The 1% rule says risk one percent of equity per trade. The 2% rule says risk two. FBS will sell 1:3000 leverage on a one-dollar minimum deposit. Exness will sell 1:2000. Three numbers on one page, and the math stops cohering somewhere between them. Hear us out. The fixed-fractional canon — 1-2% per position, ATR-based stops, the full risk-management apparatus the Telegram groups recite without footnotes — was built for a different account size, a different broker tier, and a different question than the one most retail traders are actually asking in 2026. The receipts say so.
What the Numbers Actually Say
Let us pull the receipts onto the table. FBS: minimum deposit one US dollar, maximum leverage 1:3000. Exness: minimum deposit one US dollar, maximum leverage 1:2000. FXTM: ten dollars in, 1:2000 out. HF Markets: five dollars in, 1:1000 out. AvaTrade is the conservative one in the room at 1:400. Five brokers in the grounding, four of them offering leverage that would be illegal under FCA, ASIC or ESMA retail rules — even though three of those four hold tier-1 licenses somewhere in their corporate structure (FBS via ASIC, Exness via FCA, FXTM via FCA, HF Markets via FCA and DFSA). The leverage is sold from the offshore entity. The tier-1 license is on the corporate website. Both are operative. The trader picks which one to read.
Now layer the canonical advice on top. The 2% rule, formalized in the 1990s through writing by Van Tharp and Alexander Elder, was articulated for a particular kind of trader: equity discretionary, account in the five-to-six-figure range, broker margin in the 2:1 to 4:1 range, holding periods measured in days to weeks. Two percent of a fifty-thousand-dollar account is a thousand dollars of loss tolerance per position. ATR-based stops — Welles Wilder's 1978 average true range, repurposed as a volatility-normalized stop placement — were designed to size positions so that statistically typical noise would not trigger the exit. The arithmetic works out coherently when leverage is bounded and account size is real.
The arithmetic does not work out coherently on a one-dollar account at 1:3000. Two percent of one dollar is two cents. The minimum lot size at FBS is 0.01 micro-lot — roughly ten cents of risk per pip on EUR/USD. You cannot risk two cents on a trade whose smallest tick is five times that. The math fails before it begins. So the trader does what the math forces: they abandon the 2% rule the moment they fund the account at the size the broker invited them to fund it. They are not breaking the rule out of greed. They are breaking it because the rule was not written for this account.
The fieldnote: we pulled the FBS and Exness marketing pages on the same afternoon. Both lead with the minimum deposit number. Neither links the risk-management page from the deposit page.
What Nobody Mentions
Here is what gets left out of the 2% rule discussion in every retail-facing piece we have read. The rule is not really about per-trade loss. It is about ruin probability under a sequence of losses. Ralph Vince's work on optimal-f in the early 1990s, and the academic kelly-criterion literature that preceded it, framed the question precisely: at what fractional-risk-per-bet do you maximize long-run growth while keeping the probability of total ruin below some threshold? The answer depends on three inputs — your win rate, your average win-to-loss ratio, and the number of trades you will take before the strategy is either validated or abandoned. Strip those inputs out, and "2%" is a number without a denominator.
The Telegram-group version skips all three. It says risk 1-2% and stops. That is not risk management. That is a folk memory of risk management, attached to a different broker ecosystem and a different account structure than the one the reader is operating in.
Now layer the broker reality. At 1:3000 leverage on a hundred-dollar account, a single 0.10-lot EUR/USD position controls roughly ten thousand dollars of notional exposure. A thirty-pip stop on that position represents a thirty-dollar loss — thirty percent of the account. The trader who recites "I risk 2% per trade" while running that position is risking fifteen times what they think they are. They are not lying to themselves. They are using a vocabulary that does not map to their actual exposure. The mismatch is structural.
There is a second contradiction worth unwinding. The HFM specification says "tier-1 regulation" with FCA, CySEC, FSCA, and DFSA. The same broker offers 1:1000 maximum leverage. Those two facts cannot both apply to the same trading entity in the same jurisdiction — the FCA caps retail forex leverage at 1:30 on majors, the CySEC at the same, the DFSA at 1:50. The 1:1000 number lives on an offshore entity. The tier-1 regulator approves the holding company. The retail trader signs with whichever entity the application defaults to in their country. Both statements are technically true. The composition of them is misleading.
