Learn
Read one primary source and summarize it in five sentences.
QUANT LEARNING CENTER
Learn the system before trading it.
A cited path from financial foundations to evidence-aware research, risk, and operations.
HOW TO LEARN
Learn · Derive · Apply · Challenge
Derive
Reproduce the central equation or concept by hand or notebook.
Apply
Complete the QT exercise with historical or paper data.
Challenge
State assumptions, failure modes, and disconfirming evidence.
A 12-week route · 5-7 hours per week
| Weeks | Focus | Evidence before advancing |
|---|---|---|
| 1-2 | Instruments + statistics | Explain returns, orders, distributions, and estimation error. |
| 3-4 | Data + time series | Produce a decision-time, leakage-free dataset note. |
| 5-6 | Portfolio + risk | Write and defend a paper-only risk budget. |
| 7-8 | Research + validation | Complete a reproducible, cost-aware report against a benchmark. |
| 9-10 | Crypto structure | Explain funding, basis, liquidation, custody, and venue risk. |
| 11-12 | Operations | Complete reconciliation, incident, and benchmark reviews. |
01 Financial system and instrumentsExplain how cash moves through markets and distinguish ownership, borrowing, and derivative exposure before interpreting a trading signal.
What to understand
- Time value of money, discounting, compounding, and inflation-adjusted returns
- Simple and logarithmic returns; nominal, real, gross, and net performance
- Cash, equities, bonds, commodities, currencies, and crypto assets
- Primary versus secondary markets; exchanges, brokers, clearing, and custody
- Order books, bid-ask spreads, market orders, limit orders, and stop orders
- Spot, futures, perpetual swaps, margin, leverage, funding, and liquidation
Experience notes
- A correct market view can still lose money when spread, fees, funding, or liquidation mechanics are misunderstood.
- Do not trade an instrument until you can draw its cash flows and state who owes what to whom in both normal and stressed markets.
QT LAB
Apply it in this project
Run `qt data sources`. Classify each dataset as price, flow, position, event, or derived indicator, then write what economic object it actually measures.
Sources [cfa-foundations][mit-finance][bis-crypto][qt-operations]
02 Probability and statisticsReason about uncertain outcomes and noisy estimates without treating a sample metric as a durable fact.
What to understand
- Random variables, probability distributions, expectation, variance, and quantiles
- Conditional probability, Bayes' rule, dependence, and independence
- Covariance, correlation, nonlinear dependence, and correlation instability
- Sampling distributions, standard errors, confidence intervals, and power
- Hypothesis tests, p-values, effect sizes, and multiple-comparison risk
- Skewness, kurtosis, fat tails, outliers, and robust summary statistics
Experience notes
- A Sharpe ratio without its sample length and uncertainty is marketing, not a complete statistical statement.
- Trying many indicators and reporting only the winner creates evidence even when no true edge exists; record every trial.
QT LAB
Apply it in this project
Calculate BTC daily-return quantiles and compare observed tail counts with a fitted normal distribution. Explain why the difference matters for stops and sizing.
Sources [mit-probability][multiple-testing][backtest-overfitting]
03 Data and time seriesConstruct a decision-time dataset that is reproducible and does not leak information from the future.
What to understand
- Prices versus returns; levels, differences, and stationarity
- Autocorrelation, volatility clustering, seasonality, and structural breaks
- Sampling frequency, asynchronous markets, resampling, and timezone alignment
- Missing values, stale observations, revisions, and publication lags
- Look-ahead, survivorship, selection, and timestamp leakage
- Data provenance, schemas, validation, lineage, and reproducible snapshots
Experience notes
- Most severe research errors enter before the strategy code: a clean chart can still contain revised or future-known data.
- For every feature, write the earliest timestamp at which a real operator could have observed it.
QT LAB
Apply it in this project
Inspect the dashboard data-source freshness table and one OHLCV file. For a chosen decision time, mark which rows and derived signals were genuinely available.
04 Portfolio and risk architectureTranslate uncertain forecasts into positions that keep losses, liquidity needs, and survival constraints explicit.
What to understand
- Expected return, variance, covariance, diversification, and concentration
- Systematic risk, beta, factor exposure, and benchmark-relative risk
- Volatility, drawdown, recovery time, and path dependence
- Value at Risk, expected shortfall, stress scenarios, and model limitations
- Position sizing, volatility targeting, risk budgets, and exposure caps
- Liquidity, counterparty risk, correlation breakdown, and risk of ruin
Experience notes
- Diversification measured in calm data can disappear during forced selling; include a shared-crash scenario.
- Sizing is a control system, not a confidence score. A compelling narrative never justifies bypassing an exposure cap.
QT LAB
Apply it in this project
Read `qt.risk` and select one paper-only limit. Predict its effect on position size and maximum loss, run the scenario, and reconcile prediction with output.
