Risk and Methodology Disclosures

VERSION 1.2 · EFFECTIVE SEPTEMBER 8, 2026 · DRAFT PENDING COUNSEL REVIEW · QUANTGPT LLC, A FLORIDA LIMITED LIABILITY COMPANY

• 1. Not investment advice
• 2. Risk of loss
• 3. Hypothetical performance
• 4. How our backtests work
• 5. What backtests leave out
• 6. Data sources and attribution
• 7. Crypto, prediction markets, sports
• 8. AI-generated text
• 9. Community content
• 10. Affiliate and advertising disclosure

// 1. NOT INVESTMENT ADVICE

QuantGPT is research and education software. Nothing on this site or in the Service is a recommendation to buy, sell, or hold any security, contract, token, or other asset, or to follow any strategy, and nothing is tailored to you. We do not know your finances, goals, tax position, or tolerance for loss. We are not a registered investment adviser, broker-dealer, commodity trading advisor, exchange, or bank, and we do not hold ourselves out as one. No fiduciary relationship exists between you and us. Before acting on anything you learn here, consider getting advice from a licensed professional who knows your circumstances.

// 2. RISK OF LOSS

Trading and investing in stocks, options, futures, crypto assets, event contracts, and any other market involve a substantial risk of loss. You can lose some or all of the money you put in, and with leverage or short positions you can lose more than you put in. Prices can move sharply and unexpectedly. Liquidity can vanish when you need it. Past performance, whether real or simulated, does not indicate or guarantee future results. Only risk money you can afford to lose.

// 3. HYPOTHETICAL PERFORMANCE

HYPOTHETICAL PERFORMANCE RESULTS HAVE MANY INHERENT LIMITATIONS, SOME OF WHICH ARE DESCRIBED BELOW. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. IN FACT, THERE ARE FREQUENTLY SHARP DIFFERENCES BETWEEN HYPOTHETICAL PERFORMANCE RESULTS AND THE ACTUAL RESULTS SUBSEQUENTLY ACHIEVED BY ANY PARTICULAR TRADING PROGRAM. ONE OF THE LIMITATIONS OF HYPOTHETICAL PERFORMANCE RESULTS IS THAT THEY ARE GENERALLY PREPARED WITH THE BENEFIT OF HINDSIGHT. IN ADDITION, HYPOTHETICAL TRADING DOES NOT INVOLVE FINANCIAL RISK, AND NO HYPOTHETICAL TRADING RECORD CAN COMPLETELY ACCOUNT FOR THE IMPACT OF FINANCIAL RISK IN ACTUAL TRADING. FOR EXAMPLE, THE ABILITY TO WITHSTAND LOSSES OR TO ADHERE TO A PARTICULAR TRADING PROGRAM IN SPITE OF TRADING LOSSES ARE MATERIAL POINTS WHICH CAN ALSO ADVERSELY AFFECT ACTUAL TRADING RESULTS. THERE ARE NUMEROUS OTHER FACTORS RELATED TO THE MARKETS IN GENERAL OR TO THE IMPLEMENTATION OF ANY SPECIFIC TRADING PROGRAM WHICH CANNOT BE FULLY ACCOUNTED FOR IN THE PREPARATION OF HYPOTHETICAL PERFORMANCE RESULTS AND ALL OF WHICH CAN ADVERSELY AFFECT ACTUAL TRADING RESULTS.

Every backtest, study, screen, library entry, and animated illustration in the Service is hypothetical. The animation on our home page is a simulated, parameterized sequence for illustration. It is not a record of actual trading.

// 4. HOW OUR BACKTESTS WORK

We publish our conventions so results can be understood and challenged. Where a specific report states different assumptions, the report governs.

• Survivorship-free universe. Our equity history includes companies that were later delisted, acquired, or went bankrupt. A stock that disappears mid-holding contributes zero return for the missing months rather than being quietly dropped. Most retail tools omit dead companies, which inflates every strategy.
• Point-in-time fundamentals. Financial statement data is joined by filing date, so a strategy only "knows" a number after it was public. Restatements are ignored in favor of what was first reported.
• No look-ahead. Signals are computed from data available at the decision date, positions lag signals by one period, and all filters are applied at formation time only. We removed an earlier full-window liquidity filter because it silently excluded stocks that later collapsed.
• Prices and rebalancing. Equity studies use month-end adjusted closes, equal weighting, and monthly or quarterly rebalancing with true buy-and-hold weights inside each holding period. Benchmarks are total-return indices where available.
• Guards against artifacts. Momentum ranks exclude names up more than 150% over the lookback by default, because those are usually untradeable micro-cap spikes. Fundamental ranks only consider companies with at least $10 million of trailing revenue. Monthly stock returns are clipped to guard against bad prints. Each guard is adjustable and disclosed in the report when applied.
• Out-of-sample framing. Where a strategy uses fixed rules, we report the 2015-onward segment separately as pseudo out-of-sample. This is weaker than a true holdout, because the rules were chosen with knowledge of that period's existence.
• Gross returns. Results are before commissions, spreads, slippage, market impact, borrowing costs, financing, and taxes, unless a page says otherwise.
• Losing periods are shown. Every report shows yearly results including losing years and the maximum drawdown. We do not hide bad periods.

