Start Here
This page is the map: eight pillars, ~200 articles, plus a 17-chapter advanced stream, in reading order.
This page is the map. Eight pillars, ~200 articles, plus a 17-chapter advanced book stream. Each pillar has a single job and a defined reading order. Read the pillar framing first. Then jump to whatever the table points at.
A word on the order. Pillar 1 is the philosophy stack. Pillars 2 through 6 are the practical engineering stack. Pillar 7 is code. Pillar 8 is the advanced book-native research frontier and assumes most of Pillars 1 through 6. Reading Pillar 8 first is a way to waste a weekend on terminology that has no operational meaning without the earlier groundwork.
The article titles below are the index. Links arrive as the articles get published. An empty link column does not mean the article is missing. It means it is queued.
Pillar 1: The Scientific Trader
Trading as an empirical science, not chart mysticism or guru intuition. Every rule is a hypothesis, every backtest is an experiment, every result has to clear a null benchmark. This pillar exists because most "quant" content is quant-flavoured marketing: formulas, no falsifiability. The reader leaves Pillar 1 asking different questions about every strategy claim made afterwards.
# | Article |
|---|---|
1 | Trading Systems Are Recipes, Not Predictions |
2 | Why Traders Lose Even When Their Ideas Are Good |
3 | The Difference Between a Trading Rule, a Strategy, and a Portfolio |
4 | Why "Quantitative" Does Not Automatically Mean Scientific |
5 | The Trader's Real Job: Control Losses, Not Predict Everything |
6 | Why the Market Does Not Repeat, But Still Rhymes |
7 | Loose Pants Fit Everyone: Why General Trading Ideas Survive Longer |
8 | The Death of the Single-System Trader |
9 | Why Trading Is a Probability Business, Not a Certainty Business |
10 | Why Good Trading Feels Boring |
11 | Technical Analysis as a Scientific Hypothesis |
12 | Backtesting Is an Experiment, Not a Screenshot |
13 | Why One Backtest Tells You Almost Nothing |
14 | The Problem with One Sample of Market History |
15 | Induction in Trading: Why Past Patterns Are Always Uncertain |
16 | The Null Hypothesis for Trading Systems |
17 | Why Benchmarks Matter in Rule Evaluation |
18 | Predictive Power vs Long Bias: The Hidden Trap in Backtests |
19 | How to Make Technical Analysis Falsifiable |
20 | The Difference Between Explanation and Prediction in Markets |
21 | Why Simplicity Is a Statistical Weapon |
22 | The Scientific Method for Building Trading Systems |
Pillar 2: Indicator Engineering
Indicator quality matters more than model sophistication. This pillar treats indicators as engineered information objects: distribution shape, tail behaviour, stationarity, entropy, lag, frequency response. It also takes apart the standard recipe of "throw raw indicators into an ML model and hope" and shows why that pipeline destroys signal at the preprocessing step. Filters get the same treatment: not magic, just operators on the data with measurable lag and frequency response.
