Reverse-Engineering JP Morgan's Logic in 7 Steps: How Macro Before the Chart Built €1.55M

Raffaello Ciotta starts with a threshold he learned from an ex-JP Morgan mentor: trading is two separate decisions, and the chart is the smaller one. Roughly 80% of an institutional position is built off-chart, through macro. Only about 20% is timing. Years spent polishing entries, candles, confluence, liquidity concepts, order blocks and fair value gaps can still leave you inconsistent because by the time price hits your chart, direction has already been decided elsewhere.

He anchors the claim in personal track record: running capital for a hedge fund, partly via CFDs under ESMA-style rules for smaller allocations where it is simply faster and cleaner. The screen shows live MetaTrader master accounts and performance-fee lines of €33k, €25k and €22k; aggregated across several brokers the fees reach €155k. At a 10% cut that implies €1.55M of investor profit. The setup is a master account operating on behalf of private investors — each line one account, one fee — and he only gets paid in the green.

Close the Chart, Open the Research — A Bank-First Start

Step one is to read what banks actually publish. He flips through Crédit Agricole and ING research — political risk, bonds, geopolitics, central-bank meetings, energy, economic data — and there is not a single chart. The point lands harder with how banks train: often 6 to 12 months as an analyst with no chart access and sometimes no Bloomberg terminal, learning to read micro before you are allowed to touch price. The research is public if you register, and checking it yourself is part of the method.

Step two is switching theory. Retail often follows a stochastic logic: project the past forward to find an edge. Ciotta compares it to southern Italian almanacs from 1793 that farmers used — if February 2nd rained in ten of the last ten years, you expect rain this year. There is some truth in seasonality, but no causal mechanism. Deterministic thinking, the way modern weather forecasting works, reads causes — wind, pressure, temperature — that later produce effects. Macro does the same, reading economic data, central-bank narrative, positioning and flows as causes.

That switch is justified by who actually runs markets. Bloomberg's annual hedge-fund leaderboard is dominated by macro, event-driven and quant — not chart technicians. Copying banks sounds intimidating, yet retail has an edge in agility: no mandate delays, no giant book to move, rapid entry and exit via CFDs and futures. The advantage is institutional reasoning with retail execution, not retail reasoning with institutional tools.

The single question every top strategy answers is fair value, defined as the price fundamentals justify. Many retail definitions borrow from auction market theory — value area or point of control derived from price itself — which is circular: using price to explain price and never knowing where price should be. The professional definition computes fair value from fundamentals, plotted as a central tendency; moves more than one standard deviation away become mean-reversion opportunities in a distribution view, not a volume profile.

Two Ways to Compute Fair Value and a 17-Driver Engine

There are two ways to compute it. The first is a quantitative model for longer horizons, inspired by Crédit Agricole's FAST FX (Financial Asset Short-Term FX) model that estimates what any FX cross should be at any point. Ciotta keeps a similar fair-value desk in his terminal for longer views. The second, his core weekly method, is scoring drivers across three layers. For each currency he scores domestic, global and structural drivers; a positive net score implies fair value above price and vice versa, and pairing the strongest with the weakest gives the weekly long and short.

Layer one is domestic / endogenous — the internal health of an economy and the heaviest weight. He opens two of seven drivers. First is PMI, the Purchasing Managers' Index, a monthly survey of large manufacturers that closely leads GDP: above 50 signals expansion. The logic is behavioral: managers answer on expectations; expecting stronger demand they produce more, hire more, employment lifts consumption, consumption lifts credit via auto loans and mortgages, fractional-reserve mechanics expand money supply, and the central bank eventually reacts. One survey carries that entire chain.

The second driver is the interest-rate differential, not the policy rate but the three-month funding cost — compounded secured overnight rates: Sonia for GBP, SOFR for USD, €STR for EUR — the real cost of funding a position for a quarter. His carry desk shows the widest spread on AUD/CHF around 450 basis points, and AUD/CHF has been one of 2026's biggest movers, up roughly 9.30% year-to-date. The principle is capital sits where it is paid most. Borrow where it costs -0.04% in CHF, convert to where it pays 4.46% in AUD, hold short-term paper and collect the spread. The position pays each day it stays open; that flow moves the pair. The job is not to predict the numbers but to score them.

