Not all hours are equal in forex. A signal that performs well during the London open can fail badly at 3am UTC. This isn't opinion — it's measurable. And if your forex signal provider isn't accounting for session timing in their methodology, you're getting half a model.
Here's what we know, what our data shows, and how to think about this properly.
---
The Four Sessions and Why They're Not Interchangeable
Forex runs 24 hours a day, five days a week. But liquidity — the volume of buyers and sellers at any given moment — isn't evenly distributed. It clusters into four overlapping windows:
- Sydney (20:00–05:00 UTC)
- Tokyo (00:00–09:00 UTC)
- London (07:00–16:00 UTC)
- New York (12:00–21:00 UTC)
The London–New York overlap (12:00–16:00 UTC) is where roughly 70% of daily forex volume concentrates. Tight spreads. Fast execution. Higher signal reliability — at least in theory.
Off-session hours (roughly 21:00–00:00 UTC) are the opposite. Thin liquidity. Wider spreads. Price moves that look meaningful but often aren't. In our data, signals fired during off-session hours had 14 percentage points lower win rate than those fired during the London–New York overlap. That gap is too large to ignore.
---
How Session Timing Affects Signal Performance
This isn't about the signal being "wrong." It's about the market conditions the signal is entering.
Algorithmic trading signals — including ours — are trained on price behaviour. That behaviour changes depending on who's active in the market. When European institutional desks are running, momentum follows through. When they're closed, the same setup often goes nowhere.
Three things shift materially between sessions:
Spread cost. A 2-pip spread on EURUSD during London is manageable. The same pair at 22:00 UTC might be 4–6 pips. That alone wipes out scalp signal edge. Scalping (short-duration trades targeting small price moves) is particularly sensitive to this.
Volume-based confirmation. Most signal models — ours included — use volume or volatility filters as confirmation layers. Off-session volume is low and noisy. Confirmations that would be valid at 09:00 UTC are structurally weaker at 23:00 UTC.
News asymmetry. Major economic data releases cluster during active sessions. Signals near those events either benefit from the move or get stopped out by the spike. Models that don't filter for high-impact news windows during off-hours will fire into gaps that don't resolve cleanly.
---
What the Session Data Actually Looks Like
Here's a simplified breakdown of typical signal performance characteristics by session. This is based on structural factors — not a guarantee of future results.
| Session | Typical Spreads | Volatility Level | Scalp Signal Fit | Swing Signal Fit |
|---|---|---|---|---|
| Sydney (20:00–05:00 UTC) | Medium–Wide | Low | Poor | Moderate |
| Tokyo (00:00–09:00 UTC) | Medium | Moderate | Moderate (JPY pairs) | Moderate |
| London (07:00–16:00 UTC) | Tight | High | Strong | Strong |
| New York (12:00–21:00 UTC) | Tight | High | Strong | Strong |
| London–NY Overlap (12:00–16:00 UTC) | Tightest | Highest | Strongest | Strong |
| Off-session (21:00–00:00 UTC) | Wide | Very Low | Very Poor | Weak |
Tokyo is worth singling out. It's a legitimate session for JPY pairs — USDJPY, EURJPY, AUDJPY. The session timing effect is pair-specific, not just global. A model that treats all forex pairs identically across all hours is, bluntly, not a well-built model.
---
How We Handle This at Trade Forge
Our models don't fire signals indiscriminately across all 24 hours. We apply session filters that adjust signal thresholds based on the active window.
During off-session hours, the minimum confidence threshold (the score a setup must hit before a signal is issued) is raised. Fewer signals fire. That's intentional. A lower volume of higher-quality signals is the goal — not signal count for its own sake.
We also track win rate and R-multiples (R-multiple: how much you made relative to what you risked — the only number that matters in evaluating a signal) by session, by pair, and by signal style. Our calibration engine retrains on live outcomes weekly. Session timing is one of the primary stratification variables in that process.
Most signal providers don't publish this level of breakdown. If your signal provider won't show you session-stratified performance, that IS the track record.
---
Who Doesn't Need to Think About This
Session timing matters most to active, short-duration traders. If you're trading daily or weekly swing setups, the London open versus the Sydney open is largely irrelevant. Your hold time is long enough that session dynamics wash out.
This also doesn't apply if you're trading assets with different liquidity profiles — some crypto markets, for example, run more consistently across 24 hours than major forex pairs do. The session effect is a forex-specific structural phenomenon.
And if you're not trading at all — if you're here to understand how algorithmic trading signals work before committing any capital — the session timing nuance is useful context, but not an immediate priority. Understanding position sizing and signal calibration (Platt calibration — the technique that turns raw model scores into actual probabilities) matters more at that stage.
---
FAQ
Does session timing matter for best forex signals 2026?
Yes. Any forex signal provider worth using should account for session conditions in signal generation and performance reporting. If they don't, their historical results likely include off-session noise that inflates apparent performance.
Should I only trade during the London–New York overlap?
Not necessarily. The overlap has the best structural conditions for scalp and day trade signals. But swing signals and specific pairs (JPY crosses during Tokyo, for example) can perform well outside that window. It depends on your style and the pairs you trade.
How do I know if my signal provider accounts for session timing?
Ask them for win rate breakdowns by session. If they can't provide that, the data either doesn't exist or they don't want you to see it. Either answer tells you something.
Do Trade Forge signals specify what session they're designed for?
Yes. Every signal we issue is tagged by style (scalp, day trade, or swing) and asset class. The session filters are embedded in the model logic — not left to you to figure out.
---
If you want to see how session-filtered algorithmic trading signals look in practice — entries, targets, stop levels, and live outcomes — our free Telegram channel posts the same signals as VIP with a short delay. No obligation. Just the data. Find us at tradingforge.io.
---
This article is for educational purposes only. It does not constitute financial advice. Trading involves risk of loss.