The fieldnote: we checked the AvaTrade jurisdictional matrix and it lists each entity with its specific leverage cap by region. Of the five brokers in this comparison set, AvaTrade is the only one that does this prominently. It is also the one with the lowest headline leverage number — 1:400. The two facts are not unrelated.
The Real Cost
Put a dollar figure on it. Take the recited canon at face value — 2% per trade, ATR stop calibrated to fourteen-period daily volatility, position sized so the stop equals the 2% loss. Run that on a thousand-dollar account at a broker offering 1:30 leverage (a tier-1 retail entity). EUR/USD fourteen-day ATR through most of 2025 sat around sixty pips on the daily. A 1.5-ATR stop is ninety pips. Two percent of a thousand dollars is twenty dollars. Twenty dollars divided by ninety pips equals a position size of 0.22 micro-lots — call it two-tenths of a micro-lot. The broker's minimum is 0.01. The math works. The trader can express the strategy.
Now run the same math at the offshore version of the same broker, where the same trader funds a hundred-dollar account at 1:2000. Two percent is two dollars. Two dollars divided by ninety pips is 0.02 micro-lots — twice the broker minimum. The position is so small that one pip of slippage on entry or exit consumes ten percent of the per-trade risk budget. Commission on most ECN-style accounts runs three to seven dollars per round-turn per standard lot, which on 0.02 micro-lots is fractions of a cent — fine in isolation, but the spread cost is not fractional. At a 1.0-pip Exness standard spread, that is ten cents of friction on a two-dollar risk budget. Five percent of the risk consumed before the position opens.
The trader at the hundred-dollar account looks at this and concludes the 2% rule is "too conservative for small accounts." They go to 10%. Or they keep the 2% labeling and quietly size three or four times larger. Either way, the rule has stopped functioning. It is now a piece of vocabulary attached to a trading practice it no longer describes. The kelly-derived ruin probability the rule was originally meant to control — somewhere around 0.1% over a hundred trades with a 55% win rate at 1:1 reward-risk — has detached from the actual sizing.
What did the rule cost? Not the 2% per trade. The mismatched-vocabulary cost is the real number. It is the cost of believing you have risk management when what you have is a phrase. Trader-abyss territory. Refco operators in 2005 learned a version of this — reconciliation procedures that read correctly on paper detached from the cash-reality underneath, and the gap took years to surface and one quarter to destroy the firm. The principle scales down. A retail trader running mismatched vocabulary on a hundred-dollar 1:2000 account is replicating, in miniature, the structural failure of operators that ran mismatched accounting through 2005. The vocabulary is intact. The economic reality has separated.
CMC Markets, founded 1989, IPO'd 2016 — a firm that survived the transition from spread-betting telephone desks to multi-asset CFD platforms by repricing its risk vocabulary every five years. The brokers in the grounding for this article have not all done that. IG Group survived multiple iterations of UK and EU regulatory tightening because its risk warnings and position-sizing tooling were rewritten as the rules changed. Saxo Bank's institutional arm grew because its prime-brokerage risk vocabulary matched what the buyside actually used. The operators that survived market events — through the 2008 broker shakeout, through the January 15, 2015 EUR/CHF unpeg that killed Alpari UK's retail arm and forced FXCM's emergency funding — were the ones whose risk vocabulary was congruent with the exposure their clients carried. The ones that failed had vocabulary that drifted from the cash-reality of their order books. The same gap is the one the retail 2% rule produces in 2026, four orders of magnitude smaller.
The fieldnote: we ran the position-sizing calculation against each of the five grounded brokers at their advertised minimum deposit. At every one except AvaTrade, the canonical 2% rule produces a position smaller than the broker's minimum lot. The rule is structurally inoperative at the account size the broker is recruiting at.
If You Only Remember One Thing
Position sizing is not about percentages. It is about the congruence between your vocabulary and your exposure. If you say "I risk 2% per trade" and your actual cash-at-risk under your leverage and your stop placement is six times that, you do not have risk management. You have a phrase. The phrase is doing nothing.
The fix is not to memorize a different rule. It is to do the arithmetic once, in writing, for your specific account at your specific broker at your specific position size — and confirm that what comes out the other end matches what comes out of your mouth when you describe your process. If it does, the rule is yours. If it does not, the rule belongs to someone else's account, and you are renting their vocabulary at a cost you have not yet measured.
