Risk Disclosure

1 The headline

1.1 Contracts for difference and other leveraged products are complex instruments and carry a high risk of losing money rapidly. Across regulated brokers the published figure is consistently that the majority of retail CFD accounts lose money — typically between 60% and 80%. Your broker publishes its own number; read it. There is no reason to assume you are the exception, and every losing trader assumed they were.

1.2 You can lose your entire balance. Never fund a live trading account with money you cannot afford to lose completely, including money you need for rent, debt, dependants or an emergency.

1.3 Nothing in Strix changes any of this. Better tooling can make you less wrong more cheaply. It cannot make a losing edge profitable, and software that trades faster on a bad idea simply loses faster.

2 Leverage and margin

2.1 Leverage means you control a position much larger than the cash you put up. A small adverse move in the market therefore produces a large loss on your capital. At 30:1, a 3.3% move against you wipes out the margin behind the position.

2.2 Leverage is symmetrical in arithmetic and asymmetric in practice: the same multiplier that magnifies a gain magnifies the loss that closes your account before the gain has time to arrive.

2.3 Margin calls and forced closure. If your equity falls below your broker’s maintenance requirement, the broker may close positions without consulting you, at whatever price is then available, and at the worst possible moment by construction. Which positions, and in what order, is your broker’s decision and not ours.

2.4 Margin requirements can change without notice — particularly around scheduled economic events, earnings, and weekends — which can force a closure even if the market has not moved.

3 Losing more than you deposited

3.1 Some brokers offer retail clients negative-balance protection, so that the most you can lose is the money in the account. Many do not, and the protection often does not extend to professional clients. Check whether your broker offers it, to your account classification, in your country. Do this before you fund the account, not after a gap.

3.2 Without it, a violent move can leave you owing your broker more than you deposited. That debt is between you and them, and it exists whether the order was placed by your hand or by the software on your instruction.

4 Volatility, gapping and slippage

4.1 Markets do not move in a continuous line. Prices gap — over weekends, over overnight sessions, and instantaneously on news — from one level to a materially different one with nothing traded in between.

4.2 No order can be filled at a price that did not exist. If the market gaps through your level, your order fills at the next available price, which can be far worse.

4.3 Slippage — the difference between the price you expected and the price you got — is normal, is usually against you when it matters most, and grows in exactly the conditions where you most want it not to.

4.4 Volatility can also widen spreads and reduce the size available at each price level, so the cost of getting out rises at precisely the moment you want to get out.

5 A stop is not a guarantee

5.1 Strix can place protective stops with your broker. An ordinary stop is an instruction to trade once a level is reached — it is not a promise of a price. In a gap or a fast market it fills below (or above) the level you set, sometimes by a lot.

5.2 Some brokers offer a guaranteed stop for an extra charge. That is your broker’s product, on your broker’s terms, and it is the only kind of stop that guarantees anything.

5.3 Do not size a position on the assumption that a stop bounds your loss. Size it on the assumption that the stop slips.

6 Overnight financing and holding costs

6.1 A leveraged position held overnight is financed, and you pay for that financing every night you hold it. Over weeks, the charge can consume a profitable position entirely.

6.2 Financing rates move with interest rates and with your broker’s own markup, and can change without notice. On some instruments the charge applies at a triple rate on one day of the week.

6.3 Dividends, index adjustments and contract rollovers each create their own cash adjustments, which may work for or against you depending on your direction.

6.4 Backtests and forecasts inside Strix model costs, but they model them with assumptions. Your broker’s actual charges are the ones that come out of your account.

7 Spreads, commissions and cost drag

7.1 Every trade starts at a loss equal to the spread, plus any commission. The more you trade, the more of your capital that drag consumes.

7.2 An automated strategy can trade far more often than you would by hand. That multiplies the cost drag. A strategy that looks profitable before costs and loses after them is the single most common way a trading idea dies, and it is the failure mode our grading is built to catch.

7.3 Costs at your broker will differ from the ones a backtest assumed. Where they differ upward, the result differs downward.

8 Liquidity and execution risk

8.1 There is no guarantee that a market will be liquid enough for you to open or close a position at a reasonable price, or at all. Thin markets, market-wide halts, circuit breakers at the exchange and instrument suspensions all happen.

8.2 Your order may be rejected, partially filled, or filled at several prices. A partial fill can leave you with exposure you did not intend.

8.3 Trading outside main market hours generally means wider spreads and thinner books, so the same order costs more and moves the price more.

9 Counterparty risk — your broker, not us

9.1 With a CFD you are not buying the underlying asset. You hold a contract with your broker, and its value depends on your broker being able to pay.

9.2 If your broker or exchange becomes insolvent, is hacked, freezes withdrawals, or has its licence withdrawn, your money is exposed to that event. Whether client funds are segregated, and whether any compensation scheme covers you, depends entirely on that firm and that jurisdiction. Find out before you fund the account.

9.3 Strix cannot help you here. We never hold your money, so there is nothing of yours for us to lose — and equally nothing for us to protect or return. We are not a party to your broker agreement, cannot intervene in a dispute inside it, and cannot recover funds from a failed venue on your behalf.

