Deciding how you will lose before deciding how you will win — this is the order shared by everyone who survives long-term in the markets, whether they trade discretionarily or run an EA (automated trading system). Nobody knows what the market will do next. That is precisely why you need to design, in advance, how small you can keep the damage when you are wrong, before you even think about how much you will gain when you are right.
This page centers on the “2% rule,” the backbone of FX money management, and explains the lot-sizing formula with worked numerical examples, where to place your stop-loss, the relationship between risk-reward (RR) and win rate, and why averaging down (nampin) is so dangerous — all in a way you can understand through numbers rather than rote memorization. Our lab discloses even the maximum floating loss of our own EA (MAC v2.0), so we will open up this “defensive framework” using real data as well. This is the single most important lesson in the discretionary STEP02 series. Reading it right after the previous lesson on basic entry patterns should make it click even more.
Why It Matters — Stop-Loss and Money Management Are the Lifeline Shared by Discretionary Trading and EAs
No matter how good your method is, you will be forced out of the market without money management. The reason is simple: losses compound. How much of your account balance a single loss represents completely changes how much you have left after a losing streak.
For example, someone who caps each loss at 2% of their balance still has about 90% of their balance left after five straight losses — plenty of room to recover. Someone who loses 10% each time, on the other hand, sees their balance cut nearly in half after just a few losses in a row, and recovering from there requires winning back many times the amount lost. Check the difference in the quick-reference table below.
| Losing Streak | Losing 2% Each Time | Losing 5% Each Time | Losing 10% Each Time |
|---|---|---|---|
| 3 losses in a row | approx. 94.1% | approx. 85.7% | approx. 72.9% |
| 5 losses in a row | approx. 90.4% | approx. 77.4% | approx. 59.0% |
| 10 losses in a row | approx. 81.7% | approx. 59.9% | approx. 34.9% |
| Profit needed to recover (to restore the original balance) | Relatively light | Somewhat heavy | After 10 losses, the balance must grow to about 2.9x just to recover |
The key point is designing for “dying small when you’re wrong” before designing for “being right.” This is exactly where speculation (in-the-moment gambling) and investing (a system you can sustain) diverge. Locking in this defensive framework before spending time studying methods may look like a detour, but it is actually the fastest path.
A Note from Our Researcher
When I once took a big floating loss during an FX intervention, the cause was not my method itself but the fact that “a single position was too large relative to my account.” Even when your direction is right, you can still be forced out because you couldn’t withstand a temporary move against you — this is the first wall beginners run into. Since then, I always decide “where I’ll cut” and “what lot size” before I ever place an order.
The Per-Trade Risk Allowance (the 2% Rule) and Lot Sizing
The 2% rule fixes the amount you allow yourself to lose on a single trade at 1-2% of your account balance. The reason for 2% is that within this range, you retain plenty of room to recover even after a losing streak, and your mental state is less likely to break down. If a single loss is too large, the urge to “win it back” on the next trade distorts your judgment, easily creating a vicious cycle that widens the damage further. The trick is fixing the limit as “a rate against your balance,” not as a fixed amount of money.
The Lot-Sizing Formula (Working Backward from Your Balance)
At the center of money management is the following lot-sizing formula.
Lot size = (account balance x allowed %) / (stop-loss width in pips x value per pip)
What matters here is the order of thinking. Rather than deciding “how many lots to enter with” first, you think in the order: “decide where to cut (the stop-loss width) first, and therefore how many lots to use.” Lot size is not a number you freely type in yourself — it is a “result” that is automatically determined by your stop-loss width and your risk allowance. Get this order backward, and no amount of studying anything else will make your money management work.
Let’s walk through a concrete numerical example: a balance of 100,000 yen, a 2% allowance (= 2,000 yen), and a 20-pip stop-loss width. Since the value of 1 pip differs by currency pair, we’ll compare USD/JPY and gold (XAUUSD) side by side.
| Item | USD/JPY Example | Gold (XAUUSD) Example |
|---|---|---|
| Account balance | 100,000 yen | 100,000 yen |
| Risk allowance (2%) | 2,000 yen | 2,000 yen |
| Stop-loss width | 20 pips | 20 pips (equivalent to about a $2.0 move) |
| Value per pip (rough guide, based on 1 lot = 10,000 units / 1 ounce) | approx. 100 yen/pip | For gold, the price range and pip definition vary by broker, and the value per pip is large |
| Allowable loss (converted to pip value) | 2,000 yen / 20 pips = up to 100 yen/pip | Even with the same 2,000 yen, the allowable lot size tends to be smaller |
| Resulting approximate lot size | approx. 10,000 units (equivalent to 0.1 lot) | Needs to be kept to a smaller lot size than USD/JPY |
Even with the same “100,000 yen balance, 20-pip stop-loss,” the appropriate lot size changes completely depending on the pip value of the instrument you’re trading. With an instrument like gold, where the pip value is large and price movement is volatile, you have to shrink the lot size considerably even with the same risk allowance, or you’ll exceed 2%. That’s why a fixed habit like “always 0.1 lot” is dangerous, and it’s essential to work backward through this formula every single time. If you don’t know the value per pip, always check your broker’s contract specifications (contract size) for the instrument you’re trading.
