Risk Disclosure

1. Trading futures can lose you more than you put in

Trading futures and options on futures involves substantial risk of loss and is not suitable for everyone. Leverage works in both directions: a small move against you can wipe out your margin, and you can lose more than your initial deposit.

You should only risk capital you can afford to lose entirely, and you should not fund a futures account with money you need for anything else. Most retail participants in these markets lose money.

2. Hypothetical performance

Every performance figure we publish or deliver is hypothetical. Most are produced by applying a rule to historical data. Some, such as the forward record on our Results and Trade Results pages, are recorded as our engine generated them during live sessions: they are still simulated, no order was filled, and the strategies were selected on historical data before the record began. None is a record of trades that happened.

Past performance is not necessarily indicative of future results.

3. What a screen tells you, and what it does not

A screen answers one narrow question: on this historical data, is this rule's record distinguishable from chance, and what would it have produced before your own costs?

It does not answer whether the pattern will exist tomorrow, whether you will execute it the way the simulation did, whether you can follow it through a losing run, or whether it suits your account, your capital and your life.

What each verdict means

DEAD

MeaningThe specification failed at least one test we treat as fatal. We are telling you not to spend more time on it. We are not telling you to do the opposite, which is a different idea that would need its own screen.

INCONCLUSIVE

MeaningThere were not enough trades, or the result was not separable from noise, to say anything either way. It is not a weak yes. It is “we do not know”.

CANDIDATE

MeaningIt survived every test we apply. This is the beginning of diligence, not the end of it. Most things that survive a back-test still fail in live trading.

These definitions apply in the same way to a screen of a Strategy Language specification and to a next screen.

Swing screens

A rule that can hold a position overnight is a swing screen. It is graded rather than judged gate by gate, and it is never inconclusive:

CANDIDATE

MeaningLikely a real edge over holding the market.

PROMISING

MeaningMore likely real than not, not yet enough to call it.

DOUBTFUL

MeaningProbably no edge over holding the market.

DEAD

MeaningNo edge over holding the market, or the causality or cost check failed.

A swing grade comes with a probability: our estimate, from the historical evidence, that the rule's edge over random entries on matched days, held the same way, is real and clears a set hurdle. The evidence is the rule's own trades, the same rule on the other markets it runs on, and what our record of swing screens knew before it ran, and the report says how much of the answer each carried. How likely the edge must be for each grade, and the hurdle it must clear, were set by a calibration study.

The probability is a statement about the historical evidence. It is not the probability that the rule will make money, or that you will. It can be wrong, and because it draws on the record of earlier swing screens, the same rule can get a different probability as that record grows.

Labels in our research

Our research notes label findings, for example “Settled”, “Open question” or “Works with conditions”. The labels describe how strong the historical evidence is. They are not recommendations.

4. The multiple-testing problem

If you test enough ideas, some will look good by chance alone. This is the single biggest reason back-tested strategies fail live, and most vendors do not mention it.

We publish the number of specifications tested through our process and the fraction that passed, and the report of every menu screen shows it beside the verdict. Measured on 28 October 2026: 105,204 specifications screened, of which 562 cleared the full gate stack — roughly one in one hundred and ninety. Most of them were cut by a first, inexpensive pass before the full tests ran; the research ledger at waypointquant.com/research/ledger breaks the count down. That number is the context for any individual pass: a strategy that clears our gates is not certain to be real, it is a survivor of a filter whose historical false-positive behaviour we measure and publish.

Every menu screen ordered through this site is counted in that denominator, including yours. That is deliberate: a pass rate that excluded customer traffic would be a flattering number rather than a true one. Screens written in the Strategy Language are not counted yet; this page will say when they are. A swing screen's report shows how many swing screens have been run; no pass rate is kept for them yet.

5. Limits of our data and our simulation

Anything computed from data inherits the data's faults. Ours include:

• Historical scope. Our tests cover a specific historical window. A rule that works within it may reflect conditions of that period only. Where we can test whether a result survives outside the period it was discovered in, we report that as a label, and many results do not survive it.
• Where the data ends. Our data stops at a fixed date, which can be months before you run a screen. Each market's page says what history it has.
• Contract rolls. Futures histories are stitched together across expiring contracts. Stitching introduces artefacts, and we correct for them, but no correction is perfect.
• Costs. We charge an assumed commission and slippage per trade. Your real costs depend on your broker, your order type, your size and the moment you trade. Real slippage is frequently worse than assumed, especially in fast markets, at the open and around scheduled economic releases.
• Fills. The simulation assumes your order fills at a modelled price. A real stop can gap through its level, and a real limit order can fail to fill on a move you would have wanted. Neither is captured fully.
• Overnight and weekend gaps. A position held overnight or over a weekend can open well beyond its stop. A screen fills that stop at the opening price; a real order can fill worse, and the loss can be larger than the stop you set, and larger than your account.
• Latency. The simulation acts the moment a condition is met. Your software, your connection and your broker take time. We have measured that the value of some strategies decays sharply within minutes of the signal.
• Capacity. Figures are quoted per contract unless a report says otherwise. Trading more contracts moves the market against you, particularly in thin instruments. Size does not scale returns linearly and past some point it destroys them. An estimate of what a rule would make at a larger size is itself hypothetical, built on a model of market impact.
• Survivorship in our own record. Our research record is a record of what we tested. It cannot account for ideas nobody ever thought to test.

6. Regime change

Market relationships stop working. Some of the effects in our own published research did not exist before a certain year, and some stopped after one. A strategy's historical record is not a promise that the conditions producing it still hold.

Any strategy you deploy needs monitoring, and it needs a rule decided in advance for when you will stop trading it.

7. Execution and technology

Automated strategies fail in ways that have nothing to do with the strategy: platform crashes, lost connections, a data feed that freezes, an order rejected by the exchange, a session that opens on the wrong contract month, a setting changed by accident.

You are responsible for monitoring anything that trades your account. Never leave an automated strategy running unattended without position limits and a loss limit you have tested.

8. Funded and prop-firm accounts

Evaluation and funded accounts have their own rules: trailing drawdown that follows your peak and never comes back, daily loss limits, consistency requirements, minimum trading days and payout caps. A strategy that is profitable in a back-test can still fail an evaluation, because these rules charge you for open excursions and for uneven daily results.

Any analysis we provide about such accounts is based on the rule parameters you enter and on historical simulation. The firm's actual rules, and their interpretation of them, control.

9. No advice, no recommendation

Waypoint Quant LLC is not a registered investment adviser, not a registered commodity trading advisor, not a broker, and not a member of the National Futures Association. Nothing we publish or deliver is a recommendation to buy or sell anything, and nothing is tailored to your circumstances. That includes the research desk's replies, which are written by an AI model and can be wrong, and research content that an AI model drafts for us.

We do not know your financial position and we do not want to. Every decision is yours. Consider consulting a licensed professional before trading.

10. Your responsibility

Do your own due diligence on anything you get from us. Verify a strategy yourself before you risk money on it. Start in simulation. Start small. Decide your maximum loss before you begin, not after.