Lesson 5 of 6

Probability and expected value

Step 3 of 6

Worked example

Part 1: real odds from 98 years of the stock market

Damodaran's table gives the S&P 500's total return (price change plus dividends) for every calendar year from 1928 to 2025: 98 years. Sorting those years into three buckets (checked 20 Nov 2026):

Bucket Years Probability Average return in those years
Lost money 26 26 ÷ 98 = 26.5% −13.49%
Gained 0–20% 36 36 ÷ 98 = 36.7% +11.23%
Gained 20% or more 36 36 ÷ 98 = 36.7% +30.78%
EV = (26 × −13.49 + 36 × 11.23 + 36 × 30.78) ÷ 98
   = (−350.74 + 404.28 + 1,108.08) ÷ 98
   = 1,161.62 ÷ 98
   = +11.85% a year

That matches the simple average of all 98 years (11.85%, lesson 03), as it should. (The compound yearly return over the same 98 years is lower, about 10%. The simple average of separate years is always a bit higher when returns bounce around.)

The lesson in one line: a positive expected return of about 12%, yet more than one year in four lost money, including −36.55% in 2008. Positive EV doesn't mean every year goes up.

Part 2: a bull/base/bear case for one stock

This part is made-up numbers to show the method. A stock trades at $100. You think about where it could be in a year:

Scenario What happens Probability Price in a year
Bull New product sells well 25% $150
Base Business keeps growing slowly 50% $110
Bear Sales fall, profits shrink 25% $60

First check the probabilities add up: 25% + 50% + 25% = 100%. Then:

EV = (0.25 × $150) + (0.50 × $110) + (0.25 × $60)
   = $37.50 + $55.00 + $15.00
   = $107.50

Expected return = 107.50 ÷ 100 − 1 = +7.5%

So, on these guesses, the stock is worth about $107.50 against a price of $100. But notice: there's a 1-in-4 chance of losing 40%. Lesson 04 showed what a loss like that takes to recover. EV tells you if the odds are in your favour; the bear case tells you whether you can live with being wrong.

The numbers are only as good as your probabilities. If you change the bear case to 40% and the bull to 10%, the EV drops to $94. Try it.