9.4 Your broker may also be your counterparty on the other side of the trade. Its interests and yours are not automatically aligned. That is a feature of the market structure you are entering, and it is worth understanding before you enter it.

10 Currency risk

10.1 If you trade an instrument denominated in a currency other than your account currency, your result depends on the exchange rate as well as on the trade. You can be right about the instrument and still lose.

10.2 Conversions carry their own spread, applied by your broker each time.

11 Technology, connectivity and data risk

11.1 Strix depends on a chain of systems: your device, your network, our hosting, our application, your broker’s API, and the exchange behind it. Any link can fail, and a failure can occur while you have a position open.

11.2 During an outage you may be unable to open, modify or close a position through Strix. Know how to reach your broker directly — its own platform, its app, its phone desk — and be ready to use it. That is the plan, and you should have it before you need it.

11.3 Market data and account information shown in Strix come from upstream sources and may be delayed, incomplete, or wrong. A price on our screen is not a fill. The only price that happened is the one your broker executed.

11.4 There is no uptime guarantee. See the Terms, section 11.

12 Automation risk

12.1 Software can malfunction. An automated trading agent is a program, and programs contain defects. A defect, a bad configuration, a corrupted data feed, or an unhandled broker response can cause the agent to trade when it should not, size a position wrongly, fail to close one, or repeat an action. We test for this and we still cannot promise it will not happen.

12.2 The guardrails are real, and they are not enough. The Service enforces position-size and notional limits, an open-position cap, daily loss and volume caps, a drawdown circuit breaker, an automatic halt after repeated consecutive errors, a circuit breaker on the broker API, an order-rate throttle, and a kill switch. Every one of them acts on data we actually received. If the broker API is unreachable, if prices are stale, if the market gaps, or if the process restarts, a guardrail can act late, act on bad information, or not act at all.

12.3 Automation is your instruction. Orders the agent places are your orders, in your account, at your risk. Turning it on is the decision; each individual order is the consequence of that decision.

12.4 Unattended is not unsupervised. Auto-trading is built to run without you watching every tick, not to run for weeks unwatched. Check it, read the journal, and stop it the moment it does something you cannot explain.

12.5 Stopping always works. The kill switch, disabling auto-trading and returning the Autopilot to paper are never gated by payment, region, or paperwork. If you are unsure, stop. You can always start again.

13 Strategy decay

13.1 A strategy that worked stops working. Edges get crowded out, market regimes change, correlations break, and the behaviour a model learned from history stops describing the present. This is the normal life cycle of a trading strategy, not an anomaly.

13.2 Strix monitors live-eligible strategies for decay and can demote one — removing its live eligibility — while it is running and while you are subscribed. When that happens the strategy keeps running on paper and stops receiving real money.

13.3 That gate is a detector, not a shield. It responds to evidence that has already accumulated, which means the losses that produced the evidence have already happened. A decay gate cannot see a regime change coming; it can only stop you compounding into one.

14 Crypto and funding-carry risk

14.1 Crypto markets trade continuously, are more volatile than most traditional markets, and are thinly regulated or unregulated in many jurisdictions.

14.2 No deposit protection. Assets held at a crypto exchange are generally covered by no deposit-guarantee scheme and no investor-compensation scheme. If the exchange fails, is hacked, halts withdrawals, or is seized, you may lose everything held there. This has repeatedly happened to large, well-known venues.

14.3 Cash-and-carry is market-neutral, not risk-neutral. The carry strategy in Strix holds spot long against a short perpetual future to harvest the funding rate. It is designed to be indifferent to price direction. It is not indifferent to anything else.

14.4 Funding can invert. The funding rate is the whole source of return, and it is not a yield — it is a payment between market participants that can turn negative and stay negative for extended periods. When it does, the position pays out instead of collecting, and the strategy loses money while behaving exactly as designed.

14.5 The legs can come apart. Basis between spot and perpetual can widen sharply. The two legs may sit in different margin systems, so one can be liquidated while the other is fine — turning a neutral book into an outright directional position at the worst moment.

14.6 Execution and capacity. The strategy rebalances across many instruments; each rebalance pays spread and fees, thin instruments slip, and the return is capacity-limited — it does not scale with the size you would like to put behind it.

14.7 Exchange-specific risk. Funding formulas, margin rules, liquidation mechanics, API behaviour and listing decisions are set by the exchange, apply to you, and can change without notice.

14.8 This strategy defaults to paper. Live execution requires keys you supply and is off until you turn it on.

15 Backtests, overfitting, and what a grade means

15.1 Past performance does not indicate future results. Every performance figure anywhere in Strix or on our public site is historical or illustrative. None of it is a forecast, a projection, a target, an offer, or a promise. Backtested and simulated results in particular are produced with the benefit of hindsight and did not risk any money.

15.2 Overfitting is the default outcome. Test enough variations against the same history and some of them will look excellent by chance alone. That is arithmetic, not skill. A backtest that was tuned until it looked good is evidence of tuning, not of edge.

