Music for Reading · 5 Pieces

On by default: it starts after your first click, and you can turn it off anytime.

Background music is produced and licensed by DingSong.

The 3-Minute Version

First the one-sentence method, then the four data charts that matter most, each with a short note. To read the whole story, follow the link into that section of the book.

The method in one sentence

Save every month, invest at regular intervals in passive, broad-market index ETFs, hold for the long run and never sell at the highs — and when markets crash, add more from new income.

1Figure 5-6

The Longer You Hold, the Better Your Odds of Making Money

Since 1928, US stocks held for 20 years or more have never lost money after inflation — but that is the US record. In Japan, someone who bought at the 1989 peak was still deeply in the red 20 years later. The past is no guarantee of the future.

The Longer You Hold US Stocks, the Better Your Odds of Beating Inflation (1928–2026)

The Longer You Hold US Stocks, the Better Your Odds of Beating Inflation (1928–2026)0%20%40%60%80%100%70%1 year76%3 years78%5 years87%10 years95%15 years100%20 years100%30 yearsSource: Robert Shiller's S&P 500 data (shillerdata.com, Oct. 2026 edition); author's calculations; with dividendsThe Longer You Hold US Stocks, the Better Your Odds of Beating Inflation (1928–2026)0%20%40%60%80%100%70%1y76%3y78%5y87%10y95%15y100%20y100%30ySource: Robert Shiller's S&P 500 data(shillerdata.com, Oct. 2026 edition); author'scalculations; with dividends

Read this section in Chapter V 

2Redrawn from Figure 5-7

"Sell When Pricey" Left Less Money

In the author's backtests, selling everything when the price reached 1.5 times its 10-year average, and buying back only when it fell to the average, left 70%–97% less money than simply holding, in the US, Japan and Taiwan alike; even with a 2-times threshold it left 31%–66% less. The market can stay pricey for many years, and people who sell often wait more than a decade. I used to do this myself; after testing it against the data, I dropped it.

"Sell When Pricey, Buy Back at the Average" vs. Simply Holding: Final Wealth

"Sell When Pricey, Buy Back at the Average" vs. Simply Holding: Final Wealth−100%−75%−50%−25%0%−97%USsell above150%−66%USsell above200%−78%Japansell above150%−47%Japansell above200%−70%Taiwansell above150%−31%Taiwansell above200%150% / 200% = sell everything when the price reaches 1.5 or 2 times its 10-year average; buy back when it falls to the average. Same amount invested each month.US 1934–2026 with dividends, after inflation; Japan 1968–2026 price index; Taiwan 2003–2026 with dividends (total-return index). Author's backtest"Sell When Pricey, Buy Back at the Average" vs. Simply Holding: Final Wealth−100%−75%−50%−25%0%−97%US150%−66%US200%−78%Japan150%−47%Japan200%−70%Taiwan150%−31%Taiwan200%150% / 200% = sell everything when the pricereaches 1.5 or 2 times its 10-year average; buyback when it falls to the average. Same amountinvested each month.US 1934–2026 with dividends, after inflation;Japan 1968–2026 price index; Taiwan 2003–2026with dividends (total-return index). Author'sbacktest

Read "Testing It with More Countries and Longer Periods" in Chapter V 

3Figure 5-3

Buying in Crashes: Different Results in the US, Japan and Taiwan

Doubling up from new income in crashes (the three "add" bars) raised total wealth in all three markets, but whether the extra money paid off varied widely: compared with dollar-cost averaging alone, each unit invested ended about 20% higher in the US and about 4% higher in Taiwan, but 11%–15% lower in Japan (the extra money in Japan still grew, just less than the regular money did). The three "hoard" bars show keeping 20% back as cash and waiting for a fall, which did worse in all three markets. A single country can lose decades, so it is best to spread across the whole world.

Final Wealth per Unit Invested, Compared with Dollar-Cost Averaging Alone

Final Wealth per Unit Invested, Compared with Dollar-Cost Averaging Alone−20%−10%0%+10%+20%+30%−2%USHoard 20% cashfor a 20% dip+21%USAdd from incomewhen down 20%−6%JapanHoard 20% cashfor a 20% dip−13%JapanAdd from incomewhen down 20%−4%TaiwanHoard 20% cashfor a 20% dip+4%TaiwanAdd from incomewhen down 20%US 1934–2026 with dividends, after inflation; Japan 1968–2026 price index; Taiwan 2003–2026 with dividends (total-return index). Author's backtestFinal Wealth per Unit Invested, Compared with Dollar-Cost Averaging Alone−20%−10%0%+10%+20%+30%−2%UShoard+21%USadd−6%Japanhoard−13%Japanadd−4%Taiwanhoard+4%TaiwanaddHoard = keep back 20% of each month's money ascash and invest it after a 20% fall. Add = investdouble from new income when the market is down20%.US 1934–2026 with dividends, after inflation;Japan 1968–2026 price index; Taiwan 2003–2026with dividends (total-return index). Author'sbacktest

Read "In a Crash, Add More from 'New Income'" in Chapter V 

4Figure 5-2

Fees Compound Too: The 50-Year Gap between 0.1% and 2.5%

A single NT$10,000 investment earning 8% a year for 50 years comes to about NT$448,000 with costs of 0.1%, but only about NT$145,000 with costs of 2.5%. The lower the cost, the more cheaply you are buying.

NT$10,000 Invested for 50 Years: The Gap between 0.1% and 2.5% Fees (NT$)

NT$10,000 Invested for 50 Years: The Gap between 0.1% and 2.5% Fees (NT$)0100K200K300K400K500K01020304050≈ NT$448K≈ NT$145KNo fees (8%)0.1% fees: index fund2.5% fees: active tradingAuthor's calculations; assumes a long-run market return of 8%NT$10,000 Invested for 50 Years: The Gap between 0.1% and 2.5% Fees (NT$)0100K200K300K400K500K01020304050≈ NT$448K≈ NT$145KNo fees (8%)0.1% fees: index fund2.5% fees: active tradingAuthor's calculations; assumes a long-run marketreturn of 8%

Read "Buying Low Also Means Low Costs" in Chapter V 

Have more time? Take the 30-minute route, or read the whole book from the preface. To run the numbers yourself, try the Data Lab.