Why Your Results Differ From the Signal Provider's

A trader we'll call Sam sent us a message in March that we've received, in one form or another, a few hundred times. He'd followed a gold signal service for two months. Took every trade, or nearly every trade. The provider's published record for the period showed a decent run. Sam's account was roughly flat, maybe slightly down after swap charges. His question was the obvious one, the one everyone eventually types into a search bar at 11pm: why do my results differ from the signal provider's?

The lazy answer is that the provider is lying. Sometimes that's true, and if a service won't show you its full closed history, losses included, you should assume it is. But Sam's provider wasn't lying. We know, because Sam's provider was us, and our closed record sits in public where anyone can audit it.

The honest answer is more useful and less comfortable. Between the moment a signal is posted and the moment your trade closes, there are six places where money leaks out. Spread. Delay. Slippage. Sizing drift. Skipped trades. Early exits. Each leak is small on its own. Stacked together, they routinely turn a winning strategy into a flat account, and a flat strategy into a losing one. The good news, and there is genuinely good news here, is that every leak can be measured, and most can be shrunk. Let's take them one at a time, with real numbers on a gold account, because vague reassurance never fixed anything.

The gap is normal. The size of it is up to you

First, a reset of expectations. There is no such thing as perfectly replicating someone else's trades. None. The provider fires an order on their platform, at their broker, at their moment, with their size. You do the same thing seconds or minutes later, at a different broker, possibly with different size, possibly while making a sandwich. Two different trades. They will produce two different results, every single time, and anyone who tells you otherwise has never actually done it.

The professional term for this is the signal execution gap, and it exists in every corner of the industry, from Telegram channels to institutional trade replication. Copy-trading platforms battle the same problem with software instead of thumbs; we wrote about that trade-off separately in signals versus copy trading, and the short version is that automation shrinks some leaks and adds others.

So the question was never "how do I make my results identical?" That's off the table. The real question is: how wide is your personal gap, which leaks are feeding it, and which of those leaks are you willing to fix? A disciplined follower with a good broker might run a gap of 10 to 15 per cent against the provider's raw numbers. A sloppy follower on a bad account can run a gap of 100 per cent or worse, meaning they lose money on a period the provider won. Same signals. Wildly different outcomes.

That range is the whole story of this article. Everything below is about moving yourself from the sloppy end toward the disciplined end, and knowing exactly which levers do the moving.

Leak 1: the spread you pay is not the spread they pay

Start with the least glamorous leak, because it's the most constant. Spread is the difference between the price you can buy at and the price you can sell at, and you pay it on every single trade, win or lose, whether you notice or not.

On gold, spreads vary enormously between brokers and account types. A raw or ECN-style account might show XAU/USD at 10 to 20 cents of spread plus a commission. A standard account at the same broker might show 35 to 50 cents with no commission. A poor broker, or a decent broker during the Asian session, can drift past 60 cents. During a news spike, all bets are off and even good accounts can flash spreads of a dollar or more for a few seconds.

Here's why this matters more than it looks. Say a signal has a $12 stop and a $24 target, a clean 1:2 trade. If the provider's account fills with an effective 25-cent round-trip cost and yours costs 55 cents, you're giving up 30 cents per trade more than they are. On a $12 risk, that's 2.5 per cent of your risk gone before the trade breathes. Doesn't sound like much. Now run 60 trades a month, which an unlimited gold signal service will easily produce. That's $18 of pure spread disadvantage per 0.01 lot per month, or a full one and a half losing trades' worth of risk, paid for nothing.

Those are illustrative ranges, not quotes; check your own broker's live spread during London and New York hours, because that's when most gold signals fire. The point stands regardless of the exact figures: spread differences between brokers are a permanent tax, and you choose your tax rate once, when you open the account. It is the single easiest leak to fix and the one people fix last, usually because they picked their broker years ago for a deposit bonus and never thought about it again.

Leak 2: the minutes between the signal and your fill

A signal lands at 14:32. Your phone buzzes. You're driving, or in a meeting, or the notification simply sits under fourteen others. You open the app at 14:41 and place the trade. Nine minutes have passed.

