Direct answer
There is no single, universally correct number of days or candles for “how far back” to look in forex. The right lookback depends on what you are trying to verify, especially when the topic is execution behavior associated with last look. A useful approach is to pick a horizon that is long enough to observe repeated patterns, but short enough that market conditions are still comparable.
If your goal is to understand execution behavior, you generally need enough history to see how outcomes repeat under similar circumstances. If your goal is to understand general market behavior around price, you can use standard technical time horizons (for example, short-, medium-, and long-term windows) and then cross-check that conclusions persist when the window changes.
Mechanics: what “look back” means in this context
In forex analysis, “looking back” means choosing a time window over which you collect and compare price and execution-related observations. The window length affects what you can reliably say because it changes three practical factors:
- Sample size: A longer window includes more observations, which can reduce the chance you are seeing one-off events.
- Regime relevance: Market dynamics can shift over time. A very long window may include conditions that no longer reflect the current environment.
- Execution framing: When last look is involved, the provider’s execution process is part of the story. You cannot infer provider-specific behavior from price history alone; you typically need to align your time window with what can be observed and verified from your own records.
A workable definition is: choose the shortest window that still provides enough repetitions for your verification goal, then test whether key observations hold when you expand or contract the window.
Example or checks
Here are independent checks you can run without assuming a predetermined “correct” horizon:
- Window-stability check: Repeat your analysis on two or three different lookbacks (for example, a shorter window and a longer one). If a conclusion appears only in one window length, it may be fragile.
- Condition-matching check: Compare periods that are meaningfully similar (for instance, comparable liquidity or volatility conditions). If the behavior changes dramatically between unmatched periods, extend your reasoning to those differences instead of averaging them away.
- Observable vs non-observable separation: Price-only history can show what happened to price, but last look mechanics are tied to provider execution decisions. Therefore, verify last-look-related assumptions by focusing on outcomes you can observe (such as whether submitted prices translate consistently into executed results in similar situations), and avoid treating price movement as proof of a provider rule.
For “how far back,” a common practical compromise is to use multiple horizons and focus on whether your observations remain consistent rather than searching for a single magic window.
Limitations: uncertainty, risks, and what to verify
Because forex markets and execution processes change, any fixed lookback length risks being wrong for a specific verification goal. Key limitations include:
- No guaranteed inference: A longer history does not guarantee that past conditions predict future behavior.
- Provider-specific execution: Last look behavior depends on how a specific provider implements it. Without documented terms and observable outcome alignment, you should treat assumptions as tentative.
- Context loss: Very old data can reflect different market structure, different liquidity, or different execution environments.
To stay verifiable, state your lookback choice explicitly, justify it as matching your verification goal (sample size and relevance), and report whether observations are stable across window changes.