Bid, Ask and Spread: The Three Prices Your Backtest Cannot Ignore

Look at a chart and you may feel as if the market has one price.

It does not.

At any tradable moment there is usually a difference between what buyers are bidding and what sellers are asking.

That gap is the spread.

For many strategies it is just a cost.

For others it is part of the strategy's logic.

Bid and Ask in plain English

Suppose a quote is:

Bid: 2345.20 Ask: 2345.35

The spread is 0.15.

A market buy interacts with the Ask side.

A market sell interacts with the Bid side.

That means the visible chart level and the executable level are not necessarily identical.

This seems obvious when written down.

It is surprisingly easy to forget when staring at a historical candle.

Why a Bid-only view can be misleading

Imagine a strategy that buys when price crosses a threshold and uses a very tight stop.

If your analysis looks only at Bid candles, you may conclude that price never reached a particular Ask level.

Or a spread expansion may cause an executable side of the market to touch a level while the chart appears relatively calm.

The narrower the expected edge of a strategy, the more important this becomes.

Spread is not always constant

One of the advantages of real Bid/Ask tick history is the ability to preserve spread changes over time.

Spreads can widen around:

• low-liquidity periods,
• rollover,
• major economic releases,
• market openings,
• sudden volatility.

MetaTrader explicitly distinguishes this in its testing documentation.

With real ticks, spread can change within the minute.

With generated Every Tick data, the tester uses the spread associated with the relevant minute bar when deriving Ask from Bid.

Those environments can produce different results.

A simple example

Strategy A aims to capture an average move of 100 points.

Its average spread is 5 points.

Spread matters, but the strategy has room.

Strategy B aims to capture 8 points.

Its average spread is 5 points.

Spread is now one of the dominant variables in the entire business model.

Both strategies can use the same chart.

They do not have the same data requirements.

Historical spread is part of historical market state

This is why replacing the original Bid/Ask relationship with an arbitrary fixed spread can be dangerous.

A test may become:

• too optimistic during historically expensive periods,
• too pessimistic during historically tight periods,
• incapable of reproducing spread filters,
• inaccurate around trade triggers.

For some research a fixed-spread assumption is acceptable.

But it should be an explicit assumption.

Last price is another concept

Exchange instruments can also use the Last traded price.

MetaTrader's tick structure can contain Bid, Ask and Last simultaneously.

Charts for some instruments may be formed from Last trades rather than Bid.

Yet MetaTrader notes that trading operations still use Bid and Ask.

That distinction is particularly important when moving between Forex, CFDs, futures and exchange-traded instruments.

What good tick history should preserve

For quote-driven instruments, a useful historical tick representation should preserve at least:

• precise timestamp,
• Bid,
• Ask.

Where meaningful and available, it may also include:

• Last,
• volume,
• tick flags.

Throwing Ask away because “the chart uses Bid” removes information that can directly affect trading outcomes.

A better question for every backtest

Before trusting the result, ask:

Which price generated my signal?

Then:

Which price triggered my order?

Then:

Which spread did the simulation use?

If those answers are unclear, the backtest still has an unresolved assumption.