Your EA went live on Monday.
By Friday it's trailing the backtest.
Not blown up. Not broken. Just… worse. A little slower, a little heavier, a little less like the chart that convinced you.
So you do what everyone does.
You start doubting the EA.
Wrong suspect.
The live account isn't underperforming the backtest. The backtest was overperforming reality — and it always does.The gap you're staring at has a structure. It has known causes, a normal size, and — this is the part nobody selling EAs wants you to learn — it has a use. The gap is the single most honest number in this entire industry, because it's the one number a marketer can't produce. Only live trading produces it.
Two requests before we start:
- Save this for the day your next EA goes live and drifts under its backtest.
- Send it to one trader who's about to judge a system by week two.
Skip this if you already know why live always runs under the backtest
Sixty-second version: a backtest is one historical path traded at modelled costs. A live account is one new path traded at paid costs. Costs are always modelled optimistically (not dishonestly — structurally: the model can't know next month's spread around news). The historical window is one particular mix of regimes; your live months are a different mix. And you start on one specific day, which makes your first quarter a lottery draw from the distribution the backtest quietly averaged over. Add it up and live-runs-slightly-worse isn't a defect. It's the expected value.
| What traders think | What's actually true |
|---|---|
| "Live is worse — the EA is broken" | Live slightly worse is the expected result of paid costs + one path |
| "A good backtest means a good EA" | A good backtest means one path looked good at modelled costs |
| "The gap should be zero" | A zero gap over months is a reason for more suspicion, not less |
| "Week two tells you something" | Week two is noise; the gap is a quarters-scale measurement |
| "Sellers with perfect curves are confident" | Sellers who quantify their own gap are the confident ones |
I — The gap is arithmetic, not betrayal
A backtest and a live account are not two attempts at the same thing. They are two different measurements. One replays the past with a cost model. The other buys the present with real fills. The difference between them decomposes into terms you can name — and once you can name them, "the EA stopped working" stops being your first theory.
We audited the report itself — the KPI grid, the screenshots, the cropped drawdowns — in Backtest Isn't Reality. This piece is about the thing the report can't hide even when it's honest: the gap.
Backtest = one path × modelled costs. Live = one new path × paid costs. The gap is baked into the definitions.II — The five layers inside it
Cost modelling error. Every backtest charges you an assumed spread and commission. Live charges you the actual one — including the ugly minutes: news, rollover, thin Asian hours. The model is a flat tax; reality is a progressive one that spikes exactly when your EA is most active, if your EA trades events or opens.
Slippage asymmetry. Backtests that model slippage at all usually model it symmetric — as likely to help as to hurt. Live fills skew against you. That skew is a permanent, one-directional slice of the gap, and no amount of optimization on the historical window can price it in.
Spread variance. An average spread in the model, a distribution of spreads in life. Strategies with tight stops or small targets pay this difference on every trade — it compounds silently.
Regime sampling. The backtest window is one particular blend of trend, chop, panic, and drift. Your live months are a different blend. Even a robust edge earns unevenly across regimes; your first live quarter samples whichever regime happens to be running.
Execution reality. Latency, requotes, partial fills, a broker's server hiccup at the worst possible second. Individually negligible. Summed over hundreds of trades, a real term in the gap.
None of these five is scandalous. All five are absent from most screenshots.
III — Start-date luck: you live one path, the backtest averaged many
Here's the layer traders feel most and understand least. A multi-year backtest doesn't show you your own next quarter. It shows the average texture of one long path. When you go live, you don't get the average — you get whatever three months you happened to start in. Start ahead of a drawdown the backtest recovered from, and your live account looks broken while behaving exactly to spec.
This is precisely what a Monte-Carlo view is for: shuffle the same trade population into thousands of orderings and watch how wide the honest range of "first quarters" really is. We built a free, in-browser simulator that does exactly this with your own win rate, reward multiple, and risk — no signup, nothing uploaded. Run it once and you will never judge an EA by week two again. The backtest showed you one path. Your live account is one different path. Variance between paths is not evidence of failure — it's the definition of variance.
IV — What a normal gap looks like
We won't hand you a magic universal number, because anyone who does is selling something. The honest shape is this: expectancy a notch under the modelled figure, drawdowns a notch deeper and longer than the single backtest path suggested, and trade-by-trade costs a consistent sliver worse — with the size of "a notch" scaling with how cost-sensitive the strategy is. Tight-stop scalpers live and die by the cost layers; slower systems feel regime sampling more than spread. What normal never looks like: a live account that mirrors the backtest tick for tick, or one that collapses to half its modelled expectancy the moment real money arrives. The first is too good to be real; the second means the backtest's cost model wasn't a model, it was a wish.