# | Article |
|---|---|
23 | The Indicator Is More Important Than the Model |
24 | Garbage Indicators, Garbage Predictions |
25 | Why Most Indicators Should Be Transformed Before Modeling |
26 | Relative Entropy as an Indicator Quality Score |
27 | Range/IQR: A Simple Test for Indicator Tail Problems |
28 | Why Predictive Power Often Lives in the Tails |
29 | How to Test Indicator Thresholds Without Fooling Yourself |
30 | Why You Should Test Long and Short Thresholds Separately |
31 | The Case Against Raw Price Indicators |
32 | How to Build Stationary Indicators from Non-Stationary Prices |
33 | Why ATR Normalization Is More Than a Volatility Trick |
34 | CMMA: A Better Momentum Primitive Than Price-minus-MA Alone |
35 | Why Indicator Histograms Matter |
36 | Taming Indicator Tails with Sigmoid Transforms |
37 | Why the Median Often Beats the Mean in Trading Features |
38 | Feature Engineering Before Machine Learning |
39 | No Filter Is Predictive: What Traders Misunderstand About Smoothing |
40 | The Hidden Cost of Every Moving Average: Lag |
41 | Why the SMA Is Often a Terrible Smoother |
42 | EMA vs SMA: Why Simplicity Still Matters |
43 | The Trader's Guide to Low-Pass Filters |
44 | High-Pass Filters for Traders |
45 | Band-Pass Filters: The Most Underused Tool in Technical Analysis |
46 | Decyclers: Extracting Trend by Removing Cycle Energy |
47 | Why Moving Averages Can Lie at Turning Points |
48 | The Frequency Response of Trading Indicators |
49 | How to Think About Indicator Lag Before Backtesting |
50 | Why Median Filters Are Useful for Volume and Outliers |
51 | Automatic Gain Control for Trading Indicators |
52 | Dominant Cycle Estimation Without Astrology |
53 | Why Market Cycles Are Evanescent |
Pillar 3: Robust Systems Lab
A strategy is not robust because it performed well in a backtest. It is robust because it survived hostile testing. Pillar 3 is the anti-overfitting lab: stationarity, regime coverage, walk-forward, CSCV, Monte Carlo, permutation tests, degrees of freedom, parameter stability, transaction-cost realism. The main enemy is false discovery. Most failed strategies fail because the research process was weak, not because the idea was unsound.
# | Article |
|---|---|
54 | Stationarity: The Word Every Trader Ignores Until It Kills the Strategy |
55 | Slow Wandering: The Most Dangerous Type of Market Change |
56 | Why Systems Work Until They Don't |
57 | How to Detect When a Trading System Is Dying |
58 | Why OOS Failure Is Often a Stationarity Failure |
59 | Volatility Regimes and Strategy Survival |
60 | Why Volatility Is More Non-Stationary Than Trend |
61 | How to Make Indicators More Stationary |
62 | When Forcing Stationarity Destroys Information |
63 | Rolling Normalization: Useful Tool or Hidden Overfit? |
64 | Regime Coverage: Why Your Backtest Needs Different Market States |
65 | The Difference Between Robustness and Optimization |
66 | Why "Works on All Markets" Is Usually a Red Flag |
67 | Market Personality: Why Gold, FX, Crypto, and Equities Need Different Systems |
68 | Optimization Comes After Testing, Not Before |
69 | Degrees of Freedom in Trading Systems |
70 | Why More Parameters Make a Strategy Easier to Sell and Easier to Break |
71 | The 10% Rule of Degrees of Freedom |
72 | How Many Trades Do You Need to Trust a Backtest? |
73 | Why 30 Trades Is Not a Strategy |
74 | Monte Carlo for Trading Systems |
75 | Permutation Tests for Indicator Significance |
76 | CSCV Explained Simply |
77 | Why Walk-Forward Testing Is Better Than One Big OOS Split |
78 | Parameter Stability Beats Best Parameter |
79 | The Hill, the Spike, and the Cliff: Reading Optimization Surfaces |
80 | When a Stop Loss Improves Risk but Destroys Edge |
81 | MAE/MFE Analysis: Seeing What Net Profit Hides |
82 | Why Profit Factor Can Lie |
83 | How to Evaluate a Strategy Beyond Net Profit |
84 | Why Transaction Costs Should Be Added Before You Fall in Love |
85 | The Backtest Integrity Checklist |
Pillar 4: Market Structure Notes
A strategy cannot be understood outside the market structure it trades. This pillar has two layers. The first is market personality: noise vs volatility, efficiency ratio, trend quality, timeframe selection. The second is cross-market structure: bonds, equities, commodities, FX, gold, crude, copper, dollar index, and how one market acts as a filter for another. There is also a hard-edged FX section that explains why retail FX execution is not the same product as wholesale FX, and why broker routing changes the economics of the same signal.
