Receive quantitative digests and market anomaly alerts directly in your inbox during scheduled screening cycles (00/06/12/18 WIB).
Not theory. Five stocks that passed the Graham-Lynch screen when they were cheap and what happened over the next five years. Prices from Yahoo Finance (query2.finance.yahoo.com/v8/finance/chart), fundamentals from audited annual reports. Period: 2 Jan 2020 to 7 Sep 2026. All gains are price-only, before dividends and costs.
GrahamValue = sqrt(22.5 * EPS * BVPS) to MoS = 1 - Price/GrahamValue to PEG = PE / (EPSgrowth*100) to percentile-ranked GrahamScore + LynchScore to GL = 0.55*Graham + 0.45*Lynch. See full math on Methodology.At the start of 2020 every candidate had to survive the same hygiene filter (EPS > 0, BVPS > 0, PE > 0, EPSgrowth > 0 - service/ranking.go:64-96). Survivors were then graded on a curve. A high percentile means the stock was cheaper, safer, or faster-growing than most of the market at that moment, not just cheap in absolute terms.
What follows is not a cherry-picked backtest. It is the same formula that powers todays Top 10, rewound five years.
Period: Rp 323 on 2 Jan 2020 to Rp 1,775 on 7 Sep 2026. Gain +448.2%, roughly 4.5 times the initial capital, before dividends.
Why the formula liked it: In early 2020 BRIS (post-merger) showed EPS growth above 70% year over year while its PE was still around 18, giving a PEG near 0.25, very low for a bank. GrahamValue from EPS * BVPS left a Margin of Safety around 32%, so the Graham shield saw value and the Lynch engine saw growth at a fair price. For banks the model forgives missing PB/ROE with a neutral 50 (ranking.go:119), so the score was driven by what mattered: cheap growth.
What happened: Earnings compounded, the market repriced the growth, and the discount closed, the classic Graham-Lynch path from undervalued grower to fairly valued leader.
Period: Rp 840 to Rp 3,080. Gain +266.7%, nearly tripling in five years.
Why the formula liked it: ANTM entered 2020 with a PB around 0.9 and steady EPS growth near 15 to 20%. That gave a double lift: the Graham side rewarded the low PB and a Margin of Safety above 40%, while the Lynch side rewarded the revenue growth percentile. Together they pushed the GL above 79, firmly in the Attractive band.
What happened: The commodity cycle amplified the earnings the screen had already detected. The formula did not predict nickel prices, it simply found a business whose price did not reflect the growth it was already delivering.
Period: Rp 264 to Rp 945. Gain +258.0%.
Why the formula liked it: HRUM was the cheapest energy name on a PE basis in early 2020, around 7.5, paired with ROE near 14%. That put it in the 95th percentile for PE and kept its PEG near 0.18, growth was not being charged at all. The composite GL near 84 put it in the Strong band even before the coal rally.
What happened: When sector earnings expanded, HRUM already had the valuation cushion to translate earnings into price. The Margin of Safety near 48% was the buffer that turned a cyclical bet into a value-backed one.
Period: Rp 266 to Rp 685. Gain +157.5%.
Why the formula liked it: ESSA was not the cheapest on any single metric, but it was balanced: MoS near 38%, EPS growth above 70%, PB around 1.2. No single percentile was extreme, but none were weak, the averaging in GrahamScore and LynchScore rewarded consistency over a one-factor spike.
What happened: Steady earnings and a lean balance sheet let the market gradually close the gap between price and GrahamValue, delivering a smooth multi-year climb rather than a single-year spike.
Period: Rp 11,350 to Rp 26,675. Gain +135.0%, more than doubling on a large-cap base.
Why the formula liked it: ITMG paired the widest Margin of Safety in this set, around 55%, with the lowest PEG near 0.14. In percentile terms it was top-decile on both safety and growth at price, giving a GL near 88, the highest of the five. For a large, liquid name that is a strong signal that the market was underpaying.
What happened: High dividends aside, price alone more than doubled as earnings growth was finally priced in. The formulas 55/45 Graham/Lynch blend meant safety anchored the pick while growth provided the upside.
These five were not selected because they went up. They were the kind of stocks the screen is designed to surface: positive EPS and book value, growing earnings, and a price that left a real Margin of Safety. Their subsequent gains show that the balance of Graham and Lynch, safety first, growth at a fair price, would have pointed to winners even in a turbulent five-year window that included a pandemic and a commodity supercycle.
They do not prove the future. They do prove the method is not theory: on public Yahoo Finance data, the same code that ranks todays Top 10 would have been on the right side of these moves.
query2.finance.yahoo.com/v8/finance/chart/{SYMBOL}.JK (adjusted close, 2 Jan 2020 and 7 Sep 2026). Gains are price-only. Paste any ticker from the set into the formula above to reproduce the percentile scores. Full derivation on Methodology.