Slippage Modelling Backtest: Why Your Results Are Optimistic

Every backtest lies a little, and slippage is the biggest liar. Slippage modelling backtest settings decide how much of the difference between the tested equity curve and the live one your results reveal in advance. This post covers how backtesters handle execution, typical slippage ranges per instrument, why scalping experts suffer most, and how to calibrate MT4 or MT5 tests honestly before live money is exposed.

Why slippage modelling backtest results are optimistic

A backtester plays a recording of the past. Every price in that recording is a quote that existed, and the tester lets your EA trade at it — or at the next bar’s open, which is even kinder. No algorithm waits on the other side of your order, no broker routes it, and no spread is crossed. A live order fills at the ask when buying, moves through the spread, and is delayed by milliseconds.

The gap between those two realities is slippage, and it is a cost, not a curiosity. The optimistic bias of slippage modelling backtest assumptions shows up quietly: every trade slightly better than it will be live, every drawdown slightly shallower. They are not dramatic on one trade, but they compound across thousands. The backtesting guide covers the other assumptions that hide in a tester, from spread handling to bar selection.

How slippage modelling backtest execution works: perfect fills versus friction

Testers offer three levels of execution honesty. At the bottom, “every tick” modes use the recorded tick stream and are the most accurate — but still fill at the recorded price with no slippage unless you add it. In the middle, “open prices only” fills at the next bar’s open, overstating fills by design. At the top, custom slippage settings deduct points from every entry and add them to every exit.

That last approach is the correct foundation: set the slippage parameter on the backtest tab and every order is adjusted. Spreads are a separate setting — current symbol spread or a fixed one — because spread and slippage are two parts of the same friction bill. The spreads and slippage guide explains how brokers quote the two and how they behave across sessions.

Typical slippage ranges per instrument

Slippage is not a number but a distribution, depending on the instrument and the hour. During London and New York trading in the majors, EURUSD, GBPUSD and USDJPY typically slip 0.1 to 0.2 points, with spreads in the 0.5 to 1.0 point range. Crosses such as EURGBP or AUDJPY add a little more: 0.2 to 0.4 points. Exotic pairs, thinly traded crosses, minor indices and smaller cryptos carry the widest spreads and the largest slippage — 0.5 points or more in normal conditions, several times that on news.

Session matters as much as instrument. The Asian session on the majors is thinner and wider, and news events produce slippage spikes that no average setting captures. The honest approach in a slippage modelling backtest is to test twice: a mild level for the average day, a worst case for news and thin sessions. The ESMA risk analysis documents how execution conditions vary across venues and sessions.

Why scalping EAs are most affected by slippage modelling backtest errors

Scalping strategies trade dozens of times per day for a small average profit. A scalping EA targeting ten points on a major, with a two-point spread and one point of slippage per round trip, pays thirty per cent of its target in friction before it starts. In a backtest with zero slippage that cost is invisible; live, the same settings produce an equity curve that slopes down instead of up.

The fix is to make the cost explicit before judging: model the spread from your broker’s published average, add a per-side slippage figure, and ask whether the profit target still clears the hurdle. If it only clears it when friction is zero, it is a fantasy — which is why the FCA’s CFD guidance warns retail traders to examine costs before committing capital.

How to add slippage to MT4 and MT5 tests

In MetaTrader 4, the strategy tester’s settings tab offers a slippage field in points; set it alongside the spread choice and run the test twice, with mild and worst-case values. MetaTrader 5 exposes the same concept, with the option of testing against real tick data when your broker supplies it. Record the slippage assumptions in the test report, because a result that does not state its friction cannot be compared with live trading.

Calibrate from your own account, not from averages. Run a week of demo trades with your exact EA and broker and measure the difference between requested and actual fills; that measured figure, not a guessed one, goes into the tester. Recalibrate when you change brokers or session focus — the CFTC’s investor education material emphasises that execution quality varies by provider.

Honest calibration: accept the friction

Honest calibration has one non-negotiable rule: if the strategy cannot survive realistic slippage in the tester, it is not a strategy you should run. Adjusting the slippage setting downwards until the backtest turns green is the same lie as leaving it at zero. The point of slippage modelling backtest settings is accuracy, not flattery.

The discipline pays for itself in one number: the difference between the tested equity curve and the live one. Model friction honestly and the difference shrinks to the unmodellable noise of real markets; leave it at zero and the difference is exactly the cost of finding out the hard way.

Frequently asked questions about slippage modelling backtests

Why does slippage make backtest results optimistic?

Backtesters execute at the recorded price or the open of the next bar, which never includes the cost of a real order moving through a live spread. Every backtested trade is therefore a little better than the same trade live; the gap compounds. Accurate slippage modelling backtest settings deduct a realistic per-trade cost so the equity curve shows the strategy as it will run, not as it would in a frictionless world.

What slippage should I add in MT4 and MT5?

Start with the spreads your broker publishes for the session you trade, add 0.1 to 0.2 points of slippage on majors during normal hours, and 0.5 points or more across news events. Backtest at two levels: a mild setting for the average day and a worst case for news days. If the strategy survives the worst case, the mild case is a bonus.

Which instruments slip the most?

Exotic pairs, thinly traded crosses, minor indices and smaller cryptos carry the widest spreads and the largest slippage, especially outside their primary sessions. Majors in London and New York hours are the cheapest and most predictable. Accurate slippage modelling backtest settings matter most where slippage is largest, because that is where tested and live results diverge.

Why are scalping EAs most affected by slippage?

Scalpers trade many times per day for a small average profit, so a per-trade cost the size of the profit target removes most of the edge. Backtests without slippage model this cost at zero, producing a strategy that looks robust and loses live. Slippage and spread modelling force the profit target to survive realistic friction.

Test with friction before you trust the curve. Add realistic slippage and spread settings to every backtest and compare the result against live fills — the automation playbook shows how to validate EAs properly before deployment.

Risk disclosure: Trading foreign exchange, commodities, CFDs and cryptocurrencies carries a high level of risk and may not be suitable for all investors. Past performance is not indicative of future results. AlgoTM provides trading tools and technology only and does not provide investment advice, portfolio management or guaranteed returns.