Are Futures Trading Signals Worth It? A Buyer's Checklist

STS ResearchPublished June 13, 2026Data as of November 6, 2026Updated November 10, 2026

Most futures trading signals are not worth paying for. The reason is simple. Most services cannot show you a documented, multi-year record with every trade, every loss, and the real drawdown included. A signal is only worth its price if the proof behind it survives a few hard questions. Most never even get asked.

So here is the buyer's checklist. Seven things a serious futures signal service should be able to show you before you pay. We use our own NQ book as the example of what "good" looks like, but the checklist is yours to point at any provider. Including us.

3,139
Trades you can see, wins and losses
>3.0
Our "real edge, not luck" score (clears the 3.0 bar)
$25,342
Our worst dollar drawdown, still open, published not hidden
13/16
Positive years, weakest shown

Whose numbers are these (read this first)

These figures come from our own book. Four systematic NQ strategies (Trend, Opening Range Break, VWAP Short and Overnight) run as one single-position portfolio, one trade on at a time. They are TradingView backtests from 2011 to 2026, sized at 1 or 2 NQ minis a trade (2 when daily ranges are small, 1 when they are large; at today's index level every trade since July 2024 has been 1 contract), with commissions and slippage included. The total is $696,892 net, +696.9% on a $100,000 nominal basis. The style is momentum and trend continuation, not mean reversion and not scalping.

That context matters. These are our system's results, not a rule about the NQ market itself. You cannot take our exact numbers and expect them from a different strategy. What you can take is the method. The checks we run on ourselves are the ones you should run on anyone selling you signals. The proof standard transfers even when the numbers do not.

What should I check before buying futures trading signals?

Seven things. Our book meets six of them today, including 15 years of record, all 3,139 trades shown, a published $25,342 drawdown, a t-stat above 3.0 and a Deflated Sharpe above 90%. The seventh, out-of-sample results for our current rules, starts with their forward record on November 7, 2026.

Before the details, here is the full list, with what our book reports for each line.

Checklist table: seven proof checks a futures signal buyer should demand, with the STS NQ book's value for each. Backtested track record 15 years; every trade visible 3,139; backtested drawdown shown $25,342; Harvey-Liu t-stat clears the 3.0 bar; Deflated Sharpe clears the 90 percent bar; Probabilistic Sharpe passes; profit factor by 4-year era rising 1.05 to 1.73. Every row is marked met. Checklist table: seven proof checks a futures signal buyer should demand, with the STS NQ book's value for each. Backtested track record 15 years; every trade visible 3,139; backtested drawdown shown $25,342; Harvey-Liu t-stat clears the 3.0 bar; Deflated Sharpe clears the 90 percent bar; Probabilistic Sharpe passes; profit factor by 4-year era rising 1.05 to 1.73. Every row is marked met.
Every proof row in this table is met: 15 years, all 3,139 trades shown, $25,342 drawdown published, t-stat above 3.0, Deflated Sharpe above 90%, per-era profit factor rising 1.05 to 1.73. Out-of-sample results for our current rules start November 7, 2026.

Now the seven, one at a time.

How long should a signal service track record be?

Years of trades, not a hot streak from last month. Our record runs 15 years, 2011 to 2026.

Ask how long the record is. A good answer is years of trades, not a hot streak from last month.

Anyone can post a great month. Markets hand out lucky streaks for free. What is hard to fake is a result that holds through a bull market, a crash, a chop year, and a few drawdowns. Our record runs 15 years, 2011 to 2026. If a service can only show you the last few weeks, you are not buying an edge. You are buying a coin flip with a nice chart. The same arithmetic cuts against us too: a month of our own signals is far too small a sample to judge our edge, which is why we publish the full record.

Years alone are not proof either. A long record can still be curve-fit to the past, which is why the checks below test exactly that.

Should a signal service show its losing trades?

Yes, all of them. Our backtested book is 3,139 trades and we win 45.3% of them, so the edge is that our average win of about $1,590 is 1.74 times our average loss of about $912.

Ask to see all the trades. Not the highlights. All of them, with the losing ones in plain sight.

Our backtested book is 3,139 trades. We win 45.3% of them. That number surprises people who expect a "winning system" to win most of the time. It does not.

The edge is not in being right often. It is in the size of the wins against the losses. Our average win is about $1,590. Our average loss is about $912. So a winning trade is worth 1.74 times a losing one. That payoff is the edge, not the hit rate.

