Methodology & Backtesting

17 years of live-like simulation, January 2009 to June 2026. 1% slippage applied to every trade. FactSet data.

Portfolio vs. S&P 500

January 2009 – June 2026 · indexed to 100

Performance Summary

MetricStrategyS&P 500
PeriodJan 2009 – Jun 2026Jan 2009 – Jun 2026
Annualized Return27.29%
Total Return6,683.50%1,282.99%
Sharpe Ratio1.62
Max Drawdown-34.16%-33.22%
Win Rate61%
Annual Turnover95.64%
Positions24
RebalancingWeekly
Slippage Applied1.0%
Data SourceFactSet
UniverseEurope · USA · Canada
Correlation w/ S&P 5000.651.00

Methodology

Investment Universe

Small and micro-cap primary listings in Europe, the United States, and Canada. This segment is structurally under-researched, creating persistent pricing inefficiencies that systematic factor investing can exploit.

Rebalancing & Sizing

The portfolio holds 25 equally-weighted positions. Scores are refreshed daily; the rebalancing action — selling positions that dropped below 94% and buying the next highest-ranked stock — occurs once a week.

Transaction Costs

A conservative 1.0% one-way slippage is applied to every trade, accounting for bid-ask spreads, market impact, and real-world execution friction in the small-cap universe.

Data Source

All factor inputs are sourced from FactSet — an institutional-grade financial data provider. The same data used in the backtest is used in the live screener, ensuring consistency.

The Ranking System

How the Score Is Produced

All stocks in the universe are ranked simultaneously using the 18-factor composite model. The ranking system produces a single composite score between 0 and 100% for each stock — reflecting where it stands relative to the entire universe across all factors combined. A score of 100% identifies the top-ranked stock in the universe at that point in time.

Score ≥ 94% — Hold
Stock ranks in the top tier of the universe. It remains in the portfolio. The 25 highest-scoring stocks are held at all times.
Score < 94% — Sell
Stock has dropped below the threshold. It is sold at the next weekly rebalance and replaced by the highest-ranked stock not already in the portfolio.

The Rotation Rule

Every week, each held position is checked against the 94% score threshold. Any position whose score has fallen below 94% is sold. The freed capital is redeployed into the highest-ranked stock not already held, maintaining exactly 25 positions at all times. Scores are refreshed daily — the rebalancing action itself happens once a week.

Rebalancing Process

Weekly Rebalancing Process

How the model builds and maintains the portfolio each week

1
Universe
~2,000
small & micro-cap stocks
Europe · USA · Canada
2
Ranking
18
factors evaluated
composite score 0–100%
3
Portfolio
Top 25
stocks purchased
equal weight
Weekly Check
Every
week, scores update
model re-ranks universe
Is each held position still among the top-ranked stocks? (Ranking > 94%)
Yes
Hold

Position remains in the portfolio. No action taken.

No
Rotate
SELLthe position that has fallen out of the top tier
BUYthe #1 ranked stock not already held in the portfolio

1% slippage applied to every buy and sell transaction · Equal weighting maintained

Portfolio Turnover

Annual turnover of 95.64% means the portfolio rotates approximately once per year. With 24 positions rebalanced weekly, each holding lasts roughly 12 months on average before a higher-ranked stock displaces it.

This is an active rotation strategy — not buy-and-hold. The 1.0% slippage assumption accounts for the transaction costs generated by this turnover. The backtest performance reflects returns after these costs.

95.64%
Annual Turnover
~12 months
Avg Hold Period
Weekly
Rebalancing

Factor Selection Process

The 18 factors were not chosen arbitrarily. The selection process began with a candidate pool of approximately 300 factors, covering every major category of quantitative signals used in academic and practitioner research.

Each factor was tested individually for alpha generation in the small and micro-cap universe. The top performers were then subjected to a pairwise correlation analysis. Factors with high mutual correlation — those effectively measuring the same underlying characteristic — were eliminated, retaining only the one with stronger standalone performance.

The result is a set of 18 factors with low pairwise correlations, each contributing independent information to the composite score. This signal diversification is what keeps performance stable across different market regimes — when one factor type underperforms, others compensate.

Data Quality — Powered by FactSet

Every factor in this model is sourced from FactSet Research Systems — one of the world's leading institutional-grade financial data providers, trusted by thousands of asset managers, hedge funds, investment banks, and research institutions globally.

