Your EA Entered at the Ask. Your Chart Pretended It Didn't.

Your EA Entered at the Ask. Your Chart Pretended It Didn't.

18 November 2026, 00:00
Lee Kai Lun
0
8

Imagine this entirely hypothetical scene: an EA opens a EURUSD Buy. The chart looks perfectly calm. The position is already negative.

The robot has been employed for roughly one second and apparently needs a performance review.

Before interrogating the strategy, look at two prices instead of one.

Your chart and your order may be watching different prices

Suppose EURUSD is quoted at:

  • Bid: 1.10000
  • Ask: 1.10018

The spread is 0.00018, or 1.8 pips for this conventional five-decimal EURUSD quote.

A market Buy normally executes at the Ask. Its immediate liquidation value is based on the Bid. A market Sell normally executes at the Bid, and buying back to close uses the Ask. With no price movement, the Buy is already facing an 1.8-pip gap, before any separately charged commission.

If the chart displays Bid candles, it might show 1.10000 while your Buy fills around 1.10018. Neither number is necessarily wrong. They answer different questions.

That distinction matters enormously for short-target EAs, stop placement and debugging apparently mysterious entries.

Why does the gap suddenly grow?

A spread is not a permanent entrance fee printed on the wall.

Liquidity may thin around session transitions. Important economic announcements can make quotes change rapidly. Liquidity providers and brokers can adjust their quotes when market conditions become uncertain. Different accounts can also have different spreads and commissions.

None of those observations proves that a particular broker manipulated a trade. A single screenshot of a wide spread is evidence of a quote at one moment, not evidence of its cause.

And beware the reassuring phrase "the spread is usually tiny." Your EA does not trade the word usually. It trades the actual quote available when its rules fire.

A five-pip target meets a three-pip spread

Consider a simplified hypothetical strategy targeting a five-pip favorable move from its entry price.

If the spread is one pip, that is a different cost environment from a three-pip spread. With all else equal, the larger spread leaves less room for the same gross market movement to translate into net profit.

But do not subtract spreads mechanically from every reported trade result: it depends on whether your target is measured from entry, Bid, Ask or the chart, and how the order is actually closed. Commission and slippage can change the final result too.

This is why "my backtest had a fantastic win rate" is not enough. What spread assumptions were used? What tick data? What commission? What execution conditions? A beautiful equity curve can be an excellent work of fiction if the cost model is unrealistic.

Three different costs that traders keep mixing up

Spread is the simultaneous difference between the quoted Bid and Ask.

Slippage is the difference between an expected or requested execution price and the actual fill, where the order type and execution model permit it.

Commission is a separately charged fee on accounts that use that pricing model.

They can occur together, but they are not synonyms. A spread filter is not a slippage guarantee, and a low displayed spread does not establish a low all-in cost.

Give your EA a proper spread check

In MQL5, you can obtain a current quote using SymbolInfoTick() and compare the two sides:

MqlTick tick;
if(!SymbolInfoTick(_Symbol, tick))
   return;

double point = SymbolInfoDouble(_Symbol, SYMBOL_POINT);
if(point <= 0.0 || tick.ask <= 0.0 || tick.bid <= 0.0)
   return;

double spread_points = (tick.ask - tick.bid) / point;
Print("Current spread in points: ", spread_points);

This snippet measures the quote. It does not place orders, and it does not guarantee a fill at those prices.

Notice the word points. Points and pips are not interchangeable across every instrument. For XAUUSD and other CFDs, check the symbol's point, tick size and contract specifications. Never assume that a threshold copied from EURUSD makes sense on gold.

A maximum-spread rule might block new entries above a chosen threshold. It will not retroactively improve an open position, eliminate slippage, or ensure that the quote remains unchanged between checking it and executing an order.

What to investigate before blaming the EA

  1. Display Bid and Ask in Market Watch and observe both during the problem period.
  2. Inspect the symbol specification, trading sessions, point and tick size.
  3. Check whether the EA uses a maximum spread setting and whether that setting is in points or pips.
  4. Read the Experts and Journal logs. Look for actual trade-server responses, not guesses.
  5. Compare costs during active sessions, quieter periods and scheduled announcements.
  6. Test with realistic available tick data and trading costs, then forward-test in a demo environment.

For an EA that appears to open at the "wrong" price, first establish which side of the quote the order uses. For one that refuses to enter, inspect its filters and trade permissions rather than automatically blaming spread.

The takeaway

A chart candle is not a complete execution report. Your EA interacts with Bid, Ask, symbol rules and actual fills, not merely a pretty line on a screen.

Before rewriting a strategy because its live trades look worse than expected, measure the trading conditions it actually experienced.

The robot might be following your instructions perfectly. The instructions might simply have forgotten that buying and selling are two different prices.