On a slow pair, nine minutes might cost you nothing. Gold is not a slow pair. Gold routinely moves $2 to $4 in nine quiet minutes and $10 or more in nine lively ones. If the signal said buy at 3,320 with a stop at 3,308, and the price is 3,324 when you finally fill, you have a genuine decision problem, and both branches cost money:

• Chase it. You enter $4 worse. Your effective stop distance is now $16 against the same $24 target, so your 1:2 trade became roughly 1:1.25. Over a hundred trades, that ratio damage compounds brutally.
• Skip it. You avoid the bad entry, but if the trade wins, you've created a gap of one full winner between your record and the provider's. Miss three winners a month this way and no amount of good execution elsewhere saves you.
• Wait for a pullback. Sometimes it comes and you fill at the original price like nothing happened. Sometimes it never comes and you watch the trade run to target without you, which does wonderful things for your discipline on the next signal. (It doesn't.)

The provider pays none of this cost. Their fill is their fill. Every minute of your delay is a coin flip weighted against you, because the trades that move immediately in the signal's direction, the best trades, are precisely the ones you fill worst on. That's not bad luck, it's selection: delay filters you into the mediocre entries and out of the sharp ones.

There's a workflow version of this leak that deserves its own mention: the double-check. Plenty of careful people receive a signal, then open a chart to "verify" it before entering. Five minutes of squinting at candles later, they either agree with the signal (and enter late) or disagree with it (and skip, feeding Leak 5). Both outcomes are worse than acting promptly. If you trust the service, the verification step adds delay and nothing else. If you don't trust the service, the verification step is a symptom, and the actual fix is finding a service you do trust or going back to your own trading. Half-trust is the most expensive possible setting.

Measure your own delay honestly. Notification timestamp to order timestamp, both visible in your platform history. If your median is under two minutes, this leak is minor. If it's over ten, this leak might be your whole gap on its own.

Leak 3: slippage, and why gold is the worst offender

Slippage is the difference between the price you asked for and the price you got, and people blame it for everything, which is convenient because it's the one leak that isn't their fault. Actual slippage on a market order during normal hours, at a reasonable broker, is small: a few cents on gold, often zero. If someone's account is bleeding, "slippage" is usually the story they tell, and sizing drift or delay is usually the truth.

But there is one place slippage genuinely bites, and gold traders live next door to it: news candles. Nonfarm payrolls, CPI, an unscheduled Fed headline, a geopolitical flash. Gold can jump $8 in a second. In that second, three things happen to a signal follower:

1. Pending orders fill wherever liquidity exists, not where you placed them. A buy stop at 3,325 can fill at 3,329 without anyone doing anything wrong.
2. Stops slip too. Your stop at 3,308 can close at 3,304. That's an extra $4 of loss on a planned $12 risk, a 33 per cent overshoot on that trade.
3. Spread balloons at the exact moment you're transacting, stacking Leak 1 on top.

The provider eats some of this too, to be fair. Their fills degrade on news candles like everyone's. But there's an asymmetry: a provider watching the calendar can time their entry seconds before or after the spike, while a follower reacting to a notification transacts mid-chaos almost by definition.

What can you actually do about it? Less than the fixable leaks, more than nothing. Know the calendar; the red-folder events are public and free. If a signal arrives four minutes before CPI, it's reasonable to wait out the print or take half size, and any provider worth following will tell you their own policy on trading through news if you ask. Ours is in the FAQ, and if a service you're evaluating has never thought about the question, that tells you something. Beyond that, accept it. Slippage on true news candles is a cost of doing business in gold, maybe 1 to 3 per cent of gross results in a normal month, and pretending you can eliminate it leads to worse decisions than budgeting for it.

Leak 4: sizing drift, the silent one

Here's the leak nobody audits, and it's frequently the biggest one on the sheet.

The arithmetic of a signal record assumes consistent risk. If the provider's month was, say, plus six units of risk (six "R"), that number only transfers to your account if every trade risks the same amount. Risk 1 per cent on each of 60 trades and you capture the record's shape faithfully. But that's not what people do. What people do looks like this:

Sam risks 1 per cent on Monday's signal. It loses. Tuesday's signal, still smarting, he risks 0.5 per cent. It wins. Wednesday he's annoyed about capturing only half a winner, so he risks 2 per cent to catch up. It loses. On paper the provider is down one trade out of three, a perfectly ordinary sequence. Sam is down 2.5 units against the provider's 1, and he took the exact same trades.

Notice what happened there: the strategy didn't fail, the sizing did. Reverse the win/loss order and drift can flatter you instead, which is worse in a way, because it teaches you that revenge sizing works. Over any decent sample it doesn't. Erratic sizing adds variance without adding expectancy, and variance is precisely what a small account can't afford.

