AI Trading: What Works and What Does Not

Artificial intelligence in trading is surrounded by more claims than evidence. Professional trading firms do use machine learning, and some use it extensively. Most products marketed to individuals as AI trading systems, though, rest on backtests that would not survive a careful look. This free course explains the difference.

Can artificial intelligence really trade the markets profitably?

In some uses, yes. Machine learning is applied to execution, to forecasting quantities such as volume and volatility, to extracting information from text, and to combining many weak signals into one. Those uses share a feature: large amounts of data and a well-defined target. Predicting the direction of prices days ahead with a single model has neither, and results there are much weaker.

Why do so many AI trading strategies fail?

Financial data is noisy, the number of truly independent observations is small, and markets change. A flexible model can fit the past almost perfectly and learn nothing useful. Add look-ahead errors and ignored trading costs, and a backtest can show large profits from a strategy that will lose money. The course shows each failure with examples.

Where does this course lead?

It is the free starting point of the machine learning track. Machine Learning for Trading is the next step, followed by the researcher-level Financial Machine Learning, which covers deep learning and reinforcement learning.

What you will be able to do

• Distinguish rules-based algorithmic systems from genuine machine learning architectures and apply five diagnostic questions to any AI trading vendor claim.
• Explain why the strongest quantitative performance records accompany the strictest secrecy and what that implies about retail products that describe their methods openly.
• Identify which classes of alternative data carry documented predictive value and describe the adoption lifecycle that compresses any working signal as it becomes widely known.
• Recognize survivorship bias, look-ahead bias, non-stationarity, and LLM training contamination as distinct failure modes that each silently inflate backtested performance.
• Evaluate a strategy's data provenance, parameter stability, and Sharpe ratio integrity using the 30-trades-per-parameter rule, multiple-testing corrections, and walk-forward validation.
• Ask the five due diligence questions that separate independently verified performance from marketing claims across robo-advisors, screeners, signal services, and social trading platforms.
• Apply the alpha-translation chain and DOME checklist to locate exactly where any AI trading claim runs out of evidence between raw signal and net return.

Syllabus, module by module

1. What AI Actually Means in Trading

   How AI, machine learning, deep learning, and algorithmic trading are precisely defined, commonly misrepresented, and distinguished from one another.

2. A Short History of Machines in Markets

   How systematic and machine-learning-driven trading evolved from rule-based origins in 1948 through the major quantitative firms that dominate markets today.

3. Where Machine Learning Genuinely Works

   Where machine learning creates genuine, documented value in trading, including execution optimization, alternative data signals, and arbitrage-free model constraints.

4. The Peculiar Traps of Financial Data

   How non-stationarity, survivorship bias, look-ahead bias, and LLM training contamination each corrupt financial datasets and inflate backtest performance.

5. Why Most AI Trading Claims Fail on Inspection

   Why overfitting, multiple testing, and Sharpe ratio inflation cause the overwhelming majority of AI trading strategies to fail when deployed with real capital.

6. The Retail AI Trading Product Landscape

   What the major retail AI trading product categories are, how each functions mechanically, and what independent evidence shows about their real-world performance.

7. How to Judge Any Strategy That Says It Uses AI

   How data quality, provider independence, in-sample and out-of-sample structure, and walk-forward testing determine whether a backtested result is trustworthy.

8. Building a Sober, Durable Perspective on AI in Markets

   How to apply a calibrated, evidence-based framework to any AI trading claim, grounded in documented verified strengths and extensively catalogued failure modes.

What every course contains

Every paid course comes with the first seven and the certificate, whichever level it sits at. The proctored exam and the credential come when the course is taken as part of a Certificate, Diploma or Fellowship. Everything is self-paced: the material is yours to study in any order, at any hour, for the whole of the license.

1. Workbooks. Written to train the thinking patterns and the important concepts of the course, with worked problems and the reasoning behind each answer, so that you learn to think the way the subject is thought.
2. Audio edition. Focused, well-presented recordings with explainers and examples of the material, made to be interesting, relevant and entertaining, so that the general subject can be learned away from the desk.
3. Notes. Pre-produced notes with every fact the course relies on, written out in full, so that nothing has to be transcribed from the films or the audio.
4. Cheat sheet. One sheet with every important learning concept of the course on it, to keep beside you during the work and to revise from before the exam.
5. Named sources. Every fact in the notes names its source, the paper, the book or the dataset, for anyone who wants to dig in further.
6. Explainer films. Focused films, in the style of the sample films on the catalog page, that walk through the most difficult concepts of the course with visual explanations of a high standard.
7. Transcript. A full transcript of every film and every audio recording, searchable and quotable.
8. Proctored exam. Taken as part of a Certificate, Diploma or Fellowship, the course ends in a proctored multiple-choice questionnaire, with a pass mark of 60% and three retakes included. A course taken on its own is assessed by its coursework and has no exam.
9. Verifiable certificate. A certificate of completion that a third party can verify: an employer sends the code on it to the credentials office and receives written confirmation of the course, the date and the result.
10. Credential of the program. The Certificate, Diploma or Fellowship of the program, track or path the course belongs to, awarded when every course in it is passed. Which one depends on the program you choose.

Assessment

Not assessed. Study notes, a cheat sheet and a workbook come with every lesson. Open courses are free, have no exam and award no certificate. Read how certification works.