Downgrades publish the week they happen, including the ones that did not work, which is why the record is free to read before you pay anything.
This will not suit everyone, and it is better to say so here.
If any of these is you, something else will serve you better.
Someone who runs capital to a professional standard and wants the same work on the hard mornings as the easy ones. The record is open.
This is what I run. You’re invited to read over my shoulder.
Nothing here tells anyone what to do with their money. This is the operation I run for my own investing, published as I use it, including the weeks it goes badly.
The work is closer to observation than prediction. The models report where the market stands, the record shows what came of it, and the process improves as that evidence builds. None of it asks anyone to be attached to a particular trade, mine included. We are watching the market from the same side of the glass.