And you can only judge it if the losers are on the table next to the winners. A service that shows you only winning trades is hiding the half of the data that matters.

What drawdown should a futures signal service disclose?

The worst peak-to-trough loss, in dollars. Ours is $25,342 on our NQ book and still open, and one reshuffled ordering in twenty ran past $51,684, the figure an account has to survive.

Ask for the worst peak-to-trough loss. The single deepest hole the account fell into, in dollars.

Ours is $25,342, about 3.1% of the $804,557 equity peak it fell from, or 25.3% of a fixed $100,000 account. It began on August 4, 2026 and is still open at the end of the record, so it is deeper than any drawdown the book has finished. Our worst in percentage terms was a separate, earlier 20.4% episode, from 2011 to 2014. We publish it on the tear sheet, and we tell buyers what a reshuffle of those same trades says about it. We shuffled the order of the trades 10,000 times. Our real drawdown sits at the 5th percentile: about 95% of the reshuffled orderings were worse. The order our trades came in was a kind one, so plan for more than our worst. The median reshuffled drawdown was $34,223. One ordering in twenty ran past $51,684, and that is the figure an account has to survive. We unpack why your live drawdown usually beats the backtest in expect a worse drawdown than your backtest. A service that talks only about returns and goes quiet on drawdown is selling you the upside and hiding the part that actually decides whether you can stick with it.

What t-stat shows a trading edge is probably not luck?

Above 3.0, the bar set by Harvey, Liu and Zhu. Below 3.0 a record could easily be luck, and our backtested book clears that bar.

Ask for the t-stat. It is a single number that says how far a record sits above pure luck.

Bar chart comparing Harvey-Liu t-stat against the real-edge hurdle. A coin-flip system reads 0.00, the academic hurdle is 3.00 (dashed amber line), and the STS NQ book clears the hurdle, sitting above 3.00, over 3,139 trades and 15 years. Bar chart comparing Harvey-Liu t-stat against the real-edge hurdle. A coin-flip system reads 0.00, the academic hurdle is 3.00 (dashed amber line), and the STS NQ book clears the hurdle, sitting above 3.00, over 3,139 trades and 15 years.
A coin-flip system scores 0.00. The academic hurdle for a real edge is 3.00. Our 3,139-trade backtested book clears it. The chart is drawn to no scale, so no exact value is implied.

Here is the rule. Below 3.0, a record could easily be luck. Above 3.0, it probably is not. That 3.0 bar comes from Harvey, Liu and Zhu, who studied thousands of strategies and found the old bar for "real" was set too low.

Our book clears that 3.0 bar. The number is adjusted so a big trade count alone cannot inflate it. Most services have never run this test and could not tell you their number if you asked. That alone tells you something.

How do I know if a trading signal strategy is overfit?

Ask for real overfitting tests. Our Deflated Sharpe clears the 90% bar even after deflating for 10,000 variations, and our Probabilistic Sharpe passes as well.

Ask whether the strategy was overfit. Overfitting is when a system is tuned so tightly to the past that it looks brilliant on old data and falls apart on new data. It is the single most common way backtests lie.

There are real tests for it, and a serious provider has run them. We did. The Deflated Sharpe ratio strips out the credit a strategy gets just for being one of many tries. Higher is better, and ours clears the 90% bar even after deflating for 10,000 variations. A second check, the Probabilistic Sharpe, passes as well: the chance the true edge sits above zero is effectively certain. We clear both.

Neither test is magic, and we run them on ourselves, so check them on the tear sheet rather than take our word for it. We go deeper on these exact tests in how to tell if a backtest is overfit. A service that has never heard of them has not done the homework you are paying for.

What out-of-sample results should a signal provider show?

Results on data the system was never built on. This is the one check our current rules cannot pass yet: they were set in November 2026 using the full history, so their out-of-sample record starts on November 7, 2026. Until it builds, the profit factor across four 4-year eras rises from 1.05 to 1.73.

Ask how it did out of sample. That means: build the system on one slice of history, then test it on a slice it never touched. If the edge only exists on the data used to build it, it is not an edge.

We hold ourselves to the same rule, and today we fall short of it. The current rules of our book were finalised in November 2026 on the full 2011 to 2026 history, so no part of that history is out of sample for them. Their out-of-sample record starts on November 7, 2026, and every signal from that date on is public on the forward record. What we can show meanwhile is weaker evidence: the profit factor rises across four consecutive 4-year eras, from 1.05 to 1.73, rather than fading. That says the edge did not decay over the record. It does not replace a test on new data, so weigh it that way.

