The Stop-Loss That Stops Gains
Why Your Drawdown Rules Are Probably Hurting You
New research shows that the risk management technique beloved by traders frequently does the opposite of what it’s supposed to do.
If you’ve been trading for any length of time, you’ve heard some version of this advice: “Cut your losses at 10% (or 15%, or 20%). Don’t let a small loss become a big one.”
It sounds sensible. It feels right. It’s the kind of rule that makes you feel like you’re in control.
And it’s almost certainly costing you money.
In a new paper published in the Journal of Portfolio Management, I examined nearly three decades of data across dozens of ETFs and found something that should trouble anyone who relies on mechanical drawdown rules: they rarely work, and when they fail, they fail in a particularly insidious way.
Death by a Thousand Cuts
Here’s what typically happens when you implement a strict drawdown rule:
Your position drops 10%. You exit, as planned. The market recovers. You miss it.
You re-enter. It drops 8%. Not quite your threshold. You hold. It drops another 3%. You exit at 11% down.
The market recovers again. You’re now nursing two losses instead of one.
This pattern—which I call “death by a thousand cuts”—can produce larger cumulative losses than simply riding out a single, larger drawdown would have. You’re not managing risk. You’re locking in losses that a subsequent recovery would have erased.
I know what you’re thinking: “But what about the times when the market doesn’t recover? What about the catastrophic drawdowns?”
Fair question. Let’s look at the data.
The COVID Crash: When Your Risk Metrics Lie to You
The most striking evidence comes from examining how traditional risk metrics behaved during the COVID crash of February-March 2020.
During the 22 trading days from February 21 to March 23, 2020, SPY produced an annualized return of -99% with a maximum drawdown of -33%. This was the fastest market crash in modern history.
Where did traditional drawdown-adjusted performance (DAP) rank this period?
The 25th percentile.
Read that again. The standard metric that most risk systems use ranked the worst market crash in decades as merely below-average—not catastrophic, not even particularly concerning. Just... meh.
It gets worse. A slightly different window (February 11 to March 12) with marginally better performance actually received a worse DAP score. The metric was literally ranking outcomes backwards.
This is the kind of thing that blows up accounts. Your risk model is telling you everything’s fine while the building is on fire.
Why This Happens (And How to Fix It)
The problem is mathematical. Traditional DAP divides return by drawdown. When returns are negative, larger drawdowns paradoxically improve the ratio. It’s like a thermometer that shows warmer temperatures as you freeze to death.
In the paper, I introduce a corrected metric called Coherent Drawdown-Adjusted Performance (CDAP):
r×|d|−sgn(r)
This looks intimidating but it’s actually simple. Here’s what each piece means:
r is your return
|d| is the absolute value of your drawdown (drawdowns are always expressed as positive numbers)
sgn(r) is just the sign of your return: +1 if positive, -1 if negative
The negative sign in front flips it: so when returns are positive, you’re raising drawdown to a negative power (which means dividing); when returns are negative, you’re raising it to a positive power (which means multiplying)
The key insight: when returns are positive, we divide by drawdown (larger drawdowns penalize the score). When returns are negative, we multiply by drawdown (larger drawdowns also penalize the score). The metric behaves logically in all conditions. I talked through this in more detail on the Titans of Tomorrow podcast.
Applied to those same COVID crash periods, CDAP correctly ranked them in the bottom 0.2% of all observations—exactly where risk metrics should place such extraordinary market stress.
Why Your S&P 500 Intuition Will Betray You
Part of why drawdown rules persist is that they seem to work for the one market most traders know best: US large-cap equities.
When I analyzed probability structures across different markets, the S&P 500 showed a pattern that superficially supports drawdown rules. Large drawdowns occur much more frequently than normal distribution theory predicts—up to 7x more often for severe drops. If you believe tail risks are that extreme, cutting losses seems prudent.
But here’s the catch: this pattern doesn’t generalize.
The Japan ETF (EWJ) shows some similarity. The UK ETF (EWU) shows no clear pattern at all. And critically, even where the probability patterns look similar, implementing drawdown cutoffs still doesn’t improve results. Something beyond conditional probability structure matters—recovery patterns, regime dynamics, reentry timing—and simple rules can’t capture it.
This is a trap I’ve watched traders fall into for 30 years. They develop their intuition on SPY or ES futures. They learn that cutting losses “works.” Then they apply the same logic to other markets, other instruments, other timeframes—and it fails. They blame themselves, tighten their rules, and make it worse.
The problem isn’t discipline. The problem is the rule itself.
The Long-Short Illusion
I also tested whether drawdown rules might work better for sophisticated long-short strategies. The initial results looked promising: across the full history of various long-short pairs, drawdown cutoffs appeared to improve performance.
Then I looked closer.
Most of these pairs generate minimal returns over their full history—which makes sense, since long-short trades are typically implemented only when a specific thesis is active. The apparent “benefit” of drawdown rules was simply limiting losses on unprofitable trades that shouldn’t have been on in the first place.
When I examined the same pairs during periods when their investment thesis was actually relevant, the benefits largely disappeared. Even with perfect hindsight about which periods to trade, drawdown rules provided no consistent edge.