The mechanics on gold, so you can check your own maths: one standard lot is 100 ounces, so a $1 move is worth $100 per lot, which means $1 of movement on 0.01 lots is $1. A $2,000 account risking 1 per cent has $20 per trade to spend. On a signal with a $12 stop, that's roughly 0.016 lots, so 0.01 or 0.02 depending on how your broker rounds. Not 0.05 because you feel confident. Not 0.01 because you're scared. The same formula, every time: account × risk per cent ÷ stop distance in dollars ÷ 100. Thirty seconds of arithmetic per trade, and it closes what is, for most followers we've audited, the largest single hole in the bucket.

Leak 5: skipped signals and cherry-picking

This one masquerades as intelligence. "I don't take every signal, I only take the good ones." Every trader who says this believes they're filtering. Almost all of them are cherry-picking, and the difference is measurable: filtering improves your win rate on the trades you take, cherry-picking just reduces your sample and lets your mood pick the trades.

The problem is structural. A signal service's edge, if it has one, is a statistical property of the whole stream. A forex signal accuracy rate of, say, 55 per cent with 1:1.5 average reward-to-risk is a solid, profitable stream, but the profit is not evenly distributed through it. It arrives in lumps. A handful of trades each month do the heavy lifting, and there is no reliable way to know in advance which ones. Skip a third of the signals at random and you don't get two-thirds of the results; you get two-thirds of the results on average, with a horrible variance around that average, and some months you'll have skipped exactly the lumps.

Run the arithmetic on a plausible month to see the shape of it. Sixty signals, of which five are the big winners that carry the month. Take 80 per cent of the trades at random and the chance you caught all five is about a one-in-three proposition; more often than not, you missed at least one, and reasonably often you missed two. The provider's month doesn't care. Yours does, a lot, because those two trades might have been half the month's profit.

And your skips are not random, which makes it worse. You skip after losses, when the next trade statistically owes you nothing but the stream is unchanged. You skip setups that "look wrong" to an eye that, respectfully, subscribed to signals precisely because its own analysis wasn't paying. You skip Friday afternoons. Meanwhile you never skip after a win, so you're systematically most exposed right after the streak that's about to mean-revert.

Take the whole stream or find a stream you trust enough to take whole. Half-following a signal service is the most expensive way to use one.

We put skipped trades and psychology under a proper microscope in the psychology of following signals, because the fix here isn't information, it's behaviour. But the audit part is simple: count the signals posted last month, count the ones you took, and calculate what the skipped ones did. Most people have never once run that calculation. Most people would rather not know.

Leak 6: early exits and moved stops

The final leak is the one committed with your own hands, in the moment, with the trade open and your pulse involved.

The signal says target 3,344. Price grinds to 3,335, stalls, ticks down twice. You close it. Fifteen dollars of planned reward, eleven collected, and it felt prudent while you did it. Then price runs to 3,344 without you and the provider's record books the full winner while yours books 73 per cent of one. Do that on a quarter of your winners, which is a conservative estimate of how often nerves win, and you've shaved something like 7 per cent off your gross before any other leak gets a turn.

Moved stops are the same sin with a worse tail. Widening a stop "to give it room" converts a planned $12 loss into an unplanned $20 one, and it only has to work a few times to install the habit permanently. The provider's record assumes the stop is the stop. The moment you renegotiate mid-trade, you're not following the service any more; you're trading your own system, one with no track record, using their entries as vague inspiration.

There's an asymmetry worth staring at here. Early exits truncate your winners. Moved stops extend your losers. Combine the two and you have manufactured, with real effort, the exact opposite of the "cut losses, let winners run" shape that makes the underlying record work. We've audited follower accounts where every entry matched ours within a dollar and the equity curve still pointed the wrong way, entirely because of exits. Every entry right. Every exit renegotiated. The signal was fine; the follower was running a different, worse strategy on top of it.

If you want one behavioural rule from this entire article, take this: set the stop and target when you place the trade, then put the phone down. The provider manages the signal; changes get posted if they're needed. Your job between entry and exit is, deliberately and completely, nothing.

Why my results differ from the signal provider's: adding it all up

Time to stack the leaks and watch what they do to a realistic month, because the compounding is the part people underestimate. Numbers below are illustrative arithmetic, not a performance claim; gold trading is high-risk and plenty of months are simply negative for everyone, follower and provider alike, which is why our own losses sit in public at /signals/history next to the wins.

