The headline statistic is familiar: a majority of retail accounts lose money trading contracts for difference. The explanation is usually reduced to an accusation that traders are undisciplined or a slogan that leverage is dangerous. Neither is sufficient. CFDs create a specific economic hurdle: a trader must correctly anticipate short-term market moves, overcome spread and execution costs, cover financing charges, avoid forced liquidation, and do all that repeatedly while the platform’s risk controls—not the trader’s preferred holding period—can determine when positions close.
New Australian evidence makes the problem measurable. In consultation material released on October 12, 2026, ASIC reports 68% of retail CFD clients made a net loss in the 2023–24 financial year, with aggregate net losses exceeding A$458 million, including A$73 million in fees. The European securities regulator’s earlier cross-jurisdiction studies found 74–89% of retail accounts typically lost money. Those statistics are from different regions and periods and cannot be averaged into a universal worldwide failure rate. But they are persistent enough to demand a structural explanation. [1][2]
First, Understand What the Loss Warning Actually Counts
UK FCA rules require providers to publish an up-to-date percentage of the firm’s retail client accounts with negative trading results over the relevant preceding 12 months, recalculated every three months. The calculation includes fees and charges, as well as realized and unrealized outcomes within the defined methodology. It is a share of accounts at that firm, not a prediction about the chance of losing on the next trade, and not a representative sample of every CFD trader worldwide. Two accounts held by one person can complicate interpretation of account-based figures. [3]
It follows that a displayed 78% does NOT mean 78% of individual positions lose. Nor does it reveal whether the typical losing account lost $50 or $50,000. The number says little about holding periods, account openings, deposit behavior, education or trader selection. ASIC’s recent evidence that 68% of Australian retail customers lost in its specified fiscal year cannot be compared directly with a broker’s current rolling 12-month warning as if they were the same population. What survives the methodological differences is the broad imbalance of outcomes.
FIGURE 1 | Original visualization; sources and scenario assumptions in article and methodology.
ASIC’s A$73 million fee figure implies fees represented about 15.9% of the A$458 million reported net loss magnitude, if the disclosed totals are compared mechanically. That does not mean fee-free trading would make all of those accounts profitable: fees and trading losses interact, and net totals are affected by gains and losses across accounts. Nor is the A$458 million necessarily a measure of total gross losing-trader losses, because profitable accounts offset some outcomes. Still, the fee share is a useful signal that the economic friction is material, not a footnote. [1]
CFDs Are Negative-Sum After Costs Even Before Skill Is Considered
Imagine a hypothetical EUR/USD CFD trader whose underlying trade, measured at mid-market, would break even on average before costs. Add the quoted spread, commissions, execution slippage and any financing. Expected net outcome becomes negative unless the trader consistently adds a sufficiently positive forecasting or execution edge. That arithmetic is true regardless of whether the broker hedges the client’s risk. ‘Zero commission’ does not make the all-in trading price zero. The dealer may earn through the spread and other charges, or pass prices and charge a commission.
For a transparent one-standard-lot EUR/USD example with $10 per pip assumed, suppose the full round-trip friction is 0.8 pip in spread ($8), $7 in commission, 0.5 pip in execution slippage ($5), and $3 in overnight finance. The total is $23 for that hypothetical trade. If a trader executes 100 such trades without a genuine gross edge, cumulative friction is $2,300; at 250 trades it becomes $5,750. These assumptions are illustrative and cannot be taken as a typical broker quote or a prediction of what any specific account will pay. Each provider’s schedule and market conditions differ.
FIGURE 2 | Original visualization; sources and scenario assumptions in article and methodology.
The critical insight is the compounding of repeated decisions and costs. A strategy might genuinely predict direction slightly better than random but still lose after friction. A trader who changes strategies weekly, enters and exits around noise or uses stop distances smaller than spread plus slippage faces a proportionally higher cost barrier. The quantity of trades can transform a modest per-trade cost into a persistent annual handicap. The FCA’s 2025 fair-value review specifically questioned inconsistent or inadequately explained overnight funding charges at some providers. [4]
The Win-Rate Trap: You Can Be Right Often and Still Lose
Expected trading value depends on both frequency and payoff. In a clean stylized model, let a winning trade make $80, a losing trade lose $80, and each trade incur $12 of all-in costs. A 45% gross win rate gives 0.45 × $80 – 0.55 × $80 – $12 = -$20 expected per trade. Even a 55% gross win rate produces $44 – $36 – $12 = -$4. Only at 57.5% does this symmetric-payoff model reach zero expected net profit. The required percentage is calculated by solving p × 80 – (1-p) × 80 – 12 = 0.
FIGURE 3 | Original visualization; sources and scenario assumptions in article and methodology.
This is not proof that 57.5% is some universal market threshold. With average winners twice as large as losers, break-even win rate can be much lower; with frequent small profits and occasional outsized losses, a 70% success rate may be worthless. The trap is selecting a system by its percentage of winning trades while ignoring average win, average loss and transaction costs. A profitable backtest that omits variable spreads, rejected orders, execution conditions and overnight financing is not a realistic live-trading test.
