Two projects with the same expected return but different variance get treated as equivalent.
Your VP pitches two projects for next quarter, both asking for $100K of Budget. You build a model and both show a 33% Expected Return. You greenlight the one with more upside potential. Three months later it ships to crickets - you get back $95K, a 5% loss. The other project would have returned at least $115K. Both had the same Expected Return. You picked the one with ten times the Variance and didn't know it.
Variance measures how spread out possible outcomes are around the Expected Return. Two Capital Investments with identical Expected Returns can have wildly different risk profiles - Variance is the number that captures this difference, and ignoring it is how operators blow holes in their P&L.
Variance is the probability-weighted average of squared deviations from the Expected Return. In plain math:
Variance = Σ pᵢ × (rᵢ - E[R])²
Where pᵢ is the probability of outcome i, rᵢ is the return in that outcome, and E[R] is the Expected Return you already know how to calculate.
The squaring does two things: it makes all deviations positive (a miss is a miss whether it's above or below the average), and it penalizes large deviations disproportionately. A return that's 20 percentage points off the mean contributes 4x more to Variance than one that's 10 points off.
The unit of Variance is "percentage points squared," which is awkward to interpret directly. That's why you'll often see Standard Deviation instead - it's the square root of Variance, which puts you back in the same units as the returns themselves.
Your P&L doesn't care about averages. It cares about what actually happens this quarter.
If your team has $200K in Fixed Obligations - salaries, tools, office lease - and you invested $200K in a project with 25% Expected Return, the expected outcome is $250K back. Comfortable.
But Expected Return is an average across possible worlds. In any single quarter, you live in exactly one world. If that project has high Variance, the actual return might be -30%, leaving you with $140K against $200K in obligations. That's a $60K Income Shortfall. No amount of "but the expected value was positive" fixes that.
Low-Variance projects let you plan. High-Variance projects force you to hold an Emergency Fund, delay commitments, or accept the risk of missing Fixed Obligations. The cost of Variance is real even when the Expected Return looks identical.
Step 1: Start with the outcomes and probabilities you used for Expected Return.
Project Steady ($100K Capital Investment):
Project Moonshot ($100K Capital Investment):
Same Expected Return. Now calculate Variance.
Step 2: For each outcome, compute the squared deviation from Expected Return.
Project Steady:
Project Moonshot:
Step 3: Take the probability-weighted average of those squared deviations.
Project Steady Variance: 0.6 × 144 + 0.4 × 324 = 86.4 + 129.6 = 216
Project Moonshot Variance: 0.4 × 3,249 + 0.6 × 1,444 = 1,299.6 + 866.4 = 2,166
Moonshot has 10x the Variance of Steady.
Step 4: Take the square root to get Standard Deviation.
Steady Standard Deviation: √216 ≈ 14.7 percentage points
Moonshot Standard Deviation: √2,166 ≈ 46.5 percentage points
This means Moonshot's actual return will typically land about 47 percentage points above or below the 33% Expected Return. That range spans roughly -14% to +80%. Steady lands within about 15 points of 33% - a range of roughly 18% to 48%. Both center on 33%, but Moonshot can lose you money while Steady never does.
Calculate Variance any time you're comparing Capital Investments, projects, or Budget allocations that have similar Expected Returns:
You have $100K to invest in one of two projects. Project Steady has a 60% chance of returning 45% and a 40% chance of returning 15%. Project Moonshot has a 40% chance of returning 90% and a 60% chance of returning -5%. Both have a 33% Expected Return.
Calculate Expected Return for both: Steady = 0.6 × 45% + 0.4 × 15% = 33%. Moonshot = 0.4 × 90% + 0.6 × (-5%) = 33%. Confirmed equal.
Compute squared deviations from the mean (33%): Steady deviations are (45-33)²=144 and (15-33)²=324. Moonshot deviations are (90-33)²=3,249 and (-5-33)²=1,444.