# | Article |
|---|---|
86 | Noise Is Not Volatility |
87 | Efficiency Ratio Explained for Traders |
88 | How to Rank Markets by Trend Quality |
89 | High Noise Markets Are Mean-Reversion Markets |
90 | Low Noise Markets Are Trend-Following Markets |
91 | Why One Indicator Should Not Be Used on Every Market |
92 | How to Choose the Right Timeframe for a Strategy |
93 | Price Density: A Visual Way to Measure Market Choppiness |
94 | The Difference Between Volatility Expansion and Directional Opportunity |
95 | Why Breakout Systems Need Low Noise Environments |
96 | Why Grid Systems Need Noise, Not Just Volatility |
97 | Matching Strategy Families to Market Conditions |
98 | Intermarket Analysis for System Traders |
99 | Why Cross-Asset Signals Beat Isolated Chart Reading |
100 | Using Bonds to Filter Equity Signals |
101 | Gold, Dollar, and Rates: A Practical Intermarket Map |
102 | Copper as an Economic Activity Indicator |
103 | Crude Oil, Inflation, and FX |
104 | Using Ratios as Trading Signals |
105 | Intermarket Divergence as a Trading Filter |
106 | Why FX Traders Must Watch Gold, Rates, and Equities |
107 | Cross-Asset Confirmation for Trend Systems |
108 | Lead-Lag Relationships in Global Markets |
109 | From Intermarket Analysis to Network Momentum |
110 | FX Is Not One Market: Retail vs Wholesale Structure |
111 | The Real Heart of FX Liquidity |
112 | Why Retail FX Execution Is Not the Same as Interbank FX |
113 | How FX PnL Actually Works |
114 | Market Orders vs Limit Orders in FX |
115 | Why Bid/Ask Bounce Matters for Intraday FX Systems |
116 | Why FX Traders Need Macro but Should Trade Systematically |
117 | Using Gold as an FX Indicator |
118 | Cross-Pair Signals: Can EUR Predict GBP? |
119 | Currency Strength Models from Pair Decomposition |
Pillar 5: Microstructure Alpha
At short horizons, edge is prediction conditional on execution. This is the HFT/MFT layer: order books, queues, markouts, adverse selection, fair value, spreads, skew, maker/taker economics. The headline result of this pillar is that a statistically real signal can be economically useless as taker flow and valuable as maker improvement. "Market making collects the spread" is the version sold in blog posts. The real problem is avoiding adverse selection and quoting around a better fair value than the public mid.
# | Article |
|---|---|
120 | Market Making Is Not Just Collecting the Spread |
121 | The Three Pillars of Market Making: Fair Price, Spread, Skew |
122 | Why Fair Value Is the Core of Market Making |
123 | Markouts: The Truth Serum of Market Making |
124 | Why Forecast Accuracy Is Not Enough in Market Making |
125 | Adverse Selection Explained for Traders |
126 | Toxic Flow vs Inventory Risk |
127 | Why Skewing Is Simpler Than People Think |
128 | Spread Widening During Volatility Expansion |
129 | Order Placement Alpha: The Forgotten Edge |
130 | How to Use Order Book Density for Better Limit Orders |
131 | Spoofing, Sturdy Liquidity, and Book Pressure |
132 | Order Book Imbalance: The First Microstructure Feature to Test |
133 | Microprice: Better Than Mid Price? |
134 | Using Trade Flow to Predict Short-Term Price Movement |
135 | Fill Probability from Trade Size CDFs |
136 | TWAP and VWAP Are Execution Models, Not Just Indicators |
137 | Why Small Alphas Matter More for Makers Than Takers |
138 | Maker vs Taker Edge: Same Signal, Different Economics |
139 | Dynamic Symbol Selection for Market Makers |
140 | Cross-Exchange Fair Value for Crypto Perps |
Pillar 6: Portfolio Construction & System Death
A good signal becomes a bad strategy with the wrong sizing. A good strategy becomes a bad portfolio with the wrong correlations. A good backtest becomes dangerous when the drawdown profile is misread. A once-good system decays when the regime changes. Pillar 6 covers three connected problems: portfolio construction, system decay, and the behavioural failures that keep traders attached to dying systems. The complexity/econophysics articles at the end frame why fat tails, anomalous diffusion, and non-Gaussian dynamics matter for sizing decisions.