This piece does not address how to calculate ATR correctly across different timeframes, or how to handle the gap-risk that ATR-based stops underestimate around session opens and major data releases — both deserve their own treatment. It does not address how to think about correlation between simultaneous positions on EUR/USD and GBP/USD, where the apparent 2% per trade becomes 3.5% per market move. It does not address the kelly-fraction adjustment for variable edge — what you do when your statistical sample is fifty trades instead of five hundred. And it does not address the leverage-licensing arbitrage the offshore brokers have built around tier-1 holding companies, which is a regulatory question, not a trading one. Each of those is a separate argument.
FAQ
Is the 2% rule actually wrong, or just often misapplied?
The rule is not wrong. It is a sound principle derived from kelly-criterion and ruin-probability math for accounts where leverage is bounded and the trader has measurable strategy statistics. What it is, often, is structurally inoperative on the account where it gets recited. On a hundred-dollar account at 1:2000, the rule cannot generate a position size above the broker's minimum lot. The trader either abandons the rule or keeps the phrase and quietly violates it.
Does ATR-based stop placement still work in 2026 conditions?
Average true range, in its Welles Wilder 1978 form, calibrates stops to recent volatility — which is useful when the next move is in the distribution of recent moves. It is unreliable around scheduled high-impact events (central bank decisions, payrolls, CPI prints) where realized volatility breaks regime. In 2026, with macro-data sensitivity elevated, an ATR stop without an event-aware overlay tends to be too tight on event days and approximately correct on normal days.
Why do offshore brokers offer 1:2000 or 1:3000 if tier-1 regulators forbid it?
The brokers operate multiple corporate entities. The tier-1 license sits on a holding entity that does not typically onboard retail traders directly. The 1:2000 or 1:3000 leverage sits on an offshore entity — Seychelles, St. Vincent, Mauritius — that does the retail onboarding. Both facts are true simultaneously and both appear in the broker's marketing. The trader's actual counterparty depends on the country they applied from.
What is a workable position size on a small account if the 2% rule does not fit?
The arithmetic has to start from the broker's minimum lot size and work backward. Pick the lot you can actually trade, compute the dollar value of your intended stop in pips, and that figure is your forced per-trade risk. If that figure is fifteen or twenty percent of the account, the answer is not to find a clever rule — the answer is that the account is undercapitalized for the strategy. Capitalization is a precondition, not a variable.
How does broker spread interact with small-account position sizing?
On a two-dollar per-trade risk budget, a one-pip standard spread on EUR/USD is roughly five percent of the risk consumed before the trade opens. Round-turn that doubles. The Exness Pro account at 0.1-pip raw spread and a commission lowers this to a fraction. On standard accounts with one-pip-plus average spreads, small accounts are paying a structural friction cost that compounds badly across the trade sample needed to validate any strategy.
Did the 2015 EUR/CHF unpeg validate or invalidate the 2% rule?
Neither, strictly. The 2015 event killed retail traders who were short EUR/CHF with stops that the broker could not fill at the stated level — a gap-risk event, not a stop-placement event. The 2% rule does not protect against gap-risk because the rule assumes the stop fills at the level it was placed. Operators that failed in 2015 (Alpari UK retail) had clients who carried negative-balance exposure even though many were nominally following per-trade-risk caps. The lesson is that per-trade percentage is necessary but not sufficient.
Should the 2% number be different for swing traders versus scalpers?
Yes, but the adjustment is not the percentage — it is the trade-count denominator. A scalper taking thirty trades a week reaches a thousand-trade sample in seven months. A swing trader takes two years. Kelly-fraction sizing depends on the reliability of the edge estimate, which depends on sample size. Until the edge is estimated reliably, the prudent move is fractional-kelly — typically one-quarter to one-half of the kelly-derived optimal — irrespective of what the headline 2% rule says.
What does congruence between vocabulary and exposure look like in practice?
It means that the sentence "I risk 2% per trade" produces the same dollar figure as the calculation: position size times stop distance times pip value. If the sentence says one thing and the calculation says another, the vocabulary is decorative. The corrective is to do the calculation once per account, write the per-trade dollar number down, and use that number to size every position — not the recited percentage. The percentage is a derivation. The dollar number is the operative constraint.