Weight by probabilities: Steady Variance = 0.6×144 + 0.4×324 = 86.4 + 129.6 = 216. Moonshot Variance = 0.4×3,249 + 0.6×1,444 = 1,299.6 + 866.4 = 2,166.
Convert to Standard Deviation: Steady SD = √216 ≈ 14.7pp. Moonshot SD = √2,166 ≈ 46.5pp.
Interpret: Moonshot has 10x the Variance. Its outcomes range from losing $5K to gaining $90K. Steady's outcomes range from gaining $15K to gaining $45K. Steady never loses money. Moonshot loses money 60% of the time.
Insight: Expected Return told you these projects were equivalent. Variance reveals that Moonshot is a fundamentally different bet - one where the most likely single outcome is a loss, despite the positive average.
Your team runs $200K per quarter in Fixed Obligations (salaries, tools, office lease). You allocated $200K to a project with 25% Expected Return to fund next quarter's operations. You're comparing a low-Variance version (50/50 chance of 35% or 15% return) against a high-Variance version (50/50 chance of 80% or -30% return).
Expected Return for both: 0.5×35% + 0.5×15% = 25%. And 0.5×80% + 0.5×(-30%) = 25%. Both equal.
Variance: Low = 0.5×(35-25)² + 0.5×(15-25)² = 50+50 = 100 (SD = 10pp). High = 0.5×(80-25)² + 0.5×(-30-25)² = 1,512.5 + 1,512.5 = 3,025 (SD = 55pp).
Map to dollar outcomes against $200K investment. Low-Variance worst case: $200K × 1.15 = $230K. High-Variance worst case: $200K × 0.70 = $140K.
Compare to Fixed Obligations. Low-Variance worst case: $230K - $200K = $30K Profit. High-Variance worst case: $140K - $200K = -$60K shortfall. You can't make payroll.
The high-Variance path has a 50% chance of an Income Shortfall that forces emergency action - cutting staff, drawing on your Emergency Fund, or taking on high-interest debt.
Insight: When your P&L has Fixed Obligations, Variance isn't just a statistical concept - it's the probability of a Cash Flow crisis. The Expected Return says everything is fine. The Variance tells you there's a coin flip between Profit and crisis.
Variance measures how far outcomes spread from the Expected Return - two investments with identical Expected Returns can have 10x different Variance, making them fundamentally different bets.
The cost of Variance is real: it forces you to hold an Emergency Fund, limits how aggressively you can commit to Fixed Obligations, and creates the possibility of outcomes your P&L cannot survive.
Always ask for the Variance (or Standard Deviation) before treating two options as equivalent based on Expected Return alone - it's the single most common gap in how operators evaluate Capital Investments.
Treating Expected Return as sufficient for comparison. Two projects that both show 25% Expected Return feel interchangeable, but if one has a Standard Deviation of 10pp and the other 55pp, they are not the same decision. The first is a predictable win; the second loses money half the time.
Ignoring Variance because you plan to 'just pick the upside scenario.' Variance is symmetric in the math but not in your P&L. A 30% upside miss means less Profit. A 30% downside miss means you might not cover Fixed Obligations. The downside matters more than the upside when you have real obligations to meet.
A Capital Investment of $50K has three possible outcomes: 40% chance of +60% return, 35% chance of +20% return, and 25% chance of -10% return. Calculate the Expected Return, Variance, and Standard Deviation.
Hint: First compute Expected Return as the probability-weighted average: Σ pᵢ × rᵢ. Then compute Variance as Σ pᵢ × (rᵢ - E[R])². Standard Deviation is the square root of Variance.
Expected Return = 0.40×60% + 0.35×20% + 0.25×(-10%) = 24% + 7% + (-2.5%) = 28.5%. Squared deviations: (60-28.5)²=992.25, (20-28.5)²=72.25, (-10-28.5)²=1,482.25. Variance = 0.40×992.25 + 0.35×72.25 + 0.25×1,482.25 = 396.9 + 25.29 + 370.56 = 792.75. Standard Deviation = √792.75 ≈ 28.2 percentage points. In dollars: best case $80K back, worst case $45K back, expected $64.25K back.