The market rarely provides independent identical bets. Volatility clusters, correlations rise in crises, and strategies can fail together. A day trader might have hundreds of outcomes that are correlated through one market regime. Therefore a neat distribution of win rates can overstate statistical confidence. When someone reports a high success percentage, ask for net results across all trades, real bid/ask fills, holding costs, the longest losing sequence and risk-adjusted performance—not screenshots of individual winners.
Leverage Changes Survival Time More Than Forecast Accuracy
Leverage does not itself alter the underlying price movement. It changes how much of a market move the account can absorb. A $2,000 deposit supporting $20,000 exposure represents tenfold gross exposure. A 5% adverse market move produces roughly $1,000 negative mark-to-market P&L, before fees, erasing half the deposit. At 20x exposure on the same deposit, the same 5% move generates approximately $2,000 gross loss, effectively consuming it. Actual margin close-out can occur sooner and depends on collateral, instruments, tiering, cross positions and regulation.
The important mechanism is path dependency. An unleveraged investor who has the capital to hold can remain in a volatile market through a temporary setback. A leveraged trader may be closed during the setback before the forecast direction materializes. The fact that the trader was ‘eventually right’ is immaterial if the broker liquidated the position first. Stop losses do not fully solve this: slippage, price gaps and differing trigger/execution prices can cause losses beyond a designed stop level.
The EU’s ESMA measures restricted retail CFD leverage by asset class, historically ranging from 30:1 on major currency pairs to 2:1 on crypto, alongside margin close-out at a defined account level, negative balance protection and standardized warnings. The UK’s FCA has corresponding domestic rules. ASIC’s 2026 extension proposal specifically notes that its existing intervention reduced quarterly aggregate client net losses by 91% in the first six months. That result reflects changes in the market and participation as well as restrictions: it should not be interpreted as a controlled causal estimate that leverage limits alone account for every percentage point. [5][6]
Margin Calls, Spread Widening and the Wrong Exit Price
Unlike an idealized textbook price chart, an executable CFD market has two sides of a quote. Long positions typically open at the offer and close at the bid; shorts face the inverse. When liquidity thins, spreads can widen just as a position approaches its stop or maintenance margin. A narrow expected move can disappear inside the bid/ask gap. Overnight, at market openings, during data releases or during asset-specific shocks, prices can jump between executable levels. The trader sees a single directional idea, while the actual result depends on several execution layers.
Broker structure matters as well. Some dealers hedge exposure externally, some internalize client flow, and some use hybrids. That determines how price risk is managed on the broker’s books; it does not automatically prove manipulation or eliminate client costs. Dave Finances’ analysis of how forex brokers hedge client trades separates order execution and counterparty risk from simplified stories about an exchange ‘taking the opposite side’ of every position. A trader’s operational diligence should include the actual legal counterparty, the execution policy and the handling of abnormal market conditions.
A useful stress example starts with an account worth $5,000 and a $50,000 notional CFD position. A sudden 4% adverse gap means roughly $2,000 loss; a 12% gap means roughly $6,000 loss before fees. Negative-balance protection in applicable retail regimes generally limits the customer’s obligation to the amount in the relevant CFD account, but does not preserve the original capital. Offshore or professional-client structures can have materially different terms. Dave Finances’ offshore forex broker risk analysis explains why jurisdiction and entity-level regulation change the protection perimeter.
Behavioral Bias Is a Mechanism, Not a Complete Explanation
The design of leveraged trading makes behavioral errors expensive. Daily P&L displays, instant execution, low minimum deposits, social leaderboards and positions that feel large relative to cash deposited can push a user toward excessive frequency and short holding periods. Loss aversion may lead to taking profits early while allowing losing positions to widen; overconfidence after a run of wins can multiply notional size. These are possible mechanisms, not a diagnosis of every losing trader. Regulator data alone does not establish their causal contribution to losses.
A trader using a stop on every position can still accumulate a series of small losses that outruns intermittent gains. Another trader could have a sound forecast but hold exposure too large for its natural daily volatility. Traders can also misread a CFD as a long-term investment substitute while ignoring daily or overnight funding. The industry often sells access in terms of tiny initial margin requirements; the economic risk should instead be measured on notional exposure and plausible adverse moves.
Professional-looking software can conceal this difference. A €300 margin requirement on a €10,000 position does not mean the trade risks only €300; it means the investor is controlling a larger notional exposure against limited collateral. Similarly, a broker’s ‘margin close-out’ is a mechanism for restricting further deterioration, not a guaranteed personalized stop-loss at the preferred price. A sensible pre-trade plan specifies notional, spread assumptions, stop execution risk, fees, time horizon, and total account loss tolerance.
