Value at Risk (VaR) is a statistical measure that answers one question every trader eventually faces: how much can I lose over a given period, and how likely is that? Formally, VaR expresses the maximum expected loss over a specific time horizon at a given confidence level — for example, a daily VaR of $500 at 95% confidence means there is only a 5% chance of losing more than $500 in a single trading day. Institutional desks have used VaR for decades; Basel III requires banks to report a 10-day, 99% VaR for market risk capital. For retail forex traders, VaR becomes relevant the moment you hold more than one position at a time.
Key Takeaways
- VaR requires three inputs: confidence level, time horizon, and a dollar loss amount — changing any one changes the output entirely.
- Per-trade risk calculations ignore correlation; VaR captures it, which matters when holding EUR/USD and GBP/USD simultaneously (typical correlation: 0.75–0.90).
- VaR tells you the loss threshold, not the severity of losses beyond it — always pair it with Conditional VaR (CVaR) for a complete picture.
How Value at Risk Works
VaR is defined by three variables:
VaR = f(confidence level, time horizon, loss amount)
Example: "1-day 95% VaR = $500"
- Confidence level: 95%
- Time horizon: 1 trading day
- Loss amount: $500
- Interpretation: 5% chance of losing more than $500 today
Three calculation methods — each with honest tradeoffs:
Historical Simulation replays actual past returns on your current portfolio. If EUR/USD fell 1.2% on 47 of the last 1,000 trading days, that 4.7th percentile becomes your 95% VaR input. No distribution assumptions required, but it is entirely backward-looking — a regime that never appeared in your sample window will not show up.
Parametric (Variance-Covariance) assumes returns follow a normal distribution. You input the mean and standard deviation of each position, plus the correlation matrix across positions, and the formula produces VaR directly. Fast and transparent, but normal distributions underestimate tail events by design — forex returns have fat tails.
Monte Carlo Simulation generates thousands of random return scenarios based on modeled distributions and correlations, then reads VaR from the resulting loss distribution. Most flexible, captures non-linear instruments, but overkill for a retail spot forex account.
Scaling VaR across timeframes uses the square root of time rule: weekly VaR ≈ daily VaR × √5. If your daily VaR at 95% confidence is $200, your weekly VaR is approximately $447. Swing traders holding positions for several days should apply this to understand true weekly exposure.
Quick Reference
| Aspect | Detail |
|---|---|
| Formula | VaR = f(confidence level, time horizon, loss amount) |
| Common confidence levels | 95% (retail), 99% (institutional) |
| Time horizons | 1 day (active traders), 10 days (Basel III) |
| Time-scaling | Weekly VaR ≈ Daily VaR × √5 |
| Critical limitation | Says nothing about loss magnitude beyond the threshold |
| Upgrade | Conditional VaR (CVaR) / Expected Shortfall |
Practical Example
A prop firm trader holds three open positions on a $50,000 FTMO account:
- Long EUR/USD, 1 lot, $10 stop → $100 risk
- Long GBP/USD, 1 lot, $10 stop → $100 risk
- Short USD/JPY, 0.5 lot, $15 stop → $75 risk
Naive total risk: $275 — well under a 1% account risk rule ($500). The problem is that EUR/USD and GBP/USD carry an 85% historical correlation. A broad USD strengthening event — a hawkish Fed statement, a strong NFP print — hits both long positions simultaneously. A 1-day 95% portfolio VaR calculation using historical daily moves for both pairs might return $380, not $275, because the correlated exposure compounds.
Now consider the FTMO 5% daily loss limit: $50,000 × 5% = $2,500. That $380 VaR looks comfortable against $2,500. But if the trader runs this same three-pair book at 5× the size, the portfolio VaR approaches $1,900 — and a bad day on correlated USD pairs could breach the limit before individual stops are even hit. Understanding VaR reveals that FTMO’s daily cap is a tighter constraint than raw per-trade math suggests.
Value at Risk is a risk measure that tells you the maximum you are likely to lose over a set period with a given level of confidence. For example, a daily 95% VaR of five hundred dollars means there is a five percent chance of losing more than that amount in one trading day.
Common Mistakes
-
Treating VaR as a worst-case number. VaR is a threshold, not a ceiling. The January 15, 2015 CHF flash crash saw EUR/CHF drop approximately 30% in minutes — a 6-sigma move that a 99% VaR model assigned near-zero probability. Losses on that day were not slightly above VaR; they were multiples of it.
-
Ignoring correlation when sizing multiple positions. EUR/USD and GBP/USD typically correlate at 0.75–0.90. Running both at full size does not create diversification — it concentrates USD exposure. Simple per-trade risk math misses this entirely; portfolio VaR captures it.
-
Not upgrading to CVaR. Conditional VaR (CVaR), also called Expected Shortfall, measures the average loss in the worst X% of scenarios — the region VaR leaves blank. If 95% VaR is $500, CVaR answers: “In the 5% of days where I lose more than $500, what do I lose on average?” That number is what matters for survival.
-
Forgetting to scale for time. A trader who checks daily VaR but holds positions for a week is measuring the wrong horizon. Apply the square root of time rule to align VaR with your actual holding period.
How PipJournal Tracks Value at Risk
PipJournal’s portfolio analytics surface correlated pair exposure across open and closed trades, so traders can see when their book is concentrated in USD-directional risk rather than genuinely diversified. The max drawdown and drawdown dashboards provide the realized loss data needed to back-calculate historical VaR from your own trading record — a more accurate input than generic market data.
