Value at Risk, CVaR, and the Omega ratio explained

#Value at Risk, CVaR, and the Omega Ratio, Explained

Max drawdown tells you the single worst peak-to-trough loss your strategy hit. It is essential, but it is one number about one moment. It says nothing about how often bad days happen or how bad the truly ugly ones get. For that, Freya's advanced results include three tail-risk metrics borrowed from professional risk management: Value at Risk, its harsher sibling CVaR, and the Omega ratio.

They sound intimidating. The ideas behind them are simple, and once you can read them you will judge a strategy's risk with far more precision than drawdown alone allows.

These build on the core metrics in interpreting backtest results. Read that first if drawdown and volatility are new.

---

#Value at Risk: the Loss You Should Expect

Value at Risk answers a plain question: on a bad day, how much should I expect to lose. Freya reports it at two confidence levels.

VaR at 95 percent is the loss threshold you stay above 95 percent of the time. If your VaR 95 is 4 percent, then on 19 days out of 20 your loss stays smaller than 4 percent. It is a measure of your normal downside, the size of an ordinary bad day.

VaR at 99 percent raises the bar to a rarer event: the loss you stay above 99 percent of the time. It is naturally larger than VaR 95, because it is describing a deeper, less frequent bad day.

VaR is useful precisely because it is intuitive. It puts a number on "how much do I bleed when things go against me" in terms you can size around.

#CVaR: How Bad the Bad Days Actually Get

Here is the flaw in VaR on its own: it tells you the threshold of the worst 5 percent of days, but not what happens beyond it. Two strategies can have the identical VaR 95 of 4 percent, and in one the days past that threshold cluster around 5 percent, while in the other they routinely reach 30 percent. VaR cannot tell them apart. That blind spot has ended real funds.

CVaR, also called Conditional VaR or expected shortfall, closes it. CVaR 95 is the average loss on the days that fall in the worst 5 percent tail. It does not ask "where does the tail begin," it asks "how deep is the tail on average once you are in it."

If your VaR 95 is 4 percent but your CVaR 95 is 18 percent, that is the number that matters. It means when you do have a bad-tail day, you lose 18 percent on average, not 4. A strategy whose CVaR sits close to its VaR has a well-behaved tail. A strategy whose CVaR towers over its VaR is hiding catastrophe in its worst days.

#The Omega Ratio: Every Gain Against Every Loss

Sharpe and Sortino, covered in interpreting backtest results, compress risk into a single standard deviation and assume returns are roughly bell-shaped. Crypto returns are not: they have fat tails and skew. The Omega ratio makes no such assumption. It weighs the entire distribution of gains against the entire distribution of losses, probability-weighted.

Read it simply:

• Omega above 1 means the weighted gains outweigh the weighted losses. The strategy is favorable.
• Omega below 1 means losses dominate. The strategy is unfavorable regardless of a pretty equity curve.
• Higher is better, and because it captures the whole shape of returns rather than just their spread, it can rank two strategies with identical Sharpe ratios that behave very differently in the tails.

#Reading Them Together

No single one of these is the answer. Their power is in the pattern they form.

#Putting It to Work

• Size for CVaR, not VaR. VaR tells you where the bad days start; CVaR tells you how far they go. Your position sizing should survive the CVaR, because that is the loss the market will eventually hand you.
• Treat a low Omega as a veto. A strategy with a beautiful backtest return but an Omega near or below 1 is not compensating you enough for the losses it takes. The return likely rests on a few outliers.
• Combine with drawdown and Monte Carlo. Max drawdown is the worst single moment, VaR and CVaR describe the daily tail, and Monte Carlo shows how those tails compound across an unlucky ordering. Together they triangulate the real risk.
• Remember these are backtest numbers. Live tails are usually worse, because slippage and thin liquidity bite hardest on exactly the violent days these metrics measure. See backtest vs live results.

#Frequently Asked Questions

#What is the difference between VaR and CVaR?

VaR (Value at Risk) is a threshold: the loss you stay under most of the time, for example 95 or 99 percent of days. CVaR (Conditional VaR or expected shortfall) is the average loss on the days that breach that threshold. VaR tells you where the bad tail begins; CVaR tells you how deep it goes. A strategy with a CVaR far above its VaR has rare but severe losses that VaR alone would hide.

#Is a higher or lower VaR better?

Lower is better, all else equal, because VaR is a measure of loss. A smaller VaR means a smaller expected loss on a bad day. But read it with CVaR: a strategy can have a low VaR and still be dangerous if its CVaR, the depth of its worst days, is large.

#What is a good Omega ratio?

Any Omega above 1 means probability-weighted gains outweigh losses, which is the baseline for a favorable strategy. Higher is better. Because Omega uses the full shape of the return distribution rather than assuming a bell curve, it is especially useful for crypto, where fat tails make Sharpe alone misleading.

#Do I need these if I already have max drawdown and Sharpe?

They add a dimension those miss. Max drawdown is the single worst peak-to-trough moment, and Sharpe assumes roughly normal returns. VaR and CVaR describe the frequency and depth of the loss tail, and Omega captures skew and fat tails Sharpe ignores. For crypto strategies, that tail detail is often where the real risk lives.