Want a number anyway? Then get your number, not a universal one: feed your win rate, reward multiple, and risk into the Monte-Carlo simulator and read the spread of outcomes it draws. That fan — not any seller's promise, ours included — is what "normal" looks like for your profile.
V — How long before you can judge?
Not in weeks. In trades.
Week two of a live account is usually five to twenty trades. At that sample size, the difference between a working edge and a broken one is statistically invisible — the error bars on your expectancy estimate are wider than the expectancy itself. Flip a fair coin twenty times and it will hand you streaks that feel like proof of something. It isn't proof. It's twenty flips.
A working rule — a discipline, not a law: give a system on the order of a hundred trades, or a full quarter, whichever comes later, before you compare live expectancy to modelled expectancy. And when you compare drawdown, compare it to the Monte-Carlo fan of paths, not to the single backtest line — the fan is what "on spec" actually looks like. If after that window the gap is still the size only a broken premise explains, stop trading it. Judging too early and never judging at all are both expensive; the window is what separates them.
Week two is a coin-flip sample. Judge in trades, not calendar days.VI — The red flags — including live doing too well
A live account running better than its backtest for months is also a tell — usually of a backtest tuned to look conservative for marketing, a survivor account cherry-picked from several, or a regime hot streak being sold as skill. Suspicious in both directions is the correct posture. The rest of the red-flag list writes itself once you hold the five layers: sellers who show a backtest but no live-vs-backtest comparison at all; "live" results with no dates, no costs, no broker; zero-gap claims; gaps explained by anything except the five layers ("the broker manipulated my fills" as a permanent excuse). Everyone's live results trail their backtest. The only people with no gap are people with no live results.
VII — The inversion: the gap is the tell
Now flip the whole thing. If the gap is universal, then a seller who shows you their gap — who publishes the backtest as modelled, labels it modelled, and talks plainly about what live degradation to expect — is handing you the one signal marketing can't counterfeit. It costs a seller nothing to show you a perfect curve. It costs them real sales to teach you why perfection is impossible. Perfection is a marketing decision. A quantified gap is an engineering decision. Buy from engineers.
VIII — The 20-minute track-record interrogation
Minutes 0–5 — Name the costs. Find what spread/commission/slippage the backtest charged. Can't find it? That's your answer.
Minutes 5–10 — Demand both curves. Backtest and live, same instrument, dated. No live at all is honest for a new system if it's labeled as such. Live that can't be compared to the backtest is a costume.
Minutes 10–15 — Size the gap. Expectancy vs expectancy, drawdown vs drawdown. Is the difference the size the five layers explain — or the size only a broken premise explains?
Minutes 15–20 — Interrogate the explanation. Ask the seller: "What gap should I expect between your backtest and my live account, and why?" Engineers answer with layers. Marketers answer with reassurance.
IX — What actually narrows it
Nothing closes the gap. Three things narrow it. Cost-first modelling: price the backtest at your ugliest realistic spread, not your broker's brochure spread — and let modelled slippage lean against you, the way live slippage will. Out-of-sample discipline: an edge that only exists in the tuned window will donate its entire gap to you live. Execution engineering: fewer, cleaner entries at hours where the cost distribution is tame. Everything else — more indicators, more filters, more optimization passes against the same window — widens the live gap while narrowing the backtest one, which is the exact opposite of what you paid for.
Where MTR sits
The honest ledger, because this article is also our own standard. MTR's published results are a ~29-month modelled backtest and they are labeled exactly that, on the page, in the open. We publish the method, the costs charged — cross-checked in a Rust tick simulator — and the full source: 16,923 lines you can read, because the only cure for the gap conversation is letting you inspect the premises yourself. What we will not publish is a promise about your live gap: your broker, your costs, your start date, your path. Evidence, not prophecy. Run the Monte-Carlo simulator with MTR's published trade profile or your own numbers and look at the fan of first quarters before you form an expectation — that fan is the honest forecast.
Disclosure
This is education about a measurement, not a promise about returns. Every number above is illustrative, every mechanism is checkable, and the one question worth asking any seller — including us — is the Minute-15 question: what gap should I expect, and why? If the answer isn't layers, walk.
Your first 20 minutes with the source
Open the risk module first. Find where costs enter the logic. Find the spread check before entries. Then open the backtest evidence page and read what was charged. If you can trace the cost premise from source to result in twenty minutes, you now know more about your gap than most buyers ever learn about theirs.
The backtest was never a promise.
It was a hypothesis, priced optimistically.
Live is the experiment.
The gap isn't where systems fail. It's where sellers do.