You're choosing between two projects for your team's $150K quarterly Budget. Project A: 70% chance of +30% return, 30% chance of +10% return. Project B: 50% chance of +55% return, 50% chance of -1% return. Your quarterly Fixed Obligations are $155K. Which project do you choose and why?
Hint: Calculate Expected Return and Variance for both. Then map worst-case dollar outcomes against the $155K Fixed Obligations threshold. The right answer depends on whether the worst case covers your obligations.
Project A: E[R] = 0.7×30% + 0.3×10% = 24%. Variance = 0.7×(30-24)² + 0.3×(10-24)² = 25.2 + 58.8 = 84 (SD ≈ 9.2pp). Worst case: $150K × 1.10 = $165K. Project B: E[R] = 0.5×55% + 0.5×(-1%) = 27%. Variance = 0.5×(55-27)² + 0.5×(-1-27)² = 392 + 392 = 784 (SD = 28pp). Worst case: $150K × 0.99 = $148.5K. Project B has higher Expected Return (27% vs 24%) but its worst case ($148.5K) falls $6.5K short of Fixed Obligations ($155K). Project A always covers obligations. Unless you have other Cash Flow to absorb a shortfall, Project A is the correct choice despite its lower Expected Return.
Your CFO presents three Capital Investment options, all with 20% Expected Return on a $500K investment. Option X has Standard Deviation of 5pp, Option Y has 25pp, and Option Z has 50pp. Your company has $400K in Fixed Obligations next quarter. For each option, calculate the break-even return needed to cover obligations, and check whether each option's worst reasonable outcome (Expected Return minus 2 Standard Deviations) clears that threshold.
Hint: Break-even return = (Fixed Obligations - Investment) / Investment. Then check if E[R] - 2×SD is above or below that break-even point. The 2-Standard-Deviation rule gives a rough worst-case bound.
Break-even return = ($400K - $500K) / $500K = -20%. You need at least -20% return to cover obligations. Now check E[R] - 2×SD for each: Option X: 20% - 2×5% = 10%. Returns roughly $550K at 2 SDs down. Clears $400K easily. Option Y: 20% - 2×25% = -30%. Returns $350K at 2 SDs down, a $50K shortfall against obligations. Option Z: 20% - 2×50% = -80%. Returns $100K at 2 SDs down, a $300K shortfall. Option X is safe. Option Y creates meaningful risk of missing obligations. Option Z is betting the company. Same Expected Return, completely different risk profiles. Variance - not Expected Return - determines which options are viable given your Fixed Obligations.
Variance builds directly on Expected Return by answering the question Expected Return deliberately ignores: how much can the actual outcome differ from the average? Where Expected Return gives you the center of the Return Distribution, Variance measures the width. This leads directly to Standard Deviation (the square root of Variance, easier to interpret since it shares units with the returns), and from there to Sharpe Ratio (which divides excess return by Standard Deviation to produce a single Risk-Adjusted Return number). Understanding Variance is also the foundation for Volatility (Standard Deviation of returns measured over time), Tail Risk (what happens in the extreme outcomes that Variance only partially captures), Skew (whether the Return Distribution is lopsided), and Bet Sizing (how much of your Budget to commit given the Variance of the opportunity). At the Portfolio level, Variance is what makes Capital Allocation non-trivial - it's the reason you can't just rank projects by Expected Return and fund them top-down.
Disclaimer: This content is for educational and informational purposes only and does not constitute financial, investment, tax, or legal advice. It is not a recommendation to buy, sell, or hold any security or financial product. You should consult a qualified financial advisor, tax professional, or attorney before making financial decisions. Past performance is not indicative of future results. The author is not a registered investment advisor, broker-dealer, or financial planner.