Sources [markowitz][sharpe-capm][carver-systematic][qt-strategy]
05 Strategy researchTurn an economic hypothesis into a falsifiable, cost-aware rule with an honest benchmark and a reason it may stop working.
What to understand
- Economic mechanism, market participant, constraint, and expected compensation
- Signal definition, decision frequency, holding period, and exit rule
- Null hypothesis, benchmark, effect size, and predeclared success criteria
- Turnover, fees, spread, slippage, funding, borrow, and tax considerations
- Parameter sensitivity, capacity, crowding, competition, and edge decay
- Research logs, versioned hypotheses, negative results, and reproducibility
Experience notes
- Begin with who pays the strategy and why that transfer might persist; an indicator pattern without a mechanism is fragile.
- Simple rules are not automatically valid, but their assumptions and failures are usually easier to inspect than a heavily tuned rule.
QT LAB
Apply it in this project
Before running code, write a one-page hypothesis for DCA, trend, carry, or capitulation: payer, mechanism, rule, costs, benchmark, and invalidation evidence.
Sources [fama-markets][lo-adaptive][carver-systematic][qt-strategy]
06 Backtesting and validationUse historical simulation to reject weak ideas rather than manufacture confidence in the best-looking trial.
What to understand
- Event timing, realistic fills, transaction costs, and portfolio accounting
- Training, validation, test sets, and true out-of-sample evaluation
- Data snooping, selection bias, multiple testing, and backtest overfitting
- Walk-forward analysis, overlapping labels, purging, and embargo concepts
- Bootstrap and Monte Carlo stress, parameter perturbation, and scenario tests
- Benchmark comparison, uncertainty intervals, reproducible artifacts, and audit trails
Experience notes
- Synthetic data proves software behavior, not trading edge. Real-history success is also only evidence—not a guarantee.
- Keep failed variants in the research ledger. Deleting them makes the surviving result look more statistically independent than it is.
QT LAB
Apply it in this project
Run a synthetic backtest as a software check, inspect the artifact, then run walk-forward and Monte Carlo checks on real history and compare with plain DCA.
Sources [backtest-overfitting][multiple-testing][lo-adaptive][qt-strategy]
07 Crypto-specific market structureExplain the venue, custody, collateral, and derivative risks that are absent from a simple spot-price chart.
What to understand
- Self-custody, exchange custody, counterparty exposure, and operational key risk
- Fragmented venues, cross-venue prices, settlement, and transfer latency
- Perpetual funding, dated-futures basis, collateral, and margin conventions
- Liquidation engines, insurance funds, auto-deleveraging, and cascade dynamics
- Stablecoin reserves, redemption, depegs, and quote-currency risk
- On-chain metrics, reflexivity, 24/7 liquidity, and changing market regimes
Experience notes
- A market-neutral position can retain venue, collateral, funding, execution, and liquidation risk; neutral delta does not mean risk-free.
- Do not treat an on-chain metric as ground truth until its entity labels, revisions, and economic interpretation are understood.
QT LAB
Apply it in this project
Trace one carry or wick opportunity from source fields through fee assumptions, risk gating, paper execution, and portfolio accounting. List every non-price failure mode.
08 Execution and operating disciplineOperate a research system with reconciliation, monitoring, and human controls instead of assuming automation creates safety.
What to understand
- Order states, partial fills, rejection, cancellation, retries, and idempotency
- Expected versus realized slippage, fee drift, and execution-quality review
- Position and cash reconciliation across broker, ledger, and strategy state
- Heartbeats, alerts, incident response, kill switches, and recovery procedures
- Paper, dry-run, small-live rollout gates, acceptance evidence, and rollback
- Behavioral bias, journals, change control, postmortems, and model governance
Experience notes
- The dangerous failure is often silent disagreement between intended, submitted, filled, and recorded positions; reconcile all four.
- A live rollout is an operational experiment. Increase scope only after predeclared evidence, never because of recent profits or fear of missing out.
QT LAB
Apply it in this project
Complete `docs/live-checklist.md` while staying in paper mode. Review four weeks of heartbeats, ledger reconciliation, fees, incidents, and DCA benchmark results.
EVIDENCE
Sources and further learning
- Official curriculum Investment Foundations Certificate
- Official curriculum Introduction to Probability and Statistics
- Official curriculum Finance Theory I
- Foundational research Portfolio Selection
- Foundational research Capital Asset Prices: A Theory of Market Equilibrium
- Foundational research Efficient Capital Markets: A Review of Theory and Empirical Work
- Foundational research The Adaptive Markets Hypothesis
- Foundational research The Probability of Backtest Overfitting
- Foundational research … and the Cross-Section of Expected Returns
- Official curriculum The crypto ecosystem: key elements and risks
- Official curriculum Crypto Assets
- Practitioner experience Systematic Trading
- Project documentation QT strategy and evidence notes
- Project documentation QT operations guide