// 5. WHAT BACKTESTS LEAVE OUT

• Costs. Real trading pays commissions, spreads, and slippage. For strategies with high turnover, costs can turn a positive backtest negative.
• Capacity and liquidity. A backtest assumes you could have bought and sold at the recorded price in any size. Small stocks cannot absorb large orders, and the price you get moves against you as you trade.
• Execution. Month-end closes are a convention; you may not be able to trade exactly there. Halts, gaps, and stale quotes are not modelled.
• Data errors. Historical data contains errors, missing values, and revisions. We guard against the worst, but we cannot guarantee any series.
• Data mining and overfitting. With enough tries, some strategy will look great by chance. Strategies discovered by searching the same history they are tested on are especially suspect. A strategy library of many entries guarantees that some entries look good for no reason.
• Regime change. Markets, rules, taxes, and participants change. Something that worked for 50 years can stop working, and often does once it is widely known.
• Behavior. A backtest never panics, never skips a rebalance, and never doubles down. You might.
• Taxes. Turnover creates taxable events that a gross backtest ignores.

// 6. DATA SOURCES AND ATTRIBUTION

Market data in the Service is licensed from third-party providers and remains their property. Current providers: [MASSIVE (equities pricing, reference, corporate actions, fundamentals), once the commercial license is executed; EODHD or other providers if used; FINRA (short volume, public); FRED, Federal Reserve Bank of St. Louis (macro, public); SEC EDGAR (filings and insider transactions, public); football-data.co.uk (historical match odds); Polymarket and Kalshi public APIs (resolved event contracts); crypto exchange public data APIs.] Providers are not responsible for the Service and do not endorse it. Data is provided as-is, may contain errors, and may be delayed. You may not redistribute it. See Section 8 of the Terms.

// 7. CRYPTO, PREDICTION MARKETS, SPORTS

The Service includes historical datasets for crypto assets, resolved prediction-market contracts, and sporting events with closing odds, so that the same honest historical analysis can be run on those markets. Additional cautions apply.

We are not a venue

We do not operate an exchange, prediction market, or sportsbook, and we are not affiliated with any. We do not accept, place, route, or facilitate wagers, trades, or contracts. Nothing in the Service directs you to any venue.

Legality depends on where you are

Crypto trading, event contracts, and sports wagering are restricted or illegal in many jurisdictions, including a number of U.S. states, and the rules change often. The availability of a dataset in the Service says nothing about whether participation is lawful for you. You are responsible for knowing and following the law where you live.

Historical odds are not predictions

A strategy that looked profitable against historical odds or contract prices may lose money in the future, and the entire amount staked can be lost. Bookmakers and market makers adjust to profitable patterns. Closing odds already embed most public information.

Age

You must be 18 to use the Service, and 21 wherever that is the legal age for wagering, to use these datasets for any wagering-related purpose.

If gambling is a problem

In the United States, call or text 1-800-GAMBLER, or visit ncpgambling.org. In the United Kingdom, GamCare is at gamcare.org.uk or 0808 8020 133. Elsewhere, your national helpline. Please reach out if you need to.

// 8. AI-GENERATED TEXT

Parts of the Service may use large language models to turn a plain-English request into a study definition, to summarize results, or to explain a concept. Language models can be wrong, can be confidently wrong, and can misstate numbers. The numbers in a report are computed by our engine, not by a language model; the prose around them may be model-generated. Treat any explanation as a starting point to verify, never as a fact to act on. Where you connect your own AI assistant through the plug-in, its output is produced by your provider, not by us.

// 9. COMMUNITY CONTENT

Forum posts, shared reports, performance cards, and directory listings are created by users. We do not verify them unless a page says the figure is read directly from a linked brokerage account, and even then the figure reflects one account over one period and is not typical, not audited, and not a prediction. Other users are not advisers. Results shown by any user are not representative of what you should expect.

// 10. AFFILIATE AND ADVERTISING DISCLOSURE

Some links in the Service, especially in the Marketplace, are affiliate links. If you sign up or buy through one, we may receive a commission at no additional cost to you. Affiliate links are marked, and we disclose this in accordance with the U.S. Federal Trade Commission's Endorsement Guides. Our editorial picks are chosen on merit and are not paid placements; if we ever accept payment for placement, it will be labeled as paid. We are not responsible for third-party products or services, and a listing is not an endorsement of their suitability for you.