15.3 What our grading tries to do about it. Strategies are evaluated walk-forward — trained on one period, tested on the next, rolling forward — after realistic costs, with held-out data that is not used for fitting, and with statistics that penalise the number of variations tried. This is a serious attempt to falsify a strategy before the market does it for you.

15.4 What a grade is. An A–F grade is our own opinion, produced by our own method, about how a strategy behaved on historical data after modelled costs. That is all it is.

15.5 What a grade is not. It is not a rating issued by a rating agency or a regulator. It is not a prediction. It is not a statement that a strategy will make money, or even that it will not lose money quickly. A live-eligible grade means a strategy survived our gauntlet — nothing further. Strategies that survive still lose.

15.6 A model’s confidence score, an AI debate, a forecast and a journal reflection are all derived from history and from language models. None of them knows what happens next.

16 Our own record

16.1 Our flagship directional model is graded F. A pre-registered study — purged walk-forward cross-validation, deflated Sharpe, locked hold-outs, matched null — found that it does not reliably beat random or buy-and-hold after costs. It is therefore locked out of live trading on our own platform, and paying us does not unlock it. We publish this because it is true and because a company that hid it would be lying to you about everything else too.

16.2 Three strategy candidates have been killed by our own research process — one was a universe-selection artefact, one drowned in execution costs, one never had an edge at all. Two of those write-ups are published on our blog with the numbers.

16.3 The one candidate that passed is the regime-gated crypto cash-and-carry sleeve, and it passed on a backtest. Its best measured out-of-sample result — 38 liquid perpetuals, June 2023 to June 2026, final 40% of the window held out and never fitted, net of 8 basis points of maker cost — was approximately +11.1% per year with a Sharpe of 1.73 and a maximum drawdown of −2.2% on that held-out slice.

16.4 Read that number as a warning, not a target. It is one sleeve, one venue, one three-year window. It is not a live track record, not an average, not a forecast, and not a return you should expect. The sleeve has never traded live money. During our last twelve-day live soak, its funding-regime gate kept it flat and it earned nothing at all.

16.5 If our best measured result would barely cover a subscription on a small account — and it would not, below roughly $3,200 — then the honest conclusion is to stay on the free plan and trade paper. The app tells you that too, and argues against its own upsell when the arithmetic says to.

17 Where this is not allowed

17.1 CFDs and other leveraged retail products are restricted or prohibited in many jurisdictions. CFDs may not be offered to US retail traders, and the United States is on our restricted list by default: a user who declares the United States as their country of residence cannot enable live trading through Strix.

17.2 The restricted list can change, and it is shown to you in the app. Demo and paper trading are never region-gated.

17.3 Whether this trading is lawful for you is your responsibility to establish, and your broker may apply its own restrictions on top of ours. Declaring a country you do not live in to get past the check is a breach of the Terms, section 3.5, and does not make the trading lawful.

18 Tax

18.1 Trading has tax consequences that depend on the instrument, your country and your circumstances. They can be significant, and a tax bill can arrive after the money that would have paid it has been lost.

18.2 We do not give tax advice, do not withhold tax, do not report to any tax authority on your behalf, and do not produce tax statements. The journal export is a record of your activity, not a tax document. Take professional advice where you live.

19 No advice, no guarantee

19.1 Strix does not know your finances, your objectives or your capacity to bear loss, and it performs no suitability or appropriateness assessment. Nothing it produces is investment advice or a personal recommendation, and nothing on our site or blog is a solicitation to trade.

19.2 The operator of Strix — NM Technology Services & Solutions (trading as NMTSS), NUIT 112059725, a business registered in Mozambique as an empresário individual (individual entrepreneur): an unincorporated business form, not a company, and a business-registration category that has nothing to do with trading or dealing in financial instruments — is not a broker, dealer, exchange, investment adviser, portfolio manager or fund, holds no licence, authorisation or registration from any financial services regulator in any jurisdiction, and does not claim one. Its Mozambican alvará (4170/07/01/PS/2020) is a general commercial licence, not a financial-services authorisation. No compensation scheme, ombudsman service or conduct-of-business protection applies to your relationship with us.

19.3 No result is guaranteed. Not by a grade, not by a guardrail, not by a circuit breaker, not by a backtest, and not by anything we have ever published.

19.4 If you are unsure whether leveraged trading is appropriate for you, consult a licensed financial adviser in your own jurisdiction before you fund a live account. That advice will cost less than the account will.

20 What to actually do about all this

20.1 Stay in demo. Everything in Strix works on paper with simulated money, free, forever. Stay there until you have months of your own evidence — not ours.

20.2 When you do go live, start smaller than feels worth it. The first live position exists to teach you how differently you behave with real money, and it will.

20.3 Fund it only with money whose complete loss would change nothing important in your life.

20.4 Read your broker’s own risk warning and its published percentage of losing accounts. It is describing its own customers, which is a more relevant sample than anything we can show you.

20.5 Keep the kill switch within reach and use it without hesitating. Stopping is always free, always allowed, and never the wrong call when you do not understand what is happening.

20.6 By accepting this document in the app, you confirm that you have read and understood it. Questions before you do: [email protected].

The other two documents: Terms of Service Privacy Policy