When you reduce capital, the same EA approaches total loss — unrealized loss should be viewed as a "percentage," not an "amount."
NAMPIN EA Verification Note: Gold Reversal Nampin EA『SENJU』 Development and verification series.
Product page here
When the floating loss reaches 300,000 yen. With 1,000,000 yen capital, that's 30%; with 300,000 yen, 100% (total loss).
Even if the amount is the same, the meaning is completely different.
Last time, I wrote why a high-win-rate Nampin EA is still scary. Most baskets end with small floating losses, while, very infrequently, there was a case where it expanded to about 585,000 yen at maximum.
However, these numbers are based on “1,000,000 yen / 0.01 lot.”
This time,What happens if you run the same EA with reduced capital: I will write about it.
I have had experiences where a net loss wiped out an account, or about 80% of it, with Nampin-type EAs. I experienced the same with other EAs about twice.
The capital I put in was a solid amount, such as 1,000,000 yen.
Nampin-type EAs include “incremental” ones where lots increase as the number of orders deepens. At first it seems fine, but later the lot size grows, and floating loss becomes heavy all at once.
Profit nearly every day, enjoyable to watch. That ends up as total loss or 80% drawdown.
Floating loss has different meaning depending on capital
A hypothetical example. Suppose floating loss becomes 300,000 yen.
For an account with 1,000,000 yen, 30%
For an account with 500,000 yen, 60%
For an account with 300,000 yen, 100% (total loss)
The “amount” of floating loss is the same. However, the percentage relative to the account varies greatly with capital.
Even with the same EA, reducing capital makes the same amount of floating loss more impactful on the account.
So, with actual EAs, what happens when you run with reduced capital? I used SENJU’s backtest data to perform virtual calculations by capital level.
This article writes four things.
If capital is reduced to 500,000 and 300,000, annual results and DD
What happened in market conditions like 2023
Run with 100,000 capital“Rotating type” rules and simulation results
My thoughts on how to determine capital
Results of running with reduced capital
The standard verification uses 1,000,000 yen / 0.01 lot. Below isnot the recommended margin, but a virtual capital simulation using the same trading data.
From the perspective of “can we reach double the capital without going bankrupt,” I reset capital each year and ran actual backtests.
| Capital | 2023 | 2024 | 2025 | 2026 (Jan–Aug) |
|---|---|---|---|---|
| 500,000 | Bankruptcy-level DD about 97.7% Not reached 1,000,000 | Not reached（End of year about 939,000） DD about 48.2% | Achieved（10/10） DD about 39.5% | Achieved（4/13） DD about 39.5% |
| 300,000 | Bankruptcy-level DD about 146.6% Not reached 600,000 | Achieved（9/6） DD about 62.5% | Achieved（6/12） DD about 58.1% | Achieved（3/10） DD about 48.5% |
※ The date in parentheses is the month/day when capital reached doubling.
The more capital you reduce, the higher the efficiency of profit (growth relative to capital). On the other hand, even with the same floating loss, the percentage relative to the account rises sharply.
In market conditions like 2023, the more you reduce capital, the closer you come to total loss.With 500,000 yen, DD was about 97.7% and did not reach 1,000,000 yen. With 300,000 yen, DD was about 146.6% (an unmanageable level), not reaching 600,000 yen.
This capital-based simulation is published on the product page of “SENJU” with conditional terms.『SENJU』.
Details of the simulation here
Difference between last time’s “about 60%” and the current sequence’s “about 47%”
Last time, I wrote that the maximum floating loss in 2023 was about 585,000 yen, “about 60% for a 1,000,000 account.”
On the other hand, if you run from 2023 through 2026 August continuously across years, the maximum margin drawdown is478,042 yen (about 47%).
The difference is in how you measure it. The former is how much floating loss existed at a certain point. The latter includes profits accumulated up to that point, showing how much the account balance could decrease at most. When run in continuity, accumulated profits offset part of the floating losses.
Neither is a guarantee of future loss limits.
Running with 100,000 yen capital using the “rotating type”
I think there is no such thing as an EA that never bankrupts. In particular, with a Nampin EA, I think the more realistic approach is to have profits exit the account when achieved (“rotation”).
Enduring floating losses of about 580,000 yen with 100,000 yen is not feasible. So, I changed the operation philosophy and then re-created it in the EA with the following rules.
Start with 100,000 yen capital
When the account reaches 200,000 yen, recover 100,000 yen outside the account
If DD at around -80% occurs, that account ends
If there is recovered capital, restart with it. If not, inject another 100,000 yen
| Year | Accident count | Withdrawal count | Cumulative external capital | Final managed assets | Difference from capital |
|---|---|---|---|---|---|
| 2023 | 3 | 1 | 300,000 | 165,501円 | -134,499円 |
| 2024 | 4 | 3 | 200,000 | 175,891円 | -24,109円 |
| 2025 | 5 | 5 | 300,000 | 396,420円 | +96,420円 |
| 2026（1–8月） | 5 | 7 | 100,000 | 298,062円 | +198,062円 |
※ The difference from the capital is the final managed assets minus cumulative external capital.
In 2023, an accident occurred before recovering profits, and even with an additional 300,000 yen, the capital was wiped out. In 2024, capital was wiped out as well, but in 2025 and 2026 it exceeded the capital.
What matters is not the number of accidents, but how many times you could escape 10,000 yen before an accident occurs.
However, this is a retrospective recreation based on past data, and there is no guarantee it will go the same way in the future. There are years where even additional投入 leads to capital loss as in 2023.
Capital determination (my view)
My view is as follows.
No matter how small you cut the lot, capital will decrease if an accident occurs
Therefore, the capital used should be “disposable funds” you can afford to lose without affecting your life .
For those with little usable capital, a “rotating type” approach can be helpful. Those with more capital can choose standard capital operation
First, divide the maximum floating loss of the EA by your own capital to see the percentage relative to the account. For the previous figure of about 585,000 yen, that is about 59% with 1,000,000 yen, about 117% with 500,000 yen, and about 195% with 300,000 yen.
Summary
Floating loss is not about the amount“as a percentage of the account”.
Even with the same EA, reducing capital makes you approachtotal loss (example from 2023)
Rotating type can work with small capital, but there are years with capital loss
Even with smaller lots, accidents cannot be prevented. Use only disposable funds for operation .
This capital philosophy is published with verification data in
『SENJU』
The SENJU product page also publishes simulations with reduced capital under certain conditions.
・Standard verification:1,000,000 yen / 0.01 lot
・Virtual simulation by capital:500,000 and 300,000 (not a recommended margin)
・10万円 “rotating type”:Rules and annual results
Next time
So far I have used the word “DD” many times.
Then, what does the number “Maximum DD 10%” measure?
Even with the same EA, the numbers change depending on how you measure. Next time, I will write about the definition of DD.
