Do Trading Signals Work? The Honest Answer, With Our Own Numbers
Sellers, predictably, say yes. We publish the number that complicates it: our own raw scanner, traded blind in a hypothetical backtest, produced a 46.6% simulated win rate and a 0.82 simulated profit factor. Research and education only — not financial advice.
The short answer
Do trading signals work? Stripped of everything around them, most raw signals perform about as well as a coin flip — and that is not cynicism, it is what our own data shows. When we ran our raw scanner output through a hypothetical backtest with no filters, no catalyst check, and no human review, the simulation produced 161 trades, a 46.6% win rate, a profit factor of 0.82, and an expectancy of roughly −2% per trade. Traded blind, the simulated account bled slowly and politely.
The signal itself turns out to be the least important part of a trading system. What happens around it — filtering, risk limits, exit rules, and honest record-keeping — is where the real variance lives. This page walks through the evidence, including the parts that flatter nobody, starting with us.
What our own hypothetical backtest found
Most signal services publish their winners and let the rest quietly disappear. We took the opposite approach and published the unfiltered number on our own scanner. The test was simple: take every card the raw scanner generated, enter at the trigger, exit at the stated levels, apply zero discretion, and tally the result.
| Metric | Raw scanner, traded blind (hypothetical) |
|---|---|
| Simulated trades | 161 |
| Win rate | 46.6% (simulated) |
| Profit factor | 0.82 (simulated) |
| Expectancy | ≈ −2% per trade (simulated) |
A profit factor below 1.0 means gross simulated losses exceeded gross simulated gains — the strategy gave back more than it made. And a 46.6% simulated win rate sits close enough to a coin flip that the gap may be nothing but noise. The full methodology and the rest of the audit live at our public record.
Why raw signals hover near a coin flip
The result surprised nobody on the desk, because the forces pushing raw signals toward 50/50 are structural:
- Competition. Obvious patterns attract capital until the edge is gone. A scanner condition simple enough to code in an afternoon is simple enough for thousands of others to have coded first.
- Costs. Spreads, commissions, and slippage convert a genuinely 50/50 process into a losing one. Our simulated −2% expectancy per trade is what "roughly random" looks like after friction.
- Regime dependence. A momentum signal built in a trending market meets a choppy one and quietly inverts. The signal did not break; the world changed.
- The marketing filter. You mostly hear about the configuration that worked, tested on the period where it worked. The nine that failed never make the sales page.
The variables that actually move the number
The honest version of "do trading signals work" is "which variables around a signal change the math." Four of them dominate.
Filtering
Our desk's full pipeline scans thousands of symbols per session, then runs each candidate through a catalyst check, an adversarial review — a second pass whose only job is to attack the idea — and a liquidity screen. Most raw signals die in that gauntlet, deliberately. A signal that survives hostile review is a different object from one that merely fired.
Risk discipline
Every card we post carries a trigger, TP1/TP2 targets, a stop, and a time-stop, all defined before the move. Pre-committed exits do not make an idea right; they define in advance what being wrong is allowed to cost and remove the mid-trade improvisation that quietly ruins accounts.
The exit, not the entry
When we tested a 21-variant grid of exit rules over the same entries, the simulated outcomes ranged from ugly to superficially spectacular. Identical signals, different exits, wildly different simulated equity curves. The entry gets the attention; the exit does the work.
Record-keeping
A signal that is not timestamped before the move is a story, not a data point. Our cards post to a public, timestamped paper record before the move happens, losses stay on the board, and corrections are posted in the open. That does not make the signals better — it makes the evaluation honest.
The seductive backtest trap
Here is the part most services would have led with. Inside that 21-variant exit grid, the best-looking cell showed +362 simulated units — a number that would look magnificent in a promotional graphic. Our own statistical audit rejected it. One ticker accounted for 61% of the simulated profit, and the result failed significance testing: remove a single hot name and the magic evaporated. The "strategy" was mostly one stock having a good month.
Had we led with that cell, nothing in the marketing would have been technically false. It would still have been meaningless. That is the trap with backtested signal performance generally — the space of testable configurations is enormous, and something always looks brilliant in hindsight.
How to judge whether any signal service holds up
- Timestamped calls, posted before the move. Screenshots assembled after the fact prove nothing.
- Losses visible on the same board as wins. A record showing only winners is a record that has been curated.
- Hypothetical and paper labels attached to every number — in the same sentence, not in a footnote.
- A described methodology. "Proprietary algorithm" is not a methodology; it is a curtain.
- No income promises. Nobody can promise trading outcomes. A service that does is telling you about its marketing, not its research.
We keep a fuller checklist in how to vet any signal room and a companion list of trading discord red flags worth reading before paying anyone — including us.
So what is a signal actually for?
Used honestly, a signal is a structured hypothesis: this instrument, this trigger, these targets, this stop, this time limit. It compresses hours of scanning into a reviewable card and forces pre-commitment, which is where most self-directed traders quietly fail. That is what our signal cards are — trigger-based research cards posted to a public paper record, a model desk with no real money attached, graded in the open.
What a signal cannot do — ours or anyone's — is promise an outcome. Markets do not sell that product, at any subscription price.
The bottom line
So, do trading signals work? As standalone money printers: the evidence, including our own published hypothetical numbers, says no — raw signals traded blind approximated a losing coin flip in our simulation. As structured, filtered, risk-defined research inputs judged against a transparent record: they can be a legitimately useful way to study what disciplined trading is supposed to look like. The distance between those two sentences is most of this industry.
Common questions
Do trading signals work without any filtering or risk rules?
What win rate makes a signal service good?
Are paid trading signals better than free ones?
How can I verify a signal service's track record?
Can any signal service guarantee profits?
Free to join · paid floors optional · research and education only
Last updated 2026-07-16 · ClaudeQuantAlgo Research Desk · research and education only.
Disclosures. ClaudeQuantAlgo is a research and education community. Nothing on this page constitutes financial, investment, legal, or tax advice, or a recommendation to buy or sell any security or derivative. No profit promises are made, ever — trading stocks, options, and forex involves substantial risk of loss; options positions can lose 100% of their value. The public scoreboard reflects a model ("paper") desk — no real money. Past performance — real, paper, or simulated — never guarantees future results. Hypothetical and simulated results have inherent limitations and no representation is made that any account will or is likely to achieve similar profits or losses. ClaudeQuantAlgo is not a registered investment adviser or broker-dealer. You are solely responsible for your own trading decisions. Never risk money you cannot afford to lose.