The evidence

A backtest of our own fair value model from 1992 to 2026, including companies that later failed, compared with the MSCI ACWI Equal Weighted from 2001. Hypothetical and chosen on the same history, with every caveat shown openly.

26 valuation modelsmodels 37 quality factorsfactors 6 AI factorsAI factors 2,000+ data pointsdata points 69 criteria per stock criteria →

11.7%

strategy per year

from 2001, median of the 12 review months

8.6%

MSCI ACWI Equal Weighted per year

after 0.25% ETF costs a year

2,361

failed companies

in the test, USA and Europe

years tested

1992 to 2026

Key findings

From 2001 to 2026 the strategy returned about 11.7%/yr (median of the 12 decision months, 100 portfolios each), the MSCI ACWI Equal Weighted after ETF costs 8.6%/yr over the same days: a gap of +3.1 pp a year. The strategy was ahead in 12 of 12 months.
Since 2009: 12.3%/yr against 8.3% for the MSCI ACWI Equal Weighted. Since 2016: 11.4%/yr against 8.1%, ahead of the index.
Over the whole test since June 1992 the strategy returned 12.9%/yr, with failed companies included. A hypothetical backtest of our own model: the rules were chosen on this same history in October 2026. Past results never guarantee the future.

The rules of the strategy

Buy

Fair Value at least 30% above the price
Quality Score of 50 or more
Uptrend: the price a month ago higher than twelve months ago

The strategy buys only stocks with all three checks. Among them, ranked 30% by upside and 70% by quality. 15 stocks, equal weight, at most 2 per sector, checked once a year.

Sell

Sell in the optimistic range: the price reaches the optimistic Fair Value
Quality break: the Quality Score falls below the median
Review after 5 years

Market: MSCI ACWI Equal Weighted, every stock with the same weight, with dividends, in US dollars, over the same days, minus 0.25% a year in ETF costs. The strategy pays 0.1% per purchase and per sale. Only stocks, no funds, ETFs or trusts. Including companies that later failed; their Fair Values are estimated, the USA in the 1990s and Europe before about 2013 are incomplete. The headline figure is the median over all 12 possible review months, 100 portfolios each.

Periods

from	Strategy	MSCI ACWI Equal Weighted	Gap	Months ahead
2001	11.7%	8.6%	+3.1 pp	12 of 12
2009	12.3%	8.3%	+4.0 pp	12 of 12
2016	11.4%	8.1%	+3.3 pp	12 of 12
1992	12.9%	no comparison: the index starts in 1997, and failed US companies before 1999 are largely missing

Per year, in US dollars, dividends reinvested, up to 2026. MSCI ACWI Equal Weighted after 0.25% ETF costs a year. Each row is the median over the 12 review months (100 portfolios each); "months ahead" counts in how many of the 12 review months the strategy was ahead of the MSCI ACWI Equal Weighted.

How it was tested

📅 Since 1992, compared from 2001Tested from June 1992 to June 2026. The comparison with the market starts in 2001: before 1999 most US companies that failed are still missing from the data, so the early years are less reliable.

🛒 Three checks to buyThe strategy buys only stocks with all three checks: Fair Value at least 30% above the price, Quality Score of 50 or more, and an uptrend (the price a month ago higher than twelve months ago).

🎯 Sold in the optimistic rangeA position is sold once the price reaches the optimistic Fair Value, when quality breaks below the median, or it is reviewed after five years. The proceeds roll into the next stock with all three checks.

🗓️ All 12 decision monthsThe portfolio is checked once a year, and the month changes the result. So we ran every month from January to December, 100 portfolios each, and report the median of the 12 months.

⚰️ Failed companies included2,361 companies from the USA and Europe that later disappeared are in the test, held until their last trading day; a bankruptcy then costs another 30 to 55%. Their Fair Values are estimated from similar companies. The USA in the 1990s and Europe before about 2013 are incomplete.

🧾 Only stocks, one quality formulaOnly stocks are tested: funds, ETFs, closed-end funds, investment trusts, VCTs and SPACs are left out, for failed and for listed companies by the same rule (201 and 188 securities). The Quality Score uses the same six measures for every company; where interest cover is missing, the average of the other five counts.

💸 Dividends reinvested, in US dollarsEvery return uses adjusted closes, so dividends and splits are reinvested. European stocks are converted to US dollars at the exchange rate of the day. Costs: 0.1% per purchase and per sale, returns before taxes.

