Trading Signals vs Quantitative Research: Are Signals Worth It?

If you trade futures, you get offered both, often under the same name. This guide explains the difference, what the academic evidence says about each, the five kinds of evidence that get mixed up in marketing, and six questions to ask before you trust any of them.

Research and signals side by side

What "quantitative" actually means

A quantitative approach uses fixed rules and historical data instead of in-the-moment judgment. Same inputs, same output, every day. That makes the work testable: you can compare the read published before the open with what the market did after.

It doesn't make the work right. Markets change, patterns fade, and a rule that held for years can stop holding. Good quant work treats every result as a hypothesis that's still being tested.

What the evidence says

Three bodies of research are worth knowing before you pay for anything.

Most day traders lose money. Fernando Chague, Rodrigo De-Losso and Bruno Giovannetti studied everyone who started day trading Brazilian equity futures between 2013 and 2015 and stuck with it for at least 300 days. In their 2019 working paper, 97% lost money, and only 0.4% made more than a bank teller's wage. They found no evidence that traders got better over time.

Technical rules often look better in tests than in real life. Cheol-Ho Park and Scott Irwin reviewed the literature in the Journal of Economic Surveys (2007). Of 95 modern studies, 56 found positive results. But most studies had problems, including data snooping (trying a bunch of rules and keeping the winners) and transaction costs that are hard to estimate.

Costs can wipe out a real edge. Carlo Zarattini and Andrew Aziz's paper Can Day Trading Really Be Profitable? reported strong results for an opening-range breakout on QQQ from 2016 to 2023, after commissions. An independent replication published in September 2026 reproduced the gross result on the Nasdaq-100, then showed it dropped to roughly zero once spread and slippage were charged.

None of this says research is useless. It says the burden of proof is on the method, and that the honest unit of evidence is a result after costs, over a long period, published before anyone knows the outcome.

Five kinds of evidence that aren't the same

Marketing often blurs these together. Keep them apart:

1. A report: a read published before the session, with a time stamp. It shows what was known ahead of time.
2. A replay: the same session looked at afterward. It shows what happened, with hindsight.
3. An illustration: one hand-picked example. It explains a concept. It proves nothing about how often something happens.
4. A backtest: rules applied to past data. It estimates how things behaved historically, before costs and real fills unless stated.
5. Real performance: actual trades in a real account, with fills, fees and slippage.

1. 01Answers

   "What does the evidence favor today?" · "Buy here, sell there"

2. 02Decision

   Yours · The provider's

3. 03Checkable

   Yes, if time-stamped and defined · Only if every signal and fill is logged

4. 04Accounts for your risk

   You size your own risk · Usually one size for everyone

5. 05Typical failure

   A read that is wrong · Losses from size, slippage and timing

A visual summary of the article's table — not a performance chart or a new backtest.

A level reached on a chart isn't a trade. A backtest isn't performance. One good day isn't a win rate.

How to read a backtest in 60 seconds

Before you look at the return, look for six things. If any of them is missing, the number on top doesn't mean much.

A backtest that shows all six, and still holds up after costs, deserves your attention. One that shows a headline return and nothing else is an ad.

Why negative results are a good sign

A provider that only publishes what worked is showing you a selection. One that publishes what failed is showing you a method.

We hold ourselves to that. When we retested two famous calendar patterns on the Nasdaq-100, neither held up: stocks didn't earn more in the even weeks of the Fed cycle from 2017 to 2026 (test), and "Sell in May" didn't work on the index from 2000 to 2026 (test). Both results are published with their data, method and limits.

If you still want signals, ask for this

Some traders prefer signals, and that's a fair choice. If you go that route, a credible track record has four things:

• Every signal is logged the moment it's sent, with a time stamp nobody can edit, including the losers.
• Results use real or realistic fills: the price you could actually have gotten, not the best price of the minute.
• Fees, slippage and the provider's own subscription cost are subtracted. A subscription is a cost like any other.
• The record covers several market regimes, at least a year and ideally more, not one strong quarter.

Most signal services can't show all four. The ones that can are rare enough that just asking is a useful filter.

Six questions to ask any research or signal provider

1. Is every read time-stamped before the open? If not, it can be edited after the fact.
2. Is the full archive available to subscribers, including bad days? A selection of wins is marketing, not evidence.
3. What's the exact session window? "It worked today" doesn't mean much without a set time frame.
4. Is a backtest labeled as a backtest? With its period, costs and assumptions stated.
5. Are illustrations labeled as illustrations? One screenshot isn't a statistic.
6. Does the provider tell you what would make the read wrong? Honest research states its limits.

Where Quantum Bias fits

Quantum Bias publishes research, not signals. The Daily gives a time-stamped directional read, a probability zone and reference levels for NQ and ES before the open. The Macro Bias gives a sourced macro read with a stated cutoff, its limits and what would change it. Every edition is time-stamped and kept in an archive available to members. Entries, stops and position size are still your call.

FAQ

What is quantitative research in trading? Using fixed rules and historical data to measure how a market behaves under set conditions, producing reads you can check against what happened.

Can retail traders use quantitative research? Yes. Backtesting tools, data feeds and published research make it accessible. The discipline is the same: set your rules first, test honestly, include costs.

Are trading signals worth it? Only if every signal is logged with a time stamp and the results include real fills and fees. Most can't show that.

What's the difference between a backtest and real results? A backtest applies rules to past data. Real results come from actual trades and include slippage, fees and the trader's own behavior.

Is quantitative trading guaranteed to be profitable? No. Patterns change and models fail. Quant methods make results measurable. They don't make them certain.

What percentage of day traders lose money? In one big study, 97% of people who day traded Brazilian equity futures for at least 300 days lost money. Other markets and periods may differ, but the burden of proof is on anyone who claims otherwise.

---

Research and education only. Not financial advice. Past results do not predict future returns. Futures trading involves substantial risk of loss.

How this article was made: drafted with AI assistance from published Quantum Bias reports and the sources linked above; facts and figures checked by Gabriel Décarie and Karel Panneton.