What the Latest Regulatory Data Actually Suggests
Australia’s newest public figures help separate claims. ASIC’s October 2026 consultation reports that 68% of retail CFD clients lost in 2023–24 and identifies A$458 million of aggregate net losses, including A$73 million in fees. It separately reports that 85% of retail clients trading CFDs over options made a net loss. It proposes to extend and amend product intervention for ten more years, while consultation remains open until November 9, 2026. These are regulator disclosures, not platform marketing data. [1]
The FCA’s October 2025 statement estimated that existing protections benefited nearly 400,000 consumers annually and provided £267 million to £451 million worth of protection. Those figures are a regulatory impact estimate, not cash compensation distributed to customers. A later FCA review criticized cases where providers had not adequately explained overnight charges or demonstrated fair value. Viewed together, the evidence points to a combination of market risk, costly implementation and product-governance weaknesses, not a single psychological failing. [4][7]
US CFTC consumer material on off-exchange forex provides a related, though not interchangeable, lens: it says approximately two in three retail OTC forex customers lose over a period and explicitly highlights dealer execution, spread and fee incentives. OTC forex and CFDs have different legal definitions across jurisdictions. The cross-market lesson is the importance of measuring net customer outcomes after the full cost stack, not extrapolating a single firm’s promotional return history. [8]
The Trader’s Full-Cost Checklist
| Question | What to calculate | Typical oversight |
| Exposure | Notional / total account equity | Margin confused with max loss |
| Entry and exit | Bid/ask spread + commission | Using mid-market chart fills |
| Execution risk | Variable slippage and gaps | Fixed-stop guarantee assumed |
| Carrying cost | Overnight financing × exposure × days | Ignoring holding-period drag |
| Edge quality | Win rate × gain − loss rate × loss − costs | Win percentage reported alone |
| Protection | Entity, regulator, client classification | Assuming all platforms identical |
TABLE 1 | Analytical framework compiled for this article.
This checklist exposes why ‘I only need to be right about direction’ is incomplete. An edge smaller than normal execution friction will not survive repeated trades. An account that ignores financing can turn a correct medium-term thesis into a negative return. And a stop that depends on filling at a price unavailable during a gap is not a reliable worst-case loss. Trading records should distinguish price-based gross P&L from all charges and cash flows. A strategy that appears profitable before these adjustments needs to be treated as unproven.
What Would Challenge the Thesis That Most Traders Are Structurally Disadvantaged?
The strongest counterexample would be independently verified net profitability over a long multi-regime sample, calculated after spreads, commissions, financing and slippage, with a representative population of surviving and exited accounts. Evidence that profits persist once trades are marked at executable prices and that returns are not concentrated in a tiny, highly leveraged tail would challenge sweeping claims about the category. Some traders may indeed have repeatable edges or use CFDs effectively for short-term hedges. The loss statistic does not prove universal incapacity.
Yet accounts that survive because their owners add more capital need separate interpretation from accounts that achieve positive trading P&L. So do professional customers, who may have wider portfolio hedges and different protections. A careful investigation avoids survivorship bias, does not select a few star accounts from social media, and does not equate more users with more successful investors. Dave Finances’ guide to European crypto perpetual regulation further shows why legal classification matters when comparing similarly leveraged products offered under different frameworks.
Bottom Line: The Price Prediction Is Only One Input
A CFD trader can lose despite forecasting direction correctly on many occasions. Too much leverage reduces the time available for a view to work; costs create a negative baseline; margin and gap risk make exits asymmetric; and high turnover compounds seemingly small frictions. The newest regulator evidence supports a serious, continuing retail-harm problem, while also showing how protections reduced aggregate losses in a specific observed setting. It does not establish that every trader loses or that leverage alone causes all losses.
For anyone considering a CFD account, the most useful question is not ‘How much can I control with $1,000?’ but ‘What is my expected net edge after realistic execution, how much capital can a gap consume, and what does the account contract permit the broker to do in a crisis?’ Without those answers, the numerical odds start against the trader before a market prediction is even made.
Methodology
Regulatory figures are from ASIC’s October 12, 2026 consultation and prior sector review, FCA rules and policy communications, and historical ESMA intervention analysis. Different populations and years are labeled rather than pooled. All example spreads, commissions, slippage, finance, win rates and trade outcomes are invented solely as transparent scenarios, not market observations. A$73m divided by A$458m is a rough aggregate ratio, not the fraction of each individual loss caused by fees. Exposure examples exclude volatility-dependent margin tiers and precise broker close-out algorithms. Research checked October 13, 2026.
Sources & primary documentation
Johan Shamshad is a financial markets writer at Dave Finances covering cryptocurrencies, trading platforms, brokers, fintech, financial regulation, and developments across global markets. He previously worked at Gulf News, adding newsroom experience to his coverage of fast-moving financial and digital-asset markets.
His work focuses on identifying market-moving events, company developments, regulatory changes, product launches, and shifts in trading and financial infrastructure.
Johan contributes news and analysis designed to help readers understand not only what happened, but why a development matters and how it may affect the wider financial landscape. You can reach out to him via his social media accounts:
Linkedin: https://www.linkedin.com/in/johan-shamshad-742851262/
X: https://x.com/Yasmine_FX
Investing: https://www.investing.com/members/contributors/279781574