The Drawdown Severity Score: A Smarter Way to Measure When Stocks Are on Sale

A 20% drawdown in Apple is a significant event. Apple's average historical max drawdown is 5.7%, so a 20% drop is roughly 3.5x what's typical. That's noteworthy.

A 20% drawdown in Bitcoin? That happens routinely. BTC's average max drawdown is much higher due to its inherent volatility. A 20% drop might not even register as unusual.

This is the problem with using raw percentage drops to assess whether a stock is "on sale." The same percentage means completely different things depending on the asset's volatility profile.

The Drawdown Severity Score solves this. It's a volatility-adjusted metric that normalizes drawdowns across all asset types, so a score of 5 means the same thing whether you're looking at Apple, Tesla, or Bitcoin.

How the Score Works

The severity score adjusts each asset's current drawdown based on its own historical volatility:

• Low-volatility assets (like Apple, with an average max drawdown of 5.7%) get stricter scoring. A 15% drop already pushes Apple toward severity 3, because drops that size are genuinely unusual for Apple.
• High-volatility assets (like Nvidia, with an average max drawdown of 8.4%) get more forgiving scoring. Nvidia needs a larger drop to reach the same score, because deeper pullbacks are more normal for a stock with Nvidia's volatility profile.

The math automatically adapts as a stock's behavior changes over time. If a previously stable company becomes more volatile, the scoring adjusts.

The Three Zones

The score maps to three color-coded zones:

🟢 Green Zone (0-2): Normal conditions. The asset is within its typical historical range. No action needed unless you're looking for new positions at fair value.

🟡 Yellow Zone (2-5): Elevated drawdown. The asset has pulled back more than usual. Worth monitoring if it's on your watchlist. This is where value-oriented investors start paying attention.

🔴 Red Zone (5+): Severe drawdown. The asset is experiencing a historically unusual decline for its volatility profile. These events are rare and have historically been associated with the best long-term entry points (though never guaranteed).

A Concrete Example

Here's what Apple looks like right now:

Apple is at severity 0.1, sitting 0.6% below its all-time high of $275.25. This is completely normal. Nothing to see here.

Now compare that to PayPal:

PayPal is at severity 17.2. It's 78.3% below its all-time high of $308.53 and has been in drawdown for 1,571 days. This is the most extreme reading in PayPal's entire history.

The raw percentages (0.6% vs 78.3%) tell you these are very different situations. But the severity scores tell you something more precise: Apple's current position is entirely within its normal range, while PayPal's is 17x what's typical for PayPal specifically.

What the Score Doesn't Do

The severity score does not predict whether a stock will recover, when it will recover, or how much upside exists. A high severity score means the drawdown is historically unusual. That's a necessary condition for a good buying opportunity, but it's not sufficient on its own.

Some stocks with high severity scores recover and make new highs. Others have permanently lower valuations because the business fundamentally changed. The score identifies the statistical anomaly. The investment decision still requires analysis of the underlying business.

Historical Context: The "What History Says" Data

Where the severity score becomes most powerful is when combined with historical event data. For every tracked asset, we calculate how many times it has experienced drawdowns of similar magnitude in the past, and how long those events typically lasted.

This table shows every completed drawdown event where Nvidia dropped 15% or more from its high. You can see the dates, the depth of each drop, and how long each one lasted. When the next severe drawdown hits Nvidia, you'll be able to compare it directly against this historical baseline.

That's the real value of the severity score system: it turns the vague question "should I buy the dip?" into a specific, data-informed question: "how does this drawdown compare to every other drawdown this stock has experienced?"