# | Article |
|---|---|
141 | Ranking Beats Forecasting for Many Trading Problems |
142 | Ranked Long/Short Systems Explained |
143 | Why Volatility-Adjusted Position Sizing Matters |
144 | Cost-Aware Ranking: The Missing Step in Cross-Sectional Strategies |
145 | When an Alpha Metric Is U-Shaped |
146 | Why Z-Scoring Makes Ranking Cleaner |
147 | From Indicator Value to Expected Value |
148 | Why Portfolio Construction Is Part of the Signal |
149 | The Difference Between Signal Quality and Portfolio Quality |
150 | How to Build a Cost-Aware RSI Ranking System |
151 | Drawdown Is Not Just a Number: It Is a Diagnosis |
152 | Average Drawdown vs Extreme Drawdown |
153 | When a Drawdown Means the System Is Broken |
154 | Expectancy: The Most Important Formula in Trading |
155 | Why Percent Profitable Is Overrated |
156 | Profit Factor, Expectancy, and the Shape of Returns |
157 | The Hidden Importance of Time in Market |
158 | Smooth Equity Curves Are Built, Not Found |
159 | When to Switch Off a Trading System |
160 | Why Loss Control Is the Only Thing You Fully Control |
161 | Why Simple Algorithms Beat Smart Humans |
162 | The Flawed Human Brain in Trading |
163 | Get-Even-Itis: The Most Expensive Disease in Trading |
164 | Why Traders Take Profits Too Early and Losses Too Late |
165 | The Illusion of Control in Active Trading |
166 | Why Systematic Trading Feels Emotionally Unsatisfying |
167 | Near Misses and Revenge Trading |
168 | Why Discretionary Traders Need Rules Even If They Hate Systems |
169 | The Discipline Premium in Trading |
170 | How to Remove Yourself from Your Trading System |
171 | Random Walk and Efficient Markets Are Not the Same Thing |
172 | Variance Ratio Tests for Traders |
173 | Fat Tails: Why Gaussian Thinking Breaks Trading Systems |
174 | Levy Distributions and Market Extremes |
175 | Entropy as a Market Concept |
176 | Anomalous Diffusion in Financial Markets |
177 | Why Financial Markets Are Complex Systems |
178 | The Three Stages of a Trading Idea: Absurd, Familiar, Inevitable |
179 | Why Traders Should Analyze Indicators Mathematically |
180 | From Econophysics to Practical Trading Signals |
Pillar 7: Python Research Notebooks
Every serious concept eventually becomes code, charts, and reproducible results. Pillar 7 is the implementation layer of the other pillars. Each notebook ships with a research question, the data setup, the calculation, the visual, the statistical test, the trading interpretation, and a failure-modes section. Notebooks here are not tutorials. They are research artifacts that show whether a claim made elsewhere on the site survives a re-run with different data.
# | Notebook |
|---|---|
181 | Build an Indicator Quality Report in Python |
182 | Build a Threshold Tester for Any Indicator |
183 | Permutation Test for Trading Indicators in Python |
184 | Monte Carlo Drawdown Simulator for Trading Systems |
185 | Optimization Surface Visualizer |
186 | MAE/MFE Analyzer in Python |
187 | Efficiency Ratio Market Screener |
188 | ATR-Normalized Momentum Indicator |
189 | Band-Pass Filter Indicator in Python |
190 | Dominant Cycle Heatmap with DFT |
191 | Order Book Imbalance Backtest |
192 | Microprice vs Mid Price: Empirical Test |
193 | Crypto Cross-Exchange Lead-Lag Study |
194 | Fill Probability Estimator Using Trade Size CDF |
195 | Cost-Aware Ranked Long/Short Strategy |
196 | Volatility-Regime Filter for Any Strategy |
197 | Stationarity Diagnostics for Trading Features |
198 | Backtest Integrity Checklist as Code |
199 | Portfolio of Systems Simulator |
200 | System Decay Detector |
Pillar 8: Physics, Geometry & Event-Driven Markets
The advanced book stream. Markets studied as evolving, event-driven, nonlinear systems rather than as fixed-time price series. Clock time is the wrong lens for most of what matters: events arrive in bursts, volatility clusters, liquidity shocks reshape the correlation structure on intra-day horizons. This pillar introduces intrinsic time, event-driven filters, market geometry, network causality, topological turbulence indicators, optimal transport regime detection, and the physics of phase transitions in many-body markets. Reading this pillar without Pillars 1 through 6 in place is the fastest way to mistake terminology for understanding.