⏳ Point-in-time, said openlyFair Value, quality and trend only use figures that were available on the decision day. What is from today: the valuation model and the rules, which were chosen on this same history.

🌍 One market: MSCI ACWI Equal WeightedThe market is the MSCI ACWI Equal Weighted: stocks from all countries, each with the same weight, as in the strategy's portfolio, with dividends, in US dollars, over exactly the same days, minus 0.25% a year for the costs of an ETF on it. Since 2016 the strategy is ahead of the MSCI ACWI Equal Weighted.

Risk, honestly

monthly portfolio value from 2001, median of all portfolios and decision months, against the MSCI ACWI Equal Weighted after ETF costs.

	Strategy	MSCI ACWI Equal Weighted
Volatility	18.7%	16.9%
Max drawdown	−57.1%	−57.4%

When we hold, when we rotate

On track Below the optimistic Fair Value, quality intactThe price has not reached the optimistic Fair Value yet and the business is sound. The position stays.

Review Optimistic Fair Value reachedThe price has reached the optimistic Fair Value. The strategy sells here and rotates into the next stock with all three checks.

Review Quality breakThe Quality Score dropped below the median, the business is weakening. The position is flagged before it becomes a value trap.

Review Held 5 yearsAfter five years a position is reviewed and rotated into a fresh stock with all three checks.

Failed companies

The test contains companies that later disappeared from the exchange (bankruptcy, takeover, delisting), from the USA and Europe. They are held until their last trading day; after a collapse another 30 to 55% is lost (Shumway 1997). Where their annual reports are missing, they take the figures of a similar failed company; their Fair Values are estimated from surviving companies with similar quality, momentum and profit. In a cross-check the real annual figures of failed companies from 2010 were hidden and estimated as well: the result changed −0.3 over the whole period, by −0.4 since 2009 and by +0.4 since 2016, in percentage points a year; the estimate can shift the result slightly.

Academic sources

Fama & French (1992, 1993): value and size premia (survivorship-free CRSP).
Jegadeesh & Titman (1993): returns to buying winners and selling losers (momentum 12-1).
Novy-Marx (2013): gross profitability premium.
Piotroski (2000): F-Score, quality of value stocks.
Asness, Frazzini & Pedersen: quality minus junk.
Bessembinder (2018): most stocks underperform T-bills; a few winners drive all wealth.
Shumway (1997): the delisting bias in stock returns.

Research & data

Dated data studies from the Fair Value engine, free to cite with attribution. Every study ships methodology, limitations and a CSV download.

Global Equity Fair Value Report November 2026
Global market vs aggregate fair value: +13.5 % Global Equity Fair Value Report October 2026
Global market vs aggregate fair value: +14.8 % Global Equity Fair Value Report September 2026
Global market vs aggregate fair value: +34.6 % Global Equity Fair Value Report August 2026
Global market vs aggregate fair value: +34.6 % Global Equity Fair Value Report July 2026
Global market vs aggregate fair value: +33.8 %

Common questions

How did the fair-value model do in the backtest?

From 2001 to 2026 about 11.7% a year, the MSCI ACWI Equal Weighted after 0.25% ETF costs a year over the same days 8.6%: a gap of +3.1 percentage points a year, dividends reinvested, failed companies included. This is the median over all 12 months in which the portfolio can be reviewed once a year.

Are failed companies in the test?

Yes. 2,361 companies from the USA and Europe that later disappeared from the exchange are included and held until their last trading day. Their Fair Values are estimated; the USA in the 1990s and Europe before about 2013 are incomplete. Only stocks are tested: funds, ETFs, closed-end funds, trusts and SPACs are left out on both sides.

How risky was the strategy?

Since 2001 volatility was 18.7% a year (MSCI ACWI Equal Weighted 16.9%), the largest drawdown -57.1% (MSCI ACWI Equal Weighted -57.4%).

Is this a guarantee of future returns?

No. It is a hypothetical backtest of our own model, the rules were chosen on the same history. Past results never guarantee the future. Not financial advice.

Not financial advice. A hypothetical backtest of our own Fair Value model, stocks only, including companies that later failed (USA and Europe, their Fair Values estimated), dividends reinvested, in US dollars, compared with the MSCI ACWI Equal Weighted after 0.25% ETF costs a year over the same days. The rules were chosen on the same history. Past results never guarantee the future.