Layer two is global / exogenous — the competitiveness between economies. Its core driver is central-bank narrative: statements, minutes and speeches. His terminal houses a full archive plus a diff tool, but everything is also free on each bank's site, for example the Federal Reserve's statements page. The task is to read the reaction function, not the decision. If the last Fed cut was framed around employment, then employment data will dominate price action until the next meeting. That explains why the same three-star release sometimes explodes and sometimes does nothing — the market tracks what will move the central bank. The second driver in this layer is risk sentiment, measured locally by an FGR gauge and an IMD-style meter akin to fear-and-greed, which helps avoid having the right currency at the wrong time. A chart snippet showing what each currency should do in risk-on versus risk-off regimes captures this filter.

The Structural Layer, the Weighting Engine and the Ceiling

Layer three is structural — where the weekly COT report lives. Published free every week by the U.S. CFTC, it details long and short contracts held by the largest U.S. funds. The classic mistake is to read it as a signal — they are big, so follow them. Ciotta's lesson after months of losses is to read it as a vote: a position is a forward bet. If a speculator buys gold, he expects gold to rise after doing research; absorbing that research is the value, not front-running size. The report shows intent, not timing; he points to a six-video playlist and a free sheet that repurposes the raw CFTC data into readable tables. Even then, COT is just one of 17 drivers.

The engine matters more than the checklist. The three layers are not fixed weights; a weighting engine rebalances them each week depending on what the market is actually responding to. In a week driven by geopolitical repricing, sentiment and positioning carry more. He visualizes fundamentals as a push and geopolitics as a ceiling. In one scenario the ceiling blocks an upward push; when the geopolitical event de-escalates before the fundamental impulse expires, price is free to surge past the prior high — delayed expression, not failed analysis. In another scenario the ceiling stays for months while weekly fundamentals roll over; by the time the ceiling lifts, the original impulse has decayed. That is not fundamentals failing but incomplete research that omitted the ceiling.

That is why the engine also checks whether the market is behaving rationally. The FGR index runs daily as a reality check on whether price respects fundamentals or hits a ceiling. When the gauge is too high, he simply stays out — being right on data while fighting a ceiling still loses. This is his argument against fully automating the ceiling: geopolitical risk does not compress into a clean number. If quant models handled every regime, banks would not pay six- to eight-figure salaries to humans on trading floors to read what models cannot.

The weekly routine is where the bias hits the chart, but only for the last 20%. Over the weekend he scores the three layers, applies the engine and lets the pair emerge as a residual — the strongest versus the weakest. This week that was sterling versus New Zealand dollar, producing GBP/NZD. Inside the scoring desk the rate differential, leverage-fund positioning and a seven-broker retail feed all align; retail is on average -40% short GBP/NZD, which as a contrarian filter is confirmation since most retail loses. The entry is marked on the 2-hour chart for about 180 pips and the placement takes roughly 15 minutes; thereafter one to two hours a day is enough because 80% of the work was done off-chart. A scoring-drift tool tracks each driver week over week, showing GBP/NZD flipping from -12 the prior week to bullish — the momentum of the score itself matters as much as the snapshot.

Step seven is the gap between a good month and a career: the journal as a data collector, not a diary. Not how you felt, but statistics: which weekday bleeds, maximum adverse excursion on winners, average holding period, average win streak. He noticed his Monday win rate was notably lower and fixed the routine around it. From those stats comes a scaling model he calls ARR: beta is the average winning streak, so he scales up on each win until beta, then resets to base size expecting mean reversion. The blurred formula in the video shows that logic applied to a live account that can be refreshed and verified. The goal, he argues, is not to nail a single call but to build a sellable equity curve without deep downside spikes — and banks have entire risk desks for exactly this reason, so fixed 1% per trade is not a professional standard.