> ## Content Index
> Fetch the complete content index at: https://blog.edgepilotai.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# Normal Drawdown or a Broken Strategy? What to Check First
- URL: https://blog.edgepilotai.io/normal-drawdown-or-broken-strategy/
- Published: 2026-09-15T07:37:07.000Z
- Updated: 2026-10-15T06:36:24.000Z
- Description: A smaller loss can reveal a bigger problem. Compare the duration, trade outcomes, costs, and settings before blaming—or defending—the strategy.
- Author: EdgePilot Team
- Tags: Risk Management, #en

Your strategy is losing again. Stop it, and you worry it will recover without you. Leave it running, and you worry the losses will keep growing. The drawdown percentage alone cannot settle that argument. Compare this losing stretch with the old records to see what changed before deciding what deserves a closer look.

## Why is staying below the worst historical drawdown not enough?

Drawdown measures how far account equity has fallen from an earlier peak. It tells you the size of the decline, not its cause.

Suppose a strategy’s historical tests included several drawdowns of 8%–12%. Now consider these two situations:

__Two losing stretches, two different questions (illustrative example)__
| Current drawdown | What you observe                                      | What to check                                                |
| ---------------- | ----------------------------------------------------- | ------------------------------------------------------------ |
| 6%               | No rule or cost change identified yet                 | Duration, losing streaks, and changes in trades              |
| 3%               | Recent gross profit per trade is below cost per trade | Whether trading profits fell or costs rose, and for how long |

The smaller drawdown already comes with a specific concern: each trade is not earning enough before costs to cover those costs. The larger drawdown is not automatically harmless either. “Nothing changed that I noticed” is a reason to inspect the records, not a diagnosis.

A shop can have survived quiet months before and still have a new problem today. If the margin on each order no longer covers delivery, “we have had slow periods before” does not pay the courier.

![Octobot compares the current drawdown with its historical range and prompts a cause check](https://storage.ghost.io/c/6d/bb/6dbbb7cc-1f93-4cd3-8d75-e3c2cfce055f/content/images/2026/09/normal-drawdown-historical-range.en-US.png)

Compare the current drawdown with its historical range, then investigate the cause.

## Which records help explain a losing streak?

Do not start by defending the strategy—or declaring it dead. Put the historical and recent records side by side and check what you can actually compare.

### Are the rules and position sizes still the same?

Look for parameter changes, larger positions, and different instruments. Mark those changes before comparing P&L. Otherwise, you may blame the market for a difference you introduced yourself.

### Are you winning less often, or losing more when you lose?

Separate the recent winners and losers. Compare how often trades win and the amount won or lost each time. Then review costs separately. A fee that is already included in net P&L should not be deducted again.

### How long has this drawdown lasted?

A 6% decline that recovers after three days and a 6% decline still unresolved after three months deserve different questions. Record when the decline began, the losing streaks, and whether recovery occurred. Two percentages leave most of that story out.

### Is the market that used to pay the strategy still there?

Did the earlier profits come during sustained rises, while the recent market keeps reversing? Are the instruments and trade directions comparable? List a possible explanation, then check it against the trades. Attaching a market label to a loss does not explain it.

### Does breaking the historical record mean you must stop?

The worst historical drawdown is the deepest decline observed in that history, not a ceiling on future losses. Decide risk limits and review triggers beforehand instead of moving them whenever the account makes you uncomfortable. Hypothetical results are not a promise of future performance. [Why historical simulations need care](https://www.cftc.gov/LearnAndProtect/AdvisoriesAndArticles/fraudadv%5Ftradingsystem.html?ref=blog.edgepilotai.io)

Acting on a pre-agreed risk limit and proving that a strategy has permanently stopped working are different tasks. You do not need to finish the second before following the first.

## How do you compare the current losses with earlier results?

“It recovered last time” is easy to remember. Last time’s parameters, position sizes, and costs are easier to forget.

Bring the historical and recent material into an EdgePilot Research conversation. AI can explain the differences and suggest which changes deserve attention first. That gives the discussion actual records to work with instead of a reassuring memory.

[Install and enable the official EdgePilot Research plugin](https://www.edgepilotai.io/en/research/?utm%5Fsource=edgepilot%5Fblog&utm%5Fmedium=article%5Fcta&utm%5Fcampaign=normal-drawdown-or-broken-strategy&utm%5Fcontent=install%5Fresearch), attach the relevant material, and ask a concrete question. For the two situations above:

> Use EdgePilot Research to analyze two illustrative scenarios. Historical tests of the same strategy included several 8%–12% drawdowns. Scenario A is currently down 6%, with no rule or cost changes identified yet. Scenario B is down 3%, but recent average gross profit per trade is below average cost per trade. Why is drawdown size alone insufficient to judge strategy failure? Which historical and recent records should I compare next? Keep the answer concise.

![EdgePilot Research conversation comparing 6% and 3% drawdown scenarios and the records to review](https://storage.ghost.io/c/6d/bb/6dbbb7cc-1f93-4cd3-8d75-e3c2cfce055f/content/images/2026/09/normal-drawdown-research-comparison.en-US.png)

The useful next step is to compare the duration of losses, trade outcomes, costs, and settings. That turns a general worry about the strategy into a question you can investigate.

__Bring the current and historical records together__
| You provide                              | Ask Research to help interpret                     | Your next step                       |
| ---------------------------------------- | -------------------------------------------------- | ------------------------------------ |
| Dated historical and recent results      | Changes in loss size, duration, and trade outcomes | Choose a change to investigate first |
| Versions, parameters, and position sizes | Whether the comparison uses the same settings      | Find the original setup              |
| Gross P&L and cost records               | Lower trading profits versus higher costs          | Check the trading rule or the bill   |

For a reviewed strategy package supported by Research, you can also run a historical backtest and retain the version, parameters, market settings, fees, and result. The next review can start from that record rather than “I think this was the setup.”

![](https://storage.ghost.io/c/6d/bb/6dbbb7cc-1f93-4cd3-8d75-e3c2cfce055f/content/images/2026/09/normal-drawdown-compare-records.en-US.png)

## What should you do with the comparison?

The 6% and 3% examples are not suggested stop levels. They show why a smaller loss can reveal a specific new problem, while staying below an old record does not settle the question.

If you are caught between stopping and waiting, save both sets of results and [review the drawdown with EdgePilot Research](https://www.edgepilotai.io/en/research/?utm%5Fsource=edgepilot%5Fblog&utm%5Fmedium=article%5Fcta&utm%5Fcampaign=normal-drawdown-or-broken-strategy&utm%5Fcontent=review%5Fevidence). Start by separating cost changes, trade performance, and setting changes. Then choose the next investigation.

If costs appear to be the issue, continue with [why trading costs can erase backtest profits](https://blog.edgepilotai.io/why-trading-costs-wipe-out-backtest-profits/). “It lost money before” should not become a permanent exemption from checking the books.

## Sources

- [CFTC: trading systems and hypothetical performance](https://www.cftc.gov/LearnAndProtect/AdvisoriesAndArticles/fraudadv%5Ftradingsystem.html?ref=blog.edgepilotai.io)
- [EdgePilot Research](https://www.edgepilotai.io/en/research/?ref=blog.edgepilotai.io)

*For trading research and software-testing education, not investment advice. Historical and simulated results do not guarantee future performance.*