FAQs
This is AI-generated analysis. Why should I trust it?
You shouldn't trust it blindly — and we don't ask you to.
Every briefing publishes the full reasoning chain: which disciplines drove the directional bias, which dissented, and what evidence led to the final synthesis. You can see whether the Institutional agent's COT reading makes sense, whether the Technical agent's level identification holds up on your own chart, whether the Options agent's skew interpretation matches what you see in the chain.
The analysis is generated by AI agents. The evaluation is done by you — an experienced trader who knows the difference between sound reasoning and confident-sounding nonsense.
We also publish a rolling track record in every report, and Mon-T's weekly review grades every call from A+ to F with specific reasoning. If the system is producing poor analysis, you'll see it in the data before we'd ever need to tell you.
How is this different from asking ChatGPT to analyse a market?
You could ask ChatGPT to analyse silver this week and get a reasonable response. The difference is what happens across 15 markets, 52 weeks a year.
MAD runs six specialist agents with distinct analytical mandates against each market, every week, using the same structured methodology. Each agent pulls specific data — COT reports, options skew, volatility term structure, economic calendar — rather than generating opinions from general knowledge. The Asset Agent synthesises them with dynamic weighting. Bias integrity rules enforce resets and conviction caps mechanically.
A one-off ChatGPT prompt gives you a single perspective with no methodology consistency, no track record, no accountability, and no way to compare this week's analysis against last week's. MAD gives you a system that runs the same process reliably across 15 markets every Sunday — whether you feel like doing the work or not.
The prompt-ready context block at the end of each briefing is designed for traders who ARE using ChatGPT or Claude in their workflow — it gives your AI assistant structured, verified context to work with instead of generating its own from scratch.
Why don't you provide trade signals or entry/exit points?
Two reasons — one practical, one principled.
The practical reason: providing specific trade signals in the UK crosses from research into regulated investment advice under FCA rules. MAD is structured as research and analysis — directional bias with reasoning — not as a signal service.
The principled reason: entry points, stop losses, and position sizing depend on your account size, risk tolerance, timeframe, and execution methodology. A stop that makes sense for a £500K account is reckless for a £10K account. A swing entry that works for a 2-week hold is wrong for a 2-day hold. Any service providing universal entry points is pretending those differences don't exist.
The desk gives you the analytical foundation. You apply your own execution on top.
Reports publish on Sunday. What if the market moves mid-week?
Sunday analysis provides the macro framework for the week — directional bias, key levels, risk scenarios, and catalysts. That framework doesn't expire on Monday close. The support/resistance zones, institutional positioning data, and volatility regime assessment remain relevant through the week unless something exceptional happens.
Mid-week updates are actively being developed — brief check-ins on what's changed since the Sunday briefing for each market. These will be added to paid tiers as they roll out.
In the meantime, the prompt-ready context block is designed for exactly this use case — paste it into your AI assistant mid-week with current price data and ask it to reassess the Sunday thesis against what's happened since.
What's your track record?
Mon-T's Weekly Review publishes every Friday, grading each call individually with honest retrospective analysis — including the desk's worst calls alongside its best.
We're a young service building our track record in public. Every week adds data. If the numbers don't hold up over time, you'll be the first to see it.
Where does the data come from?
The agents analyse data from multiple sources including Yahoo Finance API for pricing and historical data, CFTC Commitments of Traders reports for institutional positioning, options chain data for volatility and flow analysis, Brave Search API for real-time market context and news, and the economic calendar for upcoming data releases and central bank decisions.
Charts are generated via the Chart-IMG API using TradingView data with MAD's custom visual styling.
Who runs Macro Agent Desk?
Bobby Shaftoe — a London-based trader and technologist who built the system for his own trading before opening it to subscribers. MAD is a solo operation with no outside investors and no sales team. More on the About page.
Who is this NOT for?
MAD is not for beginners. The reports assume familiarity with futures markets, COT reports, volume & volatility terminology, and general macro concepts. Nothing is explained from scratch, this is not a guru or educational service.
MAD is not a signal service. If you want someone to tell you exactly when to enter and exit, look elsewhere.
MAD is not a community. There's no Discord, no live commentary, no group discussion. It's institutional-grade research at accessible pricing delivered by agents to your inbox. There is also a Telegram group where you can talk with Mon-T directly.
MAD is for experienced swing traders & macro investors who already know how to analyse markets and want that analysis done comprehensively across more markets than they can cover alone.