Probability, Not Prediction — TraderzLife Education Series Book 22

LEVEL 6 · ADVANCED · BOOK 22

Nobody can tell you what the next price will be. This book shows why, using sixty years of testing and nearly twenty-eight years of daily exchange-rate data. It then shows why that is good news.

Prediction, meaning a claim about the next outcome, is not available. The small amount of structure that has been measured can only be used as a distribution, never as a forecast. Probability is not a weaker substitute for prediction. It is where survival and profit actually come from.

• Two charts, one real. Every pattern a trader learns to see also appears in a coin's chart. The book shows why that matters, and what it costs to trade a coin.
• What the tests found. Correlations and variance ratios for the three largest dollar pairs since 1999, and forty years of currency forecasting against the random walk.
• Everything that has survived testing. Trend, carry, long-horizon reversion, market predictors, machine learning and the best trading firms, each with its effect size. The strongest case against the book is put in full, and not one item on the list is a forecast.
• What "70% of the time" really claims. A coin wins 70% of the time with the right bracket. An "80% accurate" signal can be wrong most of the time it fires.
• Fat tails. Four-sigma days happen dozens of times more often than the bell curve allows. The book shows what that does to a drawdown stop, and why the drawdown layers matter even more than the thin-tailed arithmetic says.
• How long before you know. A modest, real edge needs about 2,400 trades to clear the noise. A hundred trades can only reveal a 60% win rate at 1:1.
• Why your backtest lies. The best of 93 rules on pure noise, EUR/USD's trend rules before and after 2013, and the deflated Sharpe ratio worked by hand.
• Checked on 42 markets. Every figure in the book was rebuilt from the same public data. The tests were then repeated on recorded prices from our own systems across 42 markets since 2010: direction close to a coin, move size forecastable in every market, fat tails everywhere, and the best backtest rule fading out of sample.

Every computed figure comes with its data and method, and the appendices give the sample-size and deflated-Sharpe calculations cell by cell for a spreadsheet. The book also checks the rest of the series against its own standard, and publishes the corrections.

12 chapters in four parts, 3 appendices, 8 figures, 170 pages.

An honest note: this book will not give you an edge. It shows you how to find out, at a cost you control, whether you have one. It also shows you how to stay in the game long enough for the answer to arrive.

Instant PDF download.

Educational purposes only — not trading, financial or legal advice. Trading leveraged products carries a high level of risk. Bacchus Analytics Limited, trading as Traderz Life.