Systematic thinking
Quantitative and algorithmic approaches replace gut feel with rules: a strategy defined so precisely that it could be written as code and tested on history. This module explains the concepts and — just as importantly — the traps, so you can read them critically.
Statistics you actually need
- Mean & standard deviation — the centre and the spread; volatility is just standard deviation of returns.
- Correlation — how two series move together (the heart of diversification).
- Distributions & fat tails — market returns have far more extreme moves than a “normal” bell curve predicts.
- Sample size & significance — a pattern from ten trades is noise; robustness needs many independent observations.
Backtesting — and its traps
A backtest applies a rule to historical data to see how it would have performed. Done carelessly it produces beautiful, meaningless results. The classic traps:
- Overfitting — tuning a strategy so tightly to the past that it captures noise, not signal.
- Look-ahead bias — accidentally using information that wasn't available at the time.
- Survivorship bias — testing only on companies that still exist, ignoring the ones that failed.
- Ignoring costs — brokerage, taxes, slippage and impact can turn a “profitable” backtest into a real-world loss.
Factors, execution & the strategy lifecycle
Factor investing tilts a portfolio toward characteristics that have historically been associated with returns — value, quality, momentum, low-volatility, size. Execution and algos are about how large orders are worked into the market (VWAP, TWAP) to reduce impact — how institutions trade, not a prediction tool.
A sound systematic idea moves through a lifecycle: hypothesis → data → backtest → out-of-sample test → small live test → monitor and retire. Skipping the sceptical steps is how people lose money confidently.
Key terms
Overfitting
Fitting a model so closely to past data that it fails on new data.
Slippage
The difference between the expected and the actual execution price.
Factor
A measurable characteristic (value, momentum, quality) historically linked to returns.
Out-of-sample test
Testing a strategy on data it was never tuned on — the real test of robustness.
Test yourself
1. Overfitting a backtest means…
Overfitting captures random noise and fails out of sample.
2. Survivorship bias comes from…
Ignoring companies that failed flatters historical results.
3. Realistic backtests must include…
Costs and slippage often turn paper profits into real losses.
FAQs
No. This module explains quantitative and algorithmic concepts — including their pitfalls — so you can think critically about them. PCJ does not provide algorithmic signals, strategies or advice, and most retail 'algo tip' offers are marketing to be wary of.
Usually because of overfitting, look-ahead bias, survivorship bias and ignored costs. A rule tuned to the past can capture noise that doesn't repeat, and real brokerage, taxes and slippage eat into results. Out-of-sample testing is the antidote.
Tilting a portfolio toward measurable characteristics — value, quality, momentum, low-volatility, size — that have historically been associated with returns. It's a framework, not a guarantee; factors go through long periods of underperformance.
No. The point is to think systematically and sceptically — to judge any 'strategy' by whether it survives costs and out-of-sample data. That mindset helps every investor, coder or not.
Educational content for general awareness only — not investment, trading or tax advice, and not a recommendation to buy or sell any security. PCJ Holdings does not provide research or advisory services. Examples and calculator outputs are hypothetical and illustrative. Investments in securities markets are subject to market risks; read all related documents carefully. Figures are indicative for FY 2025-26 and may change.