In AI Terms
Lot sizing comes down to subtraction: “First decide how much you’re allowed to lose (2% of your balance). Next decide where you’ll give up (the stop-loss width). Once these two are fixed, the size of the position you enter is automatically determined as a single number.” At our lab, we have our AI read the actual stop-loss width and lot size for every trade, and record afterward, as verification data, whether the intended risk really did stay within 2%.*This is an AI-generated interpretation and does not guarantee future performance.
How to Decide Where to Place Your Stop-Loss (Amount / Rate / Pips / Technical)
There are broadly four ways to decide where to place your stop-loss. Each has strengths and weaknesses, and for beginners we recommend the technical basis (placing it outside the market structure).
| Method | Description | Strengths | Weaknesses |
|---|---|---|---|
| Fixed by loss amount | Fixed as an amount, e.g. “up to 2,000 yen per trade” | Intuitive and easy to understand | Ignores market structure, so you’re easily stopped out at a meaningless location |
| Fixed by loss rate | Fixed as a rate, e.g. “up to 2% of balance” | Pairs well with the 2% rule | On its own, doesn’t determine “how many pips to cut at” |
| Fixed by pips | A fixed width, e.g. “always 20 pips” | Directly feeds into lot sizing | Ignoring market volatility makes you easy to hunt |
| Fixed by technical levels (recommended) | Placed slightly outside support/resistance or the most recent swing point | Has the rationale that “if this level breaks, my scenario is invalidated” | Requires first learning how to read market structure |
Why You Place It “Slightly Outside” Support/Resistance or the Most Recent Swing
Under the technical basis, you place your stop-loss a bit further outside the most recent high or low (swing point) identified via support and resistance lines or Dow theory. The reason is that this is a place where you can objectively judge, “if this breaks, my scenario has fallen apart.” It becomes easier to place your stop-loss if you think of it not as admitting defeat, but as setting up “the signal that tells you the market thinks your premise was wrong.”
One thing to watch out for: placing your stop-loss right at a round number or exactly on a line makes it “easy to hunt.” When many people place their stop-losses at the same spot, closing orders pile up there. The market can briefly break through, sweeping up that pile of orders, and then reverse right back. This is what SMC calls a liquidity sweep (or stop hunt). The topic of stop-loss placement leads directly into SMC — we cover this in detail in the SMC/ICT introductory roadmap. This is exactly why you tuck your stop “slightly outside” rather than right on the line.
Risk-Reward (RR) — The Break-Even Point With Win Rate
RR (risk-reward) is a ratio expressing how many multiples the profit you target on a single win is of the loss you allow on a single loss. It’s calculated as “profit width / loss width.” With a 20-pip stop-loss and a 40-pip profit target, RR is 40 / 20 = 2, in other words 1:2.
Once you know your RR, you can see the minimum win rate (the break-even win rate) that trade needs in order to be viable over the long run. Take a look at the quick-reference table below.
| RR (Loss:Profit) | Win Rate Needed to Break Even | Meaning |
|---|---|---|
| 1:1 | 50% | Win 1 out of every 2 trades to break even |
| 1:2 | approx. 33% | Win 1 out of every 3 trades to break even |
| 1:3 | 25% | Win 1 out of every 4 trades to break even |
| 3:1 (loss bigger than profit) | 75% | You need to win 3 out of every 4 trades just to break even |
This is where beginners easily fall into the misconception that “a high win rate equals a good method.” When your RR is unfavorable (a single loss is bigger than a single win), your overall P&L tips negative even with a high win rate. Conversely, with an RR of 1:2 or 1:3, you can end up net positive even with a win rate of only 30-40%. Keep this perspective in mind — it sets up the next section, which explains the averaging-down style that “looks like a high win rate but still gets you forced out of the market.” For an eye toward spotting dangerous designs, the dangerous EA checklist is also a useful reference.
The Dangers of Averaging Down (Nampin) — Why No Hard Stop-Loss Is So Frightening
Averaging down (nampin, a martingale-style scaling-in technique) is a method where, when the price moves against you, you add more positions in the same direction, pulling your average entry price toward a more favorable level. Because it only takes a small retracement in price to reach your average entry, the win rate alone can look extremely high. This is the true nature of the illusion that it “looks like you win a lot.”
The problem arises when the adverse move continues. In a design with no hard stop-loss (SL), positions keep stacking up with each adverse move, and the floating loss expands in stages. It stays quiet while you’re winning, but when a trend beyond what you anticipated arrives, you approach ruin all at once — that is the risk structure unique to averaging down.