Suppose the provider's stream produces +6R gross in a month across 60 signals, with your R set at $20 on a $2,000 account. Perfect replication would be +$120. Now leak by leak:

• Spread disadvantage (Leak 1): 30 cents extra per trade on a ~$12 average stop is about 2.5 per cent of risk per trade, roughly −1.5R across 60 trades. Running total: +4.5R.
• Entry delay (Leak 2): a median 6-minute delay on gold costs conservatively 0.05R per trade in worsened entries and missed fills, −3R. Running total: +1.5R.
• News slippage (Leak 3): two bad prints in the month, −0.5R. Running total: +1R.
• Sizing drift (Leak 4): unstable risk turns your realised +1R into anything from −1R to +2R. Call it −0.5R of expectation for the drift itself. Running total: +0.5R.
• Skipped signals (Leak 5): you took 48 of 60, and the skips happened to include two full winners, −2R against the record. Running total: −1.5R.
• Early exits (Leak 6): four winners cut at 60 per cent of target, −0.6R. Final: about −2R.

Read that again, slowly. The provider had a legitimately good month. The follower, doing what felt like following, lost money, and every single step of the loss was ordinary. No scam, no lies, no bad luck worth the name. That is the signal execution gap in full costume, and it is why "the signals don't work" and "the signals work" are so often both true, about the same signals, in the same month.

It's also, in fairness, why you should be suspicious of any service that implies its published numbers are what you will earn. They aren't, ever, for anyone. The honest framing is that the published record is the ceiling and your execution decides how far below it you live.

A two-week audit that puts a number on your gap

Enough theory. Here's the practical core of this piece: a fourteen-day audit that tells you, in numbers, where your personal gap comes from. It costs nothing except honesty and about ten minutes a day.

For every signal posted over two weeks, record eleven things in a spreadsheet:

1. Signal timestamp and the provider's stated entry, stop, and target.
2. Whether you took the trade. If not, why, in five words or fewer. ("Was asleep" and "didn't like it" are different diseases.)
3. Your order timestamp. The delta from item 1 is your delay.
4. Your actual fill price versus the signalled entry. The delta is your entry cost.
5. Your position size, and the risk per cent it implies against your stop. Watch this column wobble.
6. The spread your broker showed at entry, if you can capture it.
7. Your exit price and time versus the signalled exit. Any early exit or moved stop gets flagged in its own column, no excuses recorded, just the flag.
8. The trade's result in R, and the provider's result in R.
9. For skipped trades: what the signal ultimately did, in R. This column will hurt.

Two weeks on an unlimited gold service gives you 25 to 35 rows, which is thin for statistics but plenty for diagnosis. Then total three numbers: the provider's gross R, your gross R, and the gap between them broken down by cause. Sort the causes descending. In our experience of walking followers through exactly this exercise, the top line is sizing drift or skips far more often than it's anything the broker did, which is uncomfortable and useful in equal measure, because those are the free fixes.

One warning about the audit itself: don't change your behaviour during it, or you'll measure your best fortnight instead of your real one. There's a well-known effect where the act of observing yourself improves the behaviour being observed, and while that sounds like a free win, it isn't; the improvement evaporates when the logging stops, and you'll have learned nothing about your normal state. So log quietly, judge nothing until day fifteen, and let the spreadsheet see the trader you actually are. Audit first. Fix second.

If a fortnight feels short, it is, for anything statistical. It's not trying to be statistical. You're not measuring whether the strategy works over two weeks, which would be meaningless; you're measuring how faithfully you transmit it, which shows up almost immediately. A delay problem is visible in five rows. A sizing wobble is visible in ten. The strategy question needs months and lives in the provider's public record, not in your fortnight.

What the provider controls, and what you control

A clean division of labour clarifies who owes you what. Worth writing down, because both sides of this industry blur it when convenient: followers blame providers for execution leaks, and bad providers blame followers for strategy failures. Neither excuse survives the split below.

The provider controls: the strategy's underlying edge, the quality and clarity of each signal (exact entry, stop, target, no "buy gold now!!" nonsense), the speed and reliability of delivery, updates when a trade needs managing, and, this is the big one, the honesty of the published record. Every closed trade public, losses included, no quiet deletions. If a provider fails on these, no amount of good execution on your end will save you, and you should leave. We put together the questions worth asking any signal service before money changes hands, and the record question is first for a reason.

You control: your broker and account type (Leak 1), your alert setup and response time (Leak 2), your news-hour policy (some of Leak 3), your sizing discipline (Leak 4, entirely), your take-every-trade discipline (Leak 5, entirely), and your hands-off-the-exit discipline (Leak 6, entirely).

Nobody controls: what gold does next. Any provider claiming otherwise is selling you a story, and any follower demanding otherwise is buying one.