Part I: Foundations
Ch | Chapter |
|---|---|
1 | Intrinsic Time and the Case for Physics in Markets |
2 | Mathematical Toolkit: Stochastic Processes, Hilbert Transform, Phase-Space, Optimal Transport |
Part II: Event-Driven Filters (the 19-family taxonomy)
Ch | Chapter |
|---|---|
3 | Change and Threshold Detection: CUSUM, BOCPD, Directional Change, CDaR |
4 | Path Geometry and Bar Anatomy: Swings, Pivots, Range Estimators, Matrix Profiles |
5 | Volatility, Jumps, Clustering: Lee-Mykland, BNS, GARCH States, Hawkes |
6 | Trend, Memory, Spectrum: Variance Ratio, Hurst, ARFIMA, Wavelets, EMD |
7 | Multi-Series, Regimes, Schedules: HMM, Cointegration, Kalman Pairs, Event Studies |
Part III: Market Geometry
Ch | Chapter |
|---|---|
8 | Correlation Done Right: Marchenko-Pastur, RMT, MST, PMFG |
9 | Network Causality: Granger Networks, Transfer Entropy, Contagion Density |
10 | Riemannian Geometry of Markets: Ollivier-Ricci, Forman-Ricci, Ricci Flow as Regime Speed |
11 | Topological Data Analysis: Persistent Homology, Betti Dynamics, TDA Turbulence Index |
12 | Optimal Transport and Distributional Regimes: Wasserstein, MF-DCCA |
Part IV: Physics of Market Dynamics
Ch | Chapter |
|---|---|
13 | Coupled Oscillators and Phase Synchronization: Kuramoto, Wavelet Coherence |
14 | Information and Thermodynamics: NMI, Permutation Entropy, Market Temperature, Tsallis |
15 | Chaos and Nonlinear Dynamics: Lyapunov, Recurrence Quantification, MF-DFA |
16 | Many-Body Markets: Ising/Spin-Glass, Magnetization, Sornette LPPL Crash Precursors |
Part V: Systematic Trading Synthesis
Ch | Chapter |
|---|---|
17 | Filter Stacking, ML Pipelines, and Production: Triple-Barrier Labels, Purged CV, Deployment |
How to read this site
Three honest paths through the material.
The new-trader path. Read Pillar 1 in order. Then read Pillar 3 in order. Skip Pillars 4 through 8 until both of those feel obvious. Most of the "I lost money on a system that looked great in backtest" content lives in Pillars 1 and 3. The expensive lessons get cheaper here than in a live account.
The signal-engineer path. Skim Pillar 1 in two evenings. Read Pillar 2 in full. Read Pillar 3 in full. Use Pillar 7 to verify any claim that surprised you. Pillar 4 and Pillar 6 are the natural follow-ups once features and validation are not the bottleneck any more.
The advanced-research path. Pillar 8 is the destination, but the prerequisites are Pillar 3 (validation), Pillar 2 (signal engineering), and the complexity articles at the tail of Pillar 6. Skipping the prerequisites turns Pillar 8 into an aesthetic exercise.
KEY POINTS
- Eight pillars. ~200 articles. 17 chapters in the advanced book stream. The numbering is fixed even when the article order on the homepage is not.
- Pillar 1 is philosophy. Pillars 2 through 6 are practical engineering. Pillar 7 is reproducible code. Pillar 8 is the book.
- Read Pillar 1 and Pillar 3 before any other pillar if the goal is to stop losing money on systems that backtested well.
- Read Pillar 2 before Pillar 7 if the goal is to ship signals that survive a permutation test.
- Read Pillar 8 last. The prerequisites are not optional.
- Empty links in the tables mean the article is queued, not missing. Subscribe to get notified as each one ships.