Our Lab’s MAC v2.0 (1.2x Multiplier, Up to 15 Steps): Disclosing the Worst-Case Scenario in Numbers
Our lab’s gold-only EA, MAC v2.0, is designed by combining SMC-based analysis with averaging down. The actual settings are: “initial lot 0.1 / averaging multiplier 1.2x / up to 15 steps / 30-pip spacing / 15-pip TP / no hard SL (managed on the EA’s side) / +0.1 lot for every 10,000 yen of capital,” and it is running live in an HFM copy-trading account (targeting roughly +10% monthly). We disclose, without hiding anything, what happens numerically if this setup were to stretch “all the way to the worst case.”
| Step | Lot at That Step (1.2x) | Cumulative Lot | Adverse Move to That Step (30-pip Spacing) |
|---|---|---|---|
| Step 1 | 0.10 | 0.10 | 0 pips |
| Step 2 | 0.12 | 0.22 | 30 pips |
| Step 3 | 0.14 | 0.36 | 60 pips |
| Step 5 | 0.21 | approx. 0.74 | 120 pips |
| Step 10 | 0.52 | approx. 2.6 | 270 pips |
| Step 15 (max) | approx. 1.28 | approx. 7.2 | 420 pips |
If it stretches to step 15, the cumulative lot size reaches approximately 7.2 lots, an exposure roughly 72 times the initial 0.1 lot. Since gold is an instrument where the monetary value of 1 pip is large, holding this much of a position through a scenario where the adverse move continues for as much as 420 pips means the projected floating loss vastly exceeds the initial capital, and the required margin also spikes. If it touches the line for a margin call or forced liquidation, the loss becomes realized right there. This is the “worst-case scenario” for averaging down with no hard SL.
We don’t want this to be misunderstood as “it’s dangerous, so never use it under any circumstances.” Our lab’s stance is “understand the risk in numbers, then use it only within a testing allocation of surplus funds you can afford to lose without affecting your life.” That’s exactly why we publish not just the wins, but also the maximum floating loss and the number of averaging-down steps, on our performance dashboard. For an averaging-down EA, everything comes down to whether you can prepare capital that can withstand this worst-case scenario. For how to think about margin, go on to check the pages on EA money management and HFM’s risk.
In AI Terms
In a single sentence, averaging down is “a mechanism where you win repeatedly on small retracements, in exchange for paying it all back at once during the occasional big adverse move.” When we have our lab’s AI read through every MAC v2.0 trade, behind the seemingly high win rate, a history of floating losses becomes visible — how many times, and how many steps deep, the position stacked up on a given day. The heart of verification isn’t just the win-rate number, but looking at its distribution — how far it went at worst.*This is an AI-generated interpretation and does not guarantee future performance.
Applying This to Discretionary Trading — The Habit of Deciding “Where You’ll Cut” Before You Enter
The biggest trick to making money management work in discretionary trading is fixing all three of SL (stop-loss), TP (take-profit), and lot size before you place the order. The order is: “decide the stop-loss position, then set the profit target from your RR, then work backward to the lot size using the 2% rule.” Only once these three points are settled do you actually enter.
This habit also plays the role of blocking “jumping in” (chasing the price at the top) — impulsively entering just because you saw the price move. We covered the entry patterns themselves in basic entry patterns, but the point is not deciding “where to cut” after you’ve waited and entered — rather, you’re able to wait precisely because where you’ll cut is already decided before you enter. This order is the essence of the defensive framework. Once you’ve fixed these three points, you execute them mechanically even if a subsequent floating loss rattles you.
A Note from Our Researcher
Early on, the reluctance to lock in a floating loss tends to make you shift your stop-loss. But the moment you move a line you decided “I’m cutting here today,” even once, every rule after that falls apart. To keep myself from moving my cut point later, I place the SL order at the same time as my entry order, and then leave the rest up to the market.
A Useful Lens for Evaluating EAs and Automated Trading — Working Backward From Max DD to Required Margin
The money-management mindset doubles directly as a tool for judging whether an EA is good or bad. When choosing an EA, look first at “max DD (drawdown)” rather than the advertised profit rate. Max DD is the largest decline your capital experienced at any point during operation.
Recommended margin is worked out backward by multiplying this max DD by a safety factor. For example, our lab’s other EA, SMC Gold Sniper, produced a profit factor of 1.87 and a max DD of 8.2% in backtesting (2018-2026), and we judge whether to adopt it based on whether we can allocate capital with enough cushion above that “8.2%.” If you run it with capital that can’t withstand the max DD from the backtest, then even if the design is profitable over the long run, you’ll be forced out first by a drawdown along the way. We go into detail on how to specifically read the numbers in how to read EA performance metrics, and on margin allocation in EA money management.
