Tested Oct 2025 – Jan 2026

We Ran
The Test.

Our DCF model says which stocks are overvalued and which are undervalued. We tracked 290 positions over 88 days to see if the market agreed. Here's what happened.

+7.2%
Alpha

88-day excess return

290
Positions

Tested

62.8%
Win Rate

Beat market

70%
SHORT Win Rate

Overvalued fell

99.7%
Confidence

Not luck

Test period: October 6, 2025 – January 2, 2026 (88 days)

Signal Breakdown

Long vs Short

Both signals generated alpha. But one stands out.

LONG Undervalued stocks
+5.1%
Avg Alpha
54.7%
Win Rate
137 positions tested
SHORT Overvalued stocks
+3.5%
Avg Alpha
69.9%
Win Rate
153 positions tested

When our model says a stock is overvalued, it underperforms the market
7 times out of 10.

That's the edge.

Statistical Validation

Is It Luck?

A 70% hit rate sounds good. But is it real?

3.01
t-statistic
>2.0 is significant
0.003
p-value
0.3% chance it's random
290
Sample Size
Independent positions

Translation: There's a 99.7% probability the alpha is real, not noise. The win rate (63% across 290 positions) has essentially zero probability of occurring by chance. The short signal win rate (70% across 153 positions) is statistically extraordinary.

It's not luck.

The 8 Factors

What The Ratings Measure

Each company is scored across 8 dimensions derived from our DCF analysis

🎯
Valuation
Price vs fair value
Quality
Margins, returns, consistency
📈
Growth
Revenue & earnings trajectory
💰
Income
Dividend yield & sustainability
🏰
Moat
Competitive advantage
🏥
Health
Balance sheet strength
Efficiency
Capital allocation & ROE
⚠️
Risk
Downside probability
Methodology

How We Tested

Signal Generation

Each stock's fair value (from our DCF model) is compared to its actual Bloomberg price. If >10% undervalued → LONG signal. If >10% overvalued → SHORT signal.

150/50 Long/Short Model

150% exposure to undervalued stocks (LONG). 50% exposure to overvalued stocks (SHORT). Net exposure: 100%. This captures alpha from both sides of the signal.

Regime-Based Tracking

Each company has one active position at a time. Positions continue until the signal changes. This avoids double-counting when a stock is re-rated.

Benchmark: ASX 300

Alpha is measured against the S&P/ASX 300 Index. During the test period, the ASX 300 returned -2.28%. The strategy returned +4.91%.

By Market Cap

Where The Edge Works

Signal performance across different market segments

Index Positions Alpha Win Rate
ASX 20 18 +3.4% 83.3%
ASX 50 39 +4.5% 74.4%
ASX 100 82 +3.5% 69.5%
ASX 200 157 +4.6% 68.2%
All Stocks 290 +4.3% 62.8%

Large caps have higher win rates. Mid-caps offer more opportunity. The edge works across the market.

If This Edge Persists

Implied Annual Alpha

Not a forecast. The edge is the relationship — if it holds across regimes.

Scenario 88-Day Alpha Implied Annual
Conservative (95% CI lower) +1.5% ~+6%
Base Case (point estimate) +4.3% ~+18%
Optimistic (95% CI upper) +7.1% ~+29%

How to read this: Alpha is the period excess return over the benchmark (industry standard). Returns can be annualized; alpha represents the edge. If the relationship between our signals and stock performance persists across market regimes, the implied annual alpha falls in this range. This assumes persistence — which is unproven over longer timeframes.

Limitations

What We Don't Know Yet

1
88 days is not enough
We need 12+ months across different market conditions. The cross-sectional breadth (290 positions) helps, but time-series depth matters too.
2
Single market regime
This period saw quality factors invert ("junk rally"). The valuation signal worked regardless, but we haven't tested bull markets, bear markets, or sector rotations.
3
No transaction costs
Results don't include brokerage, slippage, or borrowing costs for shorts. The conservative +6% estimate should absorb typical costs, but it's not proven.
4
Paper test only
These are simulated results. Real implementation involves execution risk, liquidity constraints, and timing differences.

Our position: The signal has statistically significant predictive power. We're not claiming it will always work. We're showing you what happened and letting you decide.

Why It Works

The Logic

Each rating is derived from the underlying analysis — the same analysis you can read for free on every company page. The ratings are the conclusions of that analysis, distilled into 8 measurable factors.

The signal is simple: when market price diverges significantly from our DCF fair value, mean reversion tends to occur. Undervalued stocks rise. Overvalued stocks fall. Not always, but often enough.

It's not magic. It's not a black box. You can read our analysis, see our fair values, and decide whether our logic makes sense. The performance data shows what happens when you systematically act on that logic.

Summary

The Story in One Breath

We tested 290 positions across 283 stocks over 88 days.

The portfolio returned +4.91% while the ASX 300 fell -2.28%.

That's +7.2% alpha. Nearly two-thirds of our calls beat the market.

The shorts hit 70% of the time.

Statistical confidence: 99.7%. Not luck.

The model identifies mispriced stocks.
The market corrects them.

You've seen the edge. Now see the signals.

The analysis behind every rating is free — read any company page and verify our logic.
The ratings themselves tell you which stocks are mispriced right now.

We showed you the data. You decide if you want in.