Capstone: Build Your Own AmiBroker Trading Research Workstation

Everything in this course has been building one thing, and this is it: a research environment you built yourself, whose every component you can explain, ending in a written report about a strategy you designed and tested.

The deliverable is not a profitable system. It is a piece of research you would be willing to defend — including, if that is what the evidence says, the conclusion that the idea does not work.

What the nine components add up to

1. 1 Market dashboardWhat kind of market is this? Benchmark trend, breadth, volatility, sector strength.
2. 2 Stock explorationWhat is in the universe right now? One row per symbol with every measurement the rest depends on.
3. 3 ScannerWhich symbols met an explicitly documented setup on the last completed bar?
4. 4 Multi-timeframe chartWhat does one candidate look like across three horizons — provably without look-ahead?
5. 5 Real-time or replay workspaceWatching a decision point arrive. Alert, human review, decision. No automation.
6. 6 The trading strategyHypothesis, universe, entry, exit, stop, sizing, constraints, costs — all written down before any test.
7. 7 and 8 Backtest and robustnessPortfolio backtest, execution audit, parameter sensitivity, out-of-sample, walk-forward.
8. 9 The research reportThirteen questions you must be able to answer about your own work.

Each component is usable on its own. Together they are a workflow: from market state, to candidates, to a specific setup, to a decision, to a tested rule set, to a written conclusion.

What this is not

It is not a template to fill in, and it is not a system to trade. Several of the formulas here are deliberately plain, because the subject is the workstation and the method — not a set of rules somebody handed you.

⚠Warning

The capstone ends at a decision, not at an order. Component 5 produces an alert; a human reads it; a human decides. Nothing in this course automates live trading, and nothing in the capstone changes that.

AmiBroker does have an automated-trading interface, and it is a legitimate tool. It is out of scope here because automating a strategy you have not validated is the fastest available way to lose money, and because validation is what the previous thirty-six parts were about.

The Level A path is complete

Every component has a version that needs no paid market-data subscription. Component 5 is the only one where the distinction is visible, and it has two formulas: a live-feed version and a Bar Replay version whose decision logic is identical line for line.

✱AmiBroker Professional required

Real-time data requires the AmiBroker Professional edition, and the GetRTData() reference states that calling it on the Standard edition gives Null values for every field.

If you are on Standard or have no feed, use the replay path. It is not a reduced version of the capstone — it is the same exercise with a different source for the next bar, and it is the path this course tests against.

How long it takes

The curriculum budgets a little over seven hours of work across the nine components. That is time on task with a database already in place, and it does not include the thinking. Realistically:

• A weekend, if you already have a database, a universe and a benchmark symbol, and you accept the specification in Component 6 as written.
• Two or three weeks of evenings, if you design your own hypothesis — which is what the exercise is actually for.

The single largest time sink is the universe, and it is worth knowing that in advance. Building something close to a point-in-time list is slow, and Component 9 asks you about it directly.

What you need before you start

• A daily end-of-day database with several years of history for twenty to fifty liquid instruments.
• A benchmark symbol in that database, sharing the same trading calendar.
• The breadth composites from Part 16, built once as a Scan.
• Your own AFL library from Part 11, if you built one.
• A research log. Not optional — Component 9 asks what is in it.

The standard this is held to

Every formula in the capstone opens with a block of assumptions, and every component asks you to record what you actually ran. That is not ceremony. It is the difference between a result and a rumour about your own computer, and it is the habit the whole course exists to build.

PartCapstone of 36Level5 — Real-Time AmiBroker UserPages9Estimated time8.4 hours

0 / 9 lessons in this part completed

• OverviewCapstone Overview and Requirements30 min
• ProjectComponent 1: Market Dashboard60 min
• ProjectComponent 2: Stock Exploration60 min
• ProjectComponent 3: Scanner55 min
• ProjectComponent 4: Multi-Timeframe Chart55 min
• ProjectComponent 5: Real-Time or Replay Workspace55 min
• ProjectComponent 6: The Trading Strategy60 min
• ProjectComponents 7 and 8: Backtest and Robustness Evaluation70 min
• ProjectComponent 9: The Research Report60 min