How do I tell if a signal service is just buy-and-hold?

Ask whether the returns depend on the index rising. Our beta to NQ is withdrawn while we re-derive it, but our book made $52,017 in 2022, a year the Nasdaq-100 fell, and one of its four strategies trades only short.

Ask whether the returns are really just the market. If a futures system only makes money when the index rises, you are paying a subscription for something a cheap index fund does for free.

Our beta to NQ is withdrawn while we re-derive the benchmark regression it came from, so we will not quote a number we cannot currently stand behind. What we can show is how the book did when the index fell. In 2022 it made $52,017, and Trend ($24,726), Opening Range Break ($17,655) and VWAP Short ($9,636) all made money that year. The book is mostly long, 2,720 of its 3,139 trades, but VWAP Short trades only short (419 trades) and Overnight trades the evening session, apart from the other three. A service whose "edge" is really just buy-and-hold in a costume is not worth a monthly fee.

What this checklist cannot do

Be honest about the limits, because we are. Passing these checks does not promise a profit. It cannot. Every number here is hypothetical backtest performance, and the future can always break a pattern that held for 15 years.

Our own record has a weak spot we say out loud. The profits are heavy in the recent years: 2020 through 2025 produced 76.1% of the net profit, and those were strong years for the Nasdaq trend our system rides. The real risk to a system like ours is not a fake edge. It is a change in market character that the past did not contain. A checklist tells you whether a provider has done honest work. It does not tell you tomorrow will look like yesterday. Anyone who promises that is the first to walk away from.

The takeaway

A futures signal is worth paying for only if the proof survives questions. Demand a multi-year record, every trade including losses, the real drawdown, a t-stat above 3.0, an overfitting test, out-of-sample results, and proof the edge is not just the market rising. If a provider cannot show these, the honest answer to "is it worth it" is no.

How we measured this

Instrument: CME Nasdaq-100 E-mini (NQ), $100,000 starting capital, no compounding. Size is 1 or 2 NQ minis a trade (2 when daily ranges are small, 1 when they are large; at today's index level every trade since July 2024 has been 1 contract), so the dollar totals already include the size that was on.

Data: the four-strategy intraday book we trade today, backtested on TradingView, 2011-11-16 through 2026-11-06, 3,139 trades, with commissions and slippage included. The book-level figures (trades, win rate, net P&L, profit factor, drawdown) match the numbers published across this site.

The validation numbers come from standard tests on those trades: the Harvey-Liu t-stat, the Deflated Sharpe ratio, the Probabilistic Sharpe, the per-era profit-factor trend, and the book's 2022 result and long/short split. The drawdown Monte-Carlo reshuffles our backtested trades 10,000 times to estimate an honest planning range.

These are derivations from backtested TradingView trades, not a fresh simulation, so they inherit the engine's fills. The limit is the one we stated: heavy reliance on a recent, trend-friendly regime.

What to do with this before you pay anyone

Take the seven checks and use them as questions in your next email to a signal provider. Ask for the multi-year record, the full trade list, the real drawdown, and the overfitting and out-of-sample work. Watch what happens. The honest ones answer with numbers. The rest change the subject to testimonials and screenshots of one good week, and a screenshot cannot be re-run, which is why backtest screenshots go stale after a contract roll.

Run the same questions on us. Our full numbers, drawdowns and all, are on the strategy page and the tear sheet, and the same record laid out against this checklist is on our NQ futures signals page, with the full 15-year record. The stop-loss and drawdown work we publish lives in how many points a stop should be and how long our drawdowns last. If you want the actual NQ entries from the four systems measured here, the pricing page has the plans. We would rather you check the proof first than take our word for it.


We trade this book live and sell access to the signals, so judge the data accordingly. This article is educational and is not investment advice. Futures trading involves substantial risk of loss and is not suitable for every investor.

Hypothetical performance disclaimer (CFTC Rule 4.41): hypothetical or simulated performance results have certain limitations. Unlike an actual performance record, simulated results do not represent actual trading. Also, since the trades have not been executed, the results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profit or losses similar to those shown. Past performance does not indicate future results.

See our forward record, every live signal we have sent, dated and public.

Hypothetical performance disclosure (CFTC Rule 4.41). These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under-or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.

Past performance is not indicative of future results. Trading futures involves substantial risk of loss.