Applying This to AI Analysis — AI That Cuts With Zero Emotion vs. Humans Who Hesitate
What makes cutting a loss difficult is not a matter of skill but of psychology. Humans hesitate over a stop-loss they already decided on, driven by emotions like “I don’t want to lock in this floating loss” or “maybe it’ll come back if I wait just a little longer.” This is where AI and EAs have a clear advantage: with no emotions, they can execute the stop-loss mechanically the moment the conditions are met.
There is, however, an important distinction to make. “Being able to cut mechanically” and “a design with no stop-loss at all, i.e. averaging down” are completely different things. What makes AI and EAs excellent is executing rules without hesitation — but if the rules themselves contain no stop-loss (as with hard-SL-free averaging down), the floating loss will still balloon even though it’s a machine. The virtue of AI and automated trading is “being able to cut without losing to emotion,” and that does not justify “a design that never cuts.” With this distinction in mind, you can see the full picture of how discretionary trading and AI divide roles in What Is Discretionary Trading? (Its Relationship With EAs and AI).
In AI Terms
It’s easiest to think of AI’s stop-loss execution as “an enforcer that never bends the rules for emotion.” At our lab, we have the AI spell out an explicit criterion — “cut when this condition is met” — and afterward verify whether it actually cut according to that criterion. AI handles the “cut without hesitation” part that humans struggle with, while humans decide “whether to build a design that cuts at all” in the first place — this division of labor is the realistic approach.*This is an AI-generated interpretation and does not guarantee future performance.
Discretionary vs. SMC vs. EA Terminology Cross-Reference
The terms surrounding stop-losses and money management go by different names and are handled differently across discretionary trading, SMC, and EAs. Here’s a reference table to keep things from getting confusing.
| Common Discretionary Term | SMC Term | Handling in EAs (Automated Trading) |
|---|---|---|
| Stop-loss | Stop position (outside the structure) | SL price fixed as a parameter, or managed by internal logic |
| Getting hunted at a round number or exactly on a line | Liquidity sweep (stop hunt) | Widening the stop distance / building it into the entry conditions |
| Risk-reward (RR) | Risk:reward (extension from the POI) | Quantified as TP width / SL width, and an optimization target |
| Averaging down / adding to a position | — (a separate lineage from SMC’s original philosophy) | Controlled via lot-multiplier, step-count, and spacing parameters |
| Whether your capital can withstand it | — | Recommended margin worked out backward from max DD x safety factor |
Summary
Stop-losses and money management are the shared “lifeline for survival” in both discretionary trading and EAs. To recap the key points: fix each single loss as a rate of 1-2% of your balance (the 2% rule); treat lot size as a result worked out from “balance x allowed % / (stop-loss width in pips x pip value),” not as an input; place your stop-loss slightly outside the structure to avoid liquidity sweeps; look at RR together with the win rate it requires so you’re not fooled by the high-win-rate trap; and understand averaging down (with no hard SL) in numbers for its worst-case scenario, handling it only within a testing allocation of surplus capital. These are the five points.
As a next step, move on to support and resistance, which underpins where you place your stop-loss, and dig deeper into margin calculations for EAs via EA money management and how to read EA performance (max DD). To check whether our lab actually keeps to this framework in practice, take a look at our performance dashboard (losing months included). The full picture of discretionary trading education is gathered on the discretionary trading fundamentals hub.
Frequently Asked Questions
- Q. What percentage is the “correct” stop-loss size?
A. Using 1-2% of your account balance as a benchmark is a widely used approach. What matters is fixing it as a “rate,” not an “amount.” Fixing it as a rate keeps your per-trade risk constant even as your balance rises and falls, leaving room to recover even after a losing streak. - Q. Are all averaging-down EAs bad?
A. Not necessarily. What matters is whether you understand the mechanism and the worst-case scenario (how many steps, cumulative lot size, projected floating loss) in numbers, and whether you can uphold the premise of handling it only with capital that can withstand it. Our lab positions MAC v2.0 as a high-risk testing allocation for surplus funds, and we publish the max floating loss as well. - Q. How much money should I start with?
A. What matters comes before the size of the amount: “can you keep each single loss fixed at 2% of your balance?” It’s safer to first check the behavior with a small amount, get into the habit of lot sizing and stop-loss placement, and only then consider your capital size. If you use an offshore broker (like HFM), start with a small testing allocation only after understanding the risks of high leverage.
Risk Disclosure
This page is not investment advice; it is analysis and verification information provided by our lab. Past performance (including backtests and forward tests) does not guarantee future profits. Offshore brokers (such as HFM) carry high-leverage risk; our lab positions them as a small, high-risk testing allocation, with domestic brokers (JFX/OANDA) as the primary focus of our operations. FX and automated trading can result in losses. Always trade with surplus funds, based on your own judgment and at your own responsibility.