Tally that up and notice the balance: three leaks fully yours, one mostly yours, one shared, one structural. Roughly 70 per cent of the execution gap lives on your side of the table. That's not a scolding, it's the most optimistic sentence in this article, because the part of the problem that's yours is the part you can actually fix without asking anyone's permission.

Fixing the fixable: broker, alerts, workflow

Now the repair list, in order of effort-to-payoff.

Broker and account type first. This is a one-time decision that repriced every future trade, so do it properly. Compare live XAU/USD spreads across two or three regulated brokers during London and New York hours, not the marketing page numbers, the actual live ones. Prefer a raw-plus-commission structure for gold if the all-in cost beats the standard account, which it usually does at reasonable volume. And check execution reviews for news behaviour, since that's when broker quality actually differentiates. If you're evaluating partner brokers anyway, our signals are free with a maintained $250 balance at a partner broker instead of the $99 monthly fee; the trade execution is the same either way, so let the spread comparison decide, not the discount.

Alerts second. Your delay leak dies with about twenty minutes of setup. Signal app notifications set to priority or "time-sensitive" so they cut through focus modes. A distinct notification sound for signals only, so your brain triages before your eyes open the phone. If your platform supports it, price alerts at signalled entries as a backstop. Then measure your median delay again after two weeks; if it hasn't dropped under three or four minutes, the problem is your day structure, not your settings, and the honest fix might be accepting you can only trade the signals that arrive in the hours you're actually reachable, taking that as a known cost rather than pretending otherwise.

Workflow third. Write your sizing formula on a card, physically, and calculate it before every entry. Pre-commit, in writing, to taking every signal for a defined period, with skips permitted only for a rule you wrote down in advance (for instance: nothing inside five minutes of red-folder news). Place stop and target with the entry, always, then close the app. None of this is sophisticated. That's rather the point; the leaks aren't sophisticated either.

Pricing in the unfixable: adjusting expectations

After all that, a residual gap remains, and this section exists so you stop being angry at it.

You cannot get the provider's fills. You cannot avoid all news slippage. You cannot be awake for every signal unless the service happens to trade only your hours. A well-run follower account, good broker, fast alerts, iron discipline, should expect to realise something like 80 to 90 per cent of the provider's R over a decent sample. Not 100. Never 100. If your audit shows 85 per cent capture, you are not doing it wrong; you are doing it about as well as it can be done, and the remaining 15 per cent is the price of not being the person pressing the original button.

Build that price into your expectations before you evaluate a service, and into your maths when you decide whether a service is worth its fee. A $99 monthly subscription on a $2,000 account needs the strategy to clear roughly 5 per cent a month after your execution haircut just to pay for itself, which is a demanding bar and worth saying out loud; that maths gets easier as the account grows or if the broker-partner route replaces the fee, and it never becomes automatic. Anyone who tells you it's automatic is lying to you. Some months the best available result is a small loss, executed well, and learning to recognise a well-executed losing month as success is, honestly, half of surviving this business.

The residual gap also changes how you should read any provider's published record, including ours. Treat the record as evidence about the strategy, not a forecast of your balance. Then treat your audit capture rate as evidence about you. Multiply the two and you have, for the first time, a realistic expectation instead of a hope.

The audit worksheet: find your biggest leak this fortnight

Here's where this leaves you. Not with a moral, with a task. Starting with the next signal you receive, run the two-week audit. To make it concrete, the checklist version:

• Days 1–14: log every signal: timestamps, fills, size, spread, exits, and results in R, yours and the provider's, including the trades you skipped.
• Day 15: total the gap and attribute it: spread, delay, slippage, sizing, skips, exits. Sort descending.
• Day 16: fix the top item only. One change. Broker comparison if it's spread, notification surgery if it's delay, the card with the formula if it's sizing, the written pre-commitment if it's skips or exits.
• Days 17–30: run the audit again and compare capture rates.

Our bet, from years of watching people do this, is that your biggest leak is not the one you'd have guessed, and that fixing your top two closes more than half of your gap inside a month. Trading gold through anyone's signals remains high-risk, losing runs are a certainty at every capture rate, and no audit changes what the market decides to do. But there is a real and large difference between losing to the market and losing to your own execution, and until you've measured, you genuinely don't know which one has been eating your account.

Sam ran the audit, since you're wondering. His top leak was skips, his second was sizing, and his broker, the thing he'd been blaming for two months, was mid-table. His gap didn't vanish. It narrowed to the point where the record he was following finally showed up, recognisably, in his own account, and that, not perfection, is what following signals well actually looks like.