Mechanism
Latency is the delay between decision and fill. In fast markets, that delay is the difference between ‘edge’ and ‘chase’.
Reactive strategies (entering after the move starts) are most exposed. By the time you enter, the good liquidity is gone and you’re paying the worst price.
Latency doesn’t show up as a single error. It shows up as a slow bleed: slightly worse entries, slightly more slippage, slightly worse expectancy.
Model risk: small parameter changes that flip your results are a warning, not a feature. Robustness is a survival requirement.
- Look for wide plateaus, not sharp peaks.
- Measure drawdown shape, not just final equity.
- Assume the future will be different in the exact way that hurts.
How it kills accounts
React to move → late fill → tight stop (because entry is worse) → stop-out → re-enter (because move continues) → repeat until you’ve paid the whole move in friction.
How it kills accounts:
- Model works on clean history.
- Regime changes and execution friction increases.
- Performance decays slowly, so you rationalize.
- You optimize parameters instead of reducing risk.
- Drawdown becomes the teacher.
Rule that survives
If you’re trading reactions, trade smaller.
Prefer pre-planned levels over impulse entries.
Measure your ‘decision-to-fill’ delay. If you can’t, assume it’s worse than you think and trade accordingly.
Rule that survives:
- Stress test tails and execution, not averages.
- Prefer robust plateaus over optimized peaks.
- When performance decays, reduce risk before “fixing” the model.
Example archetype
You buy a breakout after the candle breaks. Your entry is 5 pips worse than the level you ‘saw’. Your stop is now too close.
You didn’t trade the breakout. You traded the post-breakout cleanup.
Tell: if small parameter tweaks flip your results, your model is fragile.
Deep dive
Upgrade path: if latency is unavoidable, your edge must be slow enough to survive it. That means longer holds, wider stops, and smaller size.
Glossary: latency, slippage, markout.
Variants merged
This page consolidates closely-related entries into one stronger canonical reference