What Actually Works
If not mechanical cutoffs, then what?
Here’s the uncomfortable truth: your risk rules need to be embedded into your trading system and tested as part of it. Not bolted on afterward. Not borrowed from someone else’s system. Not copied from a trading book.
The drawdown rule fails because it’s generic. It doesn’t know anything about your entries, your exits, your timeframes, your instruments. It’s a one-size-fits-all solution applied to a problem that’s specific to your system.
Here’s something else most traders get backwards: almost anyone can, with a little bit of work, come up with entries. When to buy. What to buy. There are a thousand YouTube videos about entry signals, and honestly, most of them are fine. Entries are the easy part.
The hard part is exits—and risk management is a gigantic piece of that. When do you get out? How much do you give back before you cut? How do you size down when things go wrong? These questions don’t have universal answers. They have answers that are specific to your system, your instruments, your timeframe, your psychology.
And that’s exactly why you have to test. Test, test, and test again. Then retest.
The paper develops a framework built on three principles:
1. Context matters. A 10% drawdown during a VIX spike means something different than a 10% drawdown in calm markets. During stress periods, drawdowns typically show greater persistence. During normal periods, you see more mean reversion. Your response should differ accordingly.
2. Cross-asset confirmation. An isolated drawdown in one position might be idiosyncratic—and potentially an opportunity. Simultaneous drawdowns across multiple assets signal systematic risk requiring action. Before you cut, ask: is this just me, or is it everywhere?
3. Dynamic thresholds. Static rules can’t adapt to changing market regimes. A percentile-based classification system provides more accurate risk categorization across different environments.
But here’s what matters more than any framework: you have to do the work.
If you’re a systematic trader, you need to backtest your risk rules as part of your system, not as an add-on. See how they actually perform across different regimes, different instruments, different timeframes.
If you’re a discretionary trader looking at charts, the process is harder but the principle is the same. You need to look at thousands of charts—without bias—and mark each one with your entries and exits. Not cherry-picked winners. Not “obvious” setups in hindsight. Thousands of them, good and bad, boring and exciting. Then, only afterward, calculate your performance. See what your drawdown rules actually did to your equity curve.
And then? Find another thousand charts and do it again. On different instruments. Different timeframes. Different market conditions. Keep doing it until you find something that works—something that’s yours, tested on your system, with your rules.
Most traders won’t do this. It’s tedious. It’s unglamorous. There’s no dopamine hit from marking up your 847th chart.
But that’s exactly why it works. Trading is 99.99% sweat and 0.01% inspiration. The traders who put in the work are the ones who survive. Everyone else is just borrowing someone else’s rules and hoping they fit.
The Psychology Trap
I want to be direct about something: this research suggests that many widely-used risk management practices don’t just fail to help—they actively harm performance.
That’s uncomfortable. These rules feel responsible. They give us agency during uncertainty. When you’re staring at a losing position, having a rule to follow feels better than sitting with the discomfort.
But the behavioral finance literature is clear about why we gravitate toward explicit rules even when they underperform: loss aversion and the psychological need for control. We prefer the illusion of risk management over actual risk management.
I talked about this on the Titans of Tomorrow podcast—why most trading systems fail long before you notice, why the risk models that look best on paper blow up when you need them most. The drawdown rule is a perfect example. It’s simple, it’s intuitive, and it’s wrong.
Real risk management isn’t about simple rules. It’s about frameworks that enhance judgment with systematic analysis—knowing when drawdowns are transitory noise versus genuine warning signals, when to stay patient versus when to act decisively.
That’s harder to implement. But the evidence suggests it’s the only approach that actually works.
The full paper, “The False Promise of Drawdown Rules: New Evidence and a Better Framework,” is available in the November 2025 issue of the Journal of Portfolio Management. It includes detailed empirical analysis, supplementary exhibits, and implementation guidelines for the CDAP framework.
I’m the founder and chief investment officer of VS Asset Management. If you caught my conversation with Waqar on Titans of Tomorrow, this paper is the rigorous version of what we discussed about risk models that fail when you need them most. Questions, pushback, or war stories about drawdown rules that burned you? Drop them in the comments.
Check out my book, The Science of Free Will for insight into how deterministic physics shapes all kinds of things: from the behavior of bees to why a Supreme Court must exist.
Disclaimer: The views and opinions expressed in this article are solely those of the author and do not necessarily reflect the official policy or position of VS Asset Management, LLC. This content is for informational and educational purposes only and should not be construed as investment advice, a recommendation, or an offer or solicitation to buy or sell any securities or financial instruments. Past performance is not indicative of future results. All investments involve risk, including the possible loss of principal. The research and analysis presented herein are based on historical data and there is no guarantee that any investment strategy or framework discussed will be successful in the future. Before making any investment decisions, you should consult with a qualified financial advisor and consider your own financial situation, investment objectives, and risk tolerance. The author and VS Asset Management, LLC disclaim any liability for any direct or indirect loss or damage arising from any use of this information or reliance on the opinions expressed herein.
DAP between 2/21, 3/23 would be -99/33 =-3, how did you arrive at this is 25th percentile?
Is it based on the DAP calculated monthly for last N months?