Why 'Accurate' Is Meaningless Without a Method
Every signal provider claims to be accurate. Almost none say how they measured it, and that omission is the whole story. Search for the most accurate paid trading signal providers and you will find a wall of percentages — 85%, 90%, 95% — with no unit attached and no dates to check them against. Accuracy can be calculated two different ways, and the two methods can point in opposite directions on the exact same set of trades. For a full rundown of what a complete signal should even contain, see our best trading signals guide.
By trade count simply divides winning trades by total trades. It is the number marketed hardest, because it is the easiest to inflate: a provider can close nine small trades in profit and one large trade at a loss, market a high win rate, and still hand subscribers a net loss for the month. By points (or by pips, or by net result) measures the actual scoreline — what was gained minus what was lost, in the same unit across every trade. It cannot be gamed the same way, because a single oversized loss shows up as exactly what it is.
The table below runs both methods on the identical ten trades, so the difference is not abstract.
The mechanism behind that gap is not luck, it is proportion. A provider that lets winners run small and losses run large can post an impressive win rate while a subscriber's account still shrinks, because trade-count accuracy ignores position size and loss size entirely. A provider that reports by points has already absorbed that mechanic into a single, harder-to-dress-up number — which is why every accuracy figure on this page, including our own, is quoted by points and dated, not left as a bare percentage.
By trade count vs by points — the same 10 trades, two different headlines
| By trade count | By points | |
|---|---|---|
| What it counts | Winning trades divided by total trades | Net points gained minus net points lost |
| The same 10 trades | 9 wins, 1 loss = 90% win rate | 9 wins of +5 points each (+45), 1 loss of -60 points = -15 net points |
| Headline it produces | 90% win rate | Net loss for the period |
| What it hides | One oversized loss can erase many small wins | Nothing — the net result is the net result |
| What we publish | Not used | 93% average weekly accuracy by points, +197,891 net points over 37 weeks |
How to Read Any Win-Rate Claim
Before trusting a headline number from any provider, including us, ask which method produced it. Most marketing pages do not say, and that silence is itself the answer.
Apply the checklist below to us, not just to competitors. Our own record is published by points, updated weekly and shows losing weeks in the same feed as winning ones — the performance page is built to survive exactly this scrutiny, and we would rather you check it than take our word for it.
- Ask directly which method was used: by trade count, or by points and net result
- Ask whether losing weeks or months are shown, not just winning ones
- Check whether the figure is dated and updated on a fixed schedule, not a one-time screenshot
- Look for the average loss size next to the win rate — a high win rate paired with rare, oversized losses is a warning sign, not a strength
- Confirm the track record covers a real time span in weeks or months, not a handful of cherry-picked trades
Which Trading Signals Have the Highest Win Rate and ROI?
These are not the same question, and providers exploit the gap between them. Win rate counts how often a trade closes in profit. ROI counts how much money actually came out the other end. A service can post a high win rate and a flat or negative ROI, and a service with a moderate win rate can post a strong ROI, because ROI depends on the size of the wins relative to the size of the losses, not just how often a trade wins.
The table below illustrates the mechanic with generic, hypothetical figures — it is not a claim about any named provider.
Win rate alone does not decide ROI — a general illustration
| Hypothetical Service A | Hypothetical Service B | |
|---|---|---|
| Win rate | 85% | 55% |
| Typical loss size vs typical win size | Losses tend to run larger than wins | Wins tend to run larger than losses |
| Likely result over many trades | Can still be net negative | Can be net positive with fewer wins |
| Lesson | A high win rate does not guarantee a high ROI | ROI depends on win rate together with win and loss size, not win rate alone |
How We Calculate Our Own Number
Since this whole page argues that method matters, our own method should be stated plainly rather than assumed. Our weekly accuracy is net points across every signal closed that week, summed first, classified second: a week counts as accurate only if that net-point total finishes positive — not if most individual trades happened to hit target. 93% is the share of the 37 published weeks that finished positive on that basis, not the share of individual trades that won. The +197,891 net points figure is the running total across all 37 weeks combined, gains and losses netted against each other, not a sum of only the winning weeks.
Two consequences follow from that. First, a week with one outsized loss and several small wins can still count as a losing week even though most trades in it 'won' individually — the points decide, not the count, which is the same logic the worked example above uses. Second, all 37 weeks are published on the performance page, losing weeks included, with the same net-point calculation applied every time rather than recalculated after the fact to look better.