Almost every trader meets the demo vs live trading gap in the same order. Weeks of tidy simulated results, then a live account that behaves like a different instrument entirely.
Nothing mysterious causes that. A demo server fills you kindly, prices you generously and asks nothing of your nerve, so the two records were never measuring the same thing.

The panel above pairs twenty simulated fills with twenty live ones. Demo lands on the mark, and live lands a shade worse almost every time.
Demo vs Live Trading: Where the Gap Comes From
Four separate forces open the gap. Execution, pricing, account rules, and the trader’s own behaviour under pressure.
Only the last one gets discussed. The first three cost more, though, and they cost it quietly on every single order.
A Demo Server Simulates the Broker, Not the Market
Prices on a demo feed usually come from the live feed. Fills do not.
Your simulated order never reaches a liquidity provider. Software accepts it at the displayed price, so the queue, the rejection and the delay never happen.
The Gap Grows With Trade Count
One trade hides the difference. Two hundred trades put it on the equity curve where nobody can miss it.
So a scalping method suffers most. A method holding positions for days barely notices, because friction spreads thinly across a long hold.
A Worked Example in Pips
Take a rule that aims for eight pips and risks six. On demo it fills at the price on screen both ways, so the quoted spread becomes the only cost it ever meets.
Now price the live version of the same trade. Half a pip of spread on entry, half a pip on exit, a fifth of a pip of drift each way, plus commission worth roughly another half pip.
Round that up and one round turn costs close to two pips. Against an eight pip target, friction now takes a quarter of every trade that goes your way.
Losing trades carry the same friction. So the two records separate on both sides of the ledger rather than only on the good days.
Run the same sum on a rule aiming for eighty pips. Two pips of cost barely registers, which is why slower rules travel across to live money far more comfortably.
Where the Gap Runs Widest
- Scalping methods. A few pips of target against a cost that arrives on every round turn, so friction eats the largest share of the result here.
- News driven entries. Spread widens and depth thins at exactly the moment the rule wants a fill, which a smoothed demo feed rarely reproduces.
- Exotic and minor pairs. Thinner books mean larger drift on the same order size, while the simulation happily fills the whole amount at one price.
- Breakout rules. Orders fire into fast movement by design, so the requested price and the received price part company more often than elsewhere.
- Larger position sizes. Depth runs out sooner than most traders expect, and a demo never charges you for collecting several prices.
- Rules holding positions for days. The friendliest case, since one round turn of cost spreads across a much larger expected move.
Both Records Still Have Value
None of this makes demo useless. It makes demo good at one job and poor at another.
Use it to prove your mechanics work. Never use it to prove your edge survives contact with real money.
Execution: the Part a Simulator Flatters
Here sits the largest measurable difference. Every live order passes through steps a demo skips entirely.

Reading down that comparison, the demo path ends two steps early. Real orders travel further, and each extra step can cost you a fraction of a pip.
Slippage
Live fills drift away from the requested price when the market moves during transit. Fast conditions make that drift larger.
Demo fills rarely drift at all. Our note on slippage in trading explains when the drift helps and when it hurts.
Requotes and Rejections
Some execution models answer a fast market with a new price rather than a fill. You then accept or decline, and the moment often passes while you decide.
Demo servers usually skip that conversation. Our guide to requotes in forex covers which execution types produce them.
Partial Fills and Depth
Large orders sometimes fill in pieces at different prices. Depth of book runs out, and the remainder fills higher or lower.
Simulated orders take the whole size at one price. That flatters any method trading size that matters.
Scale the test down instead. Trade a tenth of the size on demo and the fills you get there start to resemble live ones.
Latency
Distance from the server costs milliseconds. Milliseconds cost pips when volatility spikes, and a home connection adds more of them than a hosted one.
Demo hides that leg almost entirely. Your order stops at a nearby simulation engine, so the round trip you measure there tells you nothing about the live one.
What Changes the Day You Fund
One thing changes on the broker side the moment real money arrives. Your order now needs a counterparty, and somebody has to carry the other half of it.
That single fact drives the rest of the list. Depth runs out, quotes move while your order travels, and a rejection turns into an ordinary outcome rather than a rare margin error.
None of this makes the firm your opponent. It makes the firm a business with a book to manage, which reads as a plainer story than most forum threads tell.
So ask which execution model your account runs on before you fund it. That answer decides whether a fast market hands you a requote or a worse fill, and the two problems need different fixes.
Where a Live Order Loses Time
- Terminal to broker. The hop from your machine to the broker’s server, which grows with distance and with any wireless link in the chain.
- Broker to liquidity. An extra leg on routed models, where the order joins a queue against quotes that keep updating while it travels.
- Price validation. A check against the current quote rather than the one you clicked, which produces either a worse fill or a requote.
- Fill assembly. The step where a larger order collects several prices, since depth at the best quote runs out sooner than most people expect.
- Confirmation home. The return leg, which decides how late your platform learns what actually happened to the position.
Execution Steps Compared
| Stage | On a demo account | On a live account |
|---|---|---|
| Order leaves the terminal | Travels to the broker’s simulation engine only | Travels to the broker, then often onward to a liquidity provider |
| Price check | Matched against the displayed quote instantly | Matched against a quote that may already have moved |
| Available size | Assumed unlimited at the quoted price | Limited by real depth, so large orders can fill in pieces |
| Rejection risk | Rare, and usually only for margin errors | Possible on price movement, on stale quotes and on thin liquidity |
| Confirmation | Returns almost immediately | Returns after the round trip, with latency added at both ends |
Pricing and Costs a Demo Often Skips
Execution explains half the gap. Pricing explains most of the rest.
Spread on the Demo Feed
Many demo servers run a smoothed or averaged spread. Live spread widens on releases, at the weekly open and through thin hours.
So a rule tested on demo has met the calm version of its own costs. Live conditions then charge it more, exactly when it trades most.
Why a Demo Feed Looks Calmer
Two habits smooth a demo feed. Some servers publish an average spread rather than the live one, and most of them fill from the quote on screen instead of a book.
So a demo shows you the market at rest. Live pricing follows real supply, which thins out at the open, around scheduled data and through the last quiet hour of a session.
Test that yourself in ten minutes. Put a demo chart and a live chart of the same pair side by side, then watch both spreads through a release.
Note the two numbers in your journal afterwards. One reading tells you little, though ten of them across a month give you a figure worth planning around.
Commission and Swap
Some demo accounts leave commission off entirely. Others apply a default rather than the schedule attached to your real tier.
Swap gets the same treatment. Check both against the live contract specification before you trust any simulated result.
Account Tier Mismatch
Demo accounts often open on the firm’s most attractive tier. Your live account may sit on a different one, with different volume steps and a different charging model.
Match the two before you compare anything. Our guide to forex account types shows which settings actually change.
Costs Worth Checking Before You Compare
- Average spread by session. Note the quoted spread during your own trading hours rather than the marketing figure taken from the quietest part of the day.
- Commission per lot, per side. Confirm whether the demo applies the same schedule your live tier uses, since many simulations quietly leave it at zero.
- Swap on both directions. Long and short cost different amounts, and a triple charge lands on one weekday for most instruments.
- Conversion on profit and loss. Applied whenever the quote currency differs from your account base currency, and easy to overlook on a demo funded in a convenient currency.
- Minimum volume step. A demo that allows a hundredth of a lot teaches sizing habits your live tier may not support.
The Same Signals, Two Equity Paths
Take one rule and run it in both places at once. Same entries, same exits, same size.

Two paths leave the same starting point and separate steadily. Slippage and spread create the whole of that gap, since the signals never disagreed.
Why the Gap Rarely Closes
Friction only pushes one way. A favourable fill happens sometimes, yet the spread charges you on every single round turn.
So the gap compounds rather than oscillates. Track it in your own record with our trade journal tool, one line per trade.
Measuring Your Own Gap
Log the price you wanted next to the price you received. Do that for fifty trades and you have a personal slippage figure.
That number beats every general estimate. It reflects your broker, your hours, your instrument and your connection.
Download the complete indicator database
Put these concepts on your charts. One email unlocks the full library of 1,380+ indicators with compiled MT4 and MT5 files, plus my TradingView scripts. No paywall, no spam, unsubscribe any time.
Get free access to my indicator database
One email unlocks 1,380+ free MT4, MT5 and TradingView indicators — the complete library. No single-tool download; you get the whole database.
What a Demo Proves and What It Cannot
Separating those two lists saves months. Most disappointment comes from asking a demo to answer a question it never could.

What It Genuinely Proves
A demo proves your rules produce the trades you expected. It proves the platform behaves, the expert advisor runs, and the alerts fire at sensible moments.
Session errors surface there too. Wrong symbol suffix, wrong timeframe, wrong server time, all cheap lessons on simulated money.
What It Cannot Establish
A demo cannot establish your true cost per trade. Nor can it establish whether you will follow the plan while real money moves.
It also cannot show how the broker behaves under stress. Widening, rejection and delay all live on the live server.
The Question to Ask of Any Simulated Record
Ask what the result would look like after realistic friction. Add a pip of cost per round turn, then look again.
Thin edges vanish under that test. Durable ones survive it with room to spare, which tells you something useful before you fund anything.
The Pressure a Demo Cannot Load
Now for the part nobody can code. Simulated money changes what you do, and it changes it in predictable directions.

Three behaviours show up on live accounts and almost never on demo ones. Each has a measurable cost.
Hesitation
Valid signals go untaken. You wait for confirmation, the price runs, and you enter late or skip the trade entirely.
On demo you take everything, because nothing hurts. So the live record contains a different set of trades from the simulated one.
Size Anxiety
Positions feel larger when the money exists. Traders cut size below plan, which shrinks results, or they move stops closer, which raises the number of stop-outs.
Either adjustment invalidates the comparison. The rule you tested is no longer the rule you traded.
Put a Number on the Behaviour Gap
Behaviour sounds soft until you count it. Two columns turn the whole thing into arithmetic anybody can check.
Column one holds every signal your rule produced. Column two holds the ones you actually took.
Divide the second by the first for your follow-through figure. A trader taking eight signals in ten traded a different rule from the one they tested, whatever the chart shows.
Then look at which two went missing. Traders usually skip straight after a loss, and those skipped trades often carry the same shape as the ones that carry the month.
Fixing that needs no new indicator. Smaller size restores follow-through faster than any tweak to the rule itself.
Interference
Manual closes arrive early. Targets get cut on a whim, losers get held past the stop, and an expert advisor gets switched off mid drawdown.
Read our overview of how brokers actually work if the urge to blame execution keeps returning. Most live divergence starts with the trader, not the server.
How to Use a Demo Properly
Given all that, a demo still earns its place. Give it the jobs it does well.
Four Jobs Worth Giving It
- Learn the platform cold. Order types, pending orders, partial closes, trailing stops and the report tabs, all practised until nothing surprises you.
- Debug an automated rule. Symbol names, lot rounding, session filters and timezone handling all break loudly on demo and cost nothing to fix.
- Rehearse the routine. Pre-trade checklist, journal entry, screenshot and review, run daily until the sequence needs no thought.
- Sanity check a new rule. Confirm the signals appear where you expected before you spend real money finding out.
Set It Up Like the Live Account
Match the balance you will actually fund. A demo carrying a huge simulated balance teaches sizing habits that will ruin a small live account.
Match the leverage, the tier and the instruments too. Every mismatch adds noise to a comparison you will want later.
Keep the Same Records
Journal demo trades exactly as you would journal live ones. Entry reason, screenshot, planned exit and outcome, with no gaps for the embarrassing ones.
That habit transfers. Without it, the live record starts from nothing at the very moment you need evidence most.
What to Log in Both Records
- Requested price and filled price. Two columns rather than one, because the difference between them becomes your personal slippage figure after fifty trades.
- Signal time and fill time. The delay between them explains far more late entries than any complaint about the broker ever will.
- Signals you skipped. Recorded with the reason, since a skipped trade changes the outcome exactly as much as a taken one.
- Spread at the moment of entry. Noted from the platform rather than the website, and worth capturing around scheduled news especially.
- Any manual override. Early closes, moved stops and disabled automation, all flagged so the record still describes the rule you tested.
Automated Rules Deserve the Same Treatment
An expert advisor hides the behaviour problem rather than solving it. Code takes every signal, so the skipped-trade gap disappears.
Cost and execution gaps remain untouched. So run the same program on both accounts for a month, then compare the two trade lists line by line.
Differences surface fast. Same entries at different prices point at execution, while different entries point at a data or settings mismatch.
Moving From Demo to Live Without a Cliff Edge
The jump causes most of the damage. Shrink the jump and the damage shrinks with it.
A Staged Transition
| Stage | What you run | What you learn |
|---|---|---|
| Demo, full size | The rules on simulated money at your planned position size | Whether the mechanics work and the signals appear as expected |
| Live, minimum size | The same rules at the smallest volume your account allows | Real spread, real slippage and your genuine reaction to a loss |
| Live, quarter size | Position size raised to a quarter of the plan | Whether discipline holds once a losing day carries weight |
| Live, full size | The planned size, unchanged from the tested rules | Whether the record still resembles the one you built earlier |
Why Minimum Size Still Teaches Everything
Minimum volume looks pointless to a trader in a hurry. It rarely is, and the reason has little to do with the money at stake.
Every lesson live money teaches arrives at the smallest size. Real spread, real drift, real delay, plus the small jolt of watching a genuine position move against you.
None of those lessons improve at full size. They only hurt more, which slows the learning down rather than speeding it up.
So sit at minimum volume until three things hold steady. You take the signals, you leave the exits alone, and your logged drift stops surprising you.
What to Compare at Each Stage
Compare three numbers only. Average slippage, the share of signals you actually took, and the average time from signal to fill.
Those three explain nearly every divergence. Profit figures explain very little at small sample sizes.
When to Step Back
Step back when you skip signals or override exits. Not because you failed, but because the record has stopped describing the rule.
Drop to the previous stage, rebuild the habit, then move up again. Our library of forex trading strategies assumes you follow the rule you chose.
Common Mistakes When Comparing the Two
Five errors ruin the comparison before it starts. Each has a simple correction.
- Trading a huge demo balance. Position sizing habits built on an unrealistic balance collapse the moment your live account holds a normal amount.
- Ignoring the execution model. Instant and market execution behave differently under stress, and our guide to instant against market execution explains which one requotes.
- Comparing different periods. A calm demo month against a volatile live month measures the calendar rather than the platform or the rule.
- Counting only closed trades. Skipped signals belong in the record too, since a trade you never took still changes the outcome.
- Switching rules after the first losing week. Variance produces losing weeks constantly, and a rule changed mid sample leaves you with two half tested rules.
A Checklist Before You Fund the Live Account
Work through this once. It takes an evening and removes most of the surprises.
| Check | Where to find the answer | Why it matters |
|---|---|---|
| Does the live tier match the demo tier? | The account comparison page, then your account agreement | Volume steps and charging models differ between tiers on most platforms |
| What commission applies per lot? | Contract specifications inside the platform | Simulations frequently leave commission at zero without saying so |
| What swap applies long and short? | Contract specifications, refreshed every few weeks | Overnight costs decide whether a slower method survives at all |
| Which execution model runs on the account? | The account terms, or a direct question to support | Instant execution requotes, market execution slips instead |
| What is the smallest position allowed? | Contract specifications, under minimum volume | Rounding a calculated position upward quietly raises risk on every trade |
| Does a negative balance policy exist? | The written terms for your client category | Gaps can take an account through zero, and policies differ by jurisdiction |
Settings Worth Copying Across
- Starting balance. Matched to the amount you will genuinely fund, since sizing habits built on a fantasy balance transfer straight into the live account.
- Leverage level. Set the same on both, because a different margin requirement changes which positions the platform will even accept.
- Instrument list. The same symbols with the same suffix, which also catches the naming errors that break automated rules on a live server.
- Chart timeframes and templates. Identical layouts, so a signal you read at a glance on demo looks the same when money is involved.
- Trading hours. The same session windows on both accounts, since spread and depth vary enormously between the quiet and busy parts of a day.
Two Numbers to Record on Day One
Write down your average demo slippage and your average demo spread. Both come from the record you already kept.
Then measure the same two numbers live after fifty trades. The difference between the pairs is the honest size of your own gap.
FAQ
How long should I trade a demo before going live?
Long enough to stop making mechanical errors, which usually takes a few weeks rather than a few months. Demo stops teaching once orders, exits and journalling feel automatic, and every extra week after that point delays the only lessons live money can deliver. Move across at minimum volume as soon as the platform holds no surprises.
Why are my demo results so much better than my live ones?
Four causes stack up. Demo fills land on the mark while live fills drift, demo spread often runs smoother than the real feed, commission and swap sometimes go missing from the simulation, and you almost certainly traded differently once real money moved. Measure the first three from your own statements before blaming the fourth.
Can a demo account prove a strategy works?
No, and that limit matters. A demo can show the rules generate the trades you expected, which is a mechanical result rather than an economic one. Cost, execution and behaviour all sit outside the simulation, so a simulated curve tells you the rule runs, never that it earns.
Do brokers deliberately make demo accounts perform better?
The difference comes from architecture rather than intent in most cases. A simulation engine has no counterparty, no depth limit and no rejection logic, so it naturally produces cleaner outcomes than a real routing chain. Worth noting, though: a demo running noticeably different spreads from the live feed deserves a question to support.
How much slippage should I expect once I go live?
No general figure fits, because the answer depends on your broker, your instrument, your hours and your connection. Measure it instead: log the requested price beside the filled price for fifty trades, then take the average. Traders working around scheduled news usually find a much larger number than traders working quiet sessions, which is useful information on its own.
Does a faster connection or a hosted server close the gap?
It closes part of it. Hosting the terminal near the broker’s servers removes some latency, which helps rules that act on fast moves and matters very little to rules holding positions for days. Spread, commission and depth stay exactly where they were, so treat hosting as a small improvement rather than a fix.
Should I keep a demo account after going live?
Yes, for one narrow purpose. Test changes to your rules there rather than on the account holding your money, and use it to check platform updates and new expert advisors before they touch anything real. Keep the two records separate, because merging them destroys the comparison you spent months building. Results are not guaranteed; past performance is not indicative of future results.
External references
- For background on this concept, see Paper Trading on Wikipedia.
- For broader market context, see Fill Ratio in the BabyPips Forexpedia.