Point-in-Time Data

FactSet's database stores financial data as it was known at each historical date — not as subsequently restated. This eliminates look-ahead bias, ensuring the backtest reflects what an investor could actually have known and acted on at each rebalancing date.

Institutional Coverage

FactSet provides coverage across 70,000+ securities globally, including the small and micro-cap segment where analyst coverage is thin and data quality from retail-grade providers deteriorates significantly. The data used here meets the same standard as that used by quantitative hedge funds.

Consistency Across Markets

Fundamental data for European, US, and Canadian listings is normalised into a consistent reporting framework, enabling true cross-market comparisons. IFRS and US GAAP line items are mapped to common definitions before any factor is computed.

Live Screener — Same Source

The live screener uses the same FactSet data feed that powered the backtest. There is no gap between the historical simulation and what subscribers see today — the inputs, definitions, and factor computations are identical.

Why this matters: Many retail quantitative tools use free or low-cost data feeds that suffer from survivorship bias, restatement errors, and inconsistent coverage of non-US markets. Using FactSet removes these failure modes — the backtest results are a realistic representation of what the strategy would have produced, not an artefact of clean data.

Factor Categories

The 18 factors are grouped into 6 independent categories. Each category captures a different dimension of stock quality — together they form a diversified, multi-signal composite.

Momentum5×

Price trend and relative strength signals across multiple timeframes

Profitability4×

ROI, ROA, ROE, and operating income growth metrics

FCF / Yield3×

Free cash flow and shareholder return signals

Valuation3×

Earnings yield, EV/Sales, and EBITDA/EV metrics

Low Volatility2×

Price deviation and financial stability filters

Quality1×

Piotroski F-Score for comprehensive fundamental quality

The 18 Factors in Detail

Each factor is equal-weighted at 5.56% of the composite score. All are sourced from FactSet and computed consistently across Europe, the US, and Canada.

1
Price / 52-Week HighMomentum

Measures how close the current price is to its 52-week high. Stocks near their highs tend to continue outperforming.

2
TRSD30DMomentum

Risk-adjusted short-term momentum. Measures total return volatility over 30 days, capturing near-term momentum with a volatility penalty.

3
ROI% TTMProfitability

Return on Investment over the trailing twelve months. Identifies companies generating strong returns on deployed capital.

4
FCF Yield (per Share)FCF / Yield

Free cash flow per share divided by price. A high ratio indicates the company generates substantial cash relative to its stock price.

5
26-Week Price Change (Industry)Momentum

26-week price change relative to industry peers. Stocks outperforming their industry group tend to maintain that edge.

6
Sub-Industry Total ReturnMomentum

Aggregate total return of the stock's sub-industry group. Riding sector momentum in addition to individual stock momentum.

7
Operating Income GrowthProfitability

Year-over-year operating income growth vs. prior year quarter. Growing operating profits signal improving business fundamentals.

8
Medium-Term Price MomentumMomentum

Ratio of the 120-day price to the 180-day price. Captures medium-term momentum without the reversal noise of very short windows.

9
Shareholder YieldFCF / Yield

Dividends paid plus net share buybacks as a percentage of market cap. Rewards companies that return capital to shareholders.

10
EV / Sales TTMValuation

Enterprise Value divided by trailing twelve-month sales. Lower values indicate cheaper companies relative to revenue.

11
Low Volatility (252-Day)Low Volatility

252-day price deviation. Lower volatility stocks are scored higher, capturing the well-documented low-volatility anomaly.

12
ROA% QuarterlyProfitability

Return on Assets for the most recent quarter. Efficient asset utilization is a hallmark of quality businesses.

13
FCF YieldFCF / Yield

Free cash flow yield based on enterprise value. High FCF yield companies are often undervalued relative to their cash generation.

14
Median ROE (12-Month)Profitability

Median return on equity over the past 12 months. Using the median rather than point-in-time reduces noise from seasonal earnings.

15
EBITDA / EVValuation

Operating income before depreciation divided by enterprise value. The inverse of EV/EBITDA — higher is more attractive.

16
Interest Coverage TTMLow Volatility

EBIT divided by interest expense. Measures financial stability — companies that can comfortably cover debt payments are less risky.

17
Earnings YieldValuation

Earnings per share divided by price (inverse P/E). High earnings yield indicates value, a cornerstone of quantitative value investing.

18
Piotroski F-ScoreQuality

Composite quality score (0–9) based on profitability, leverage, and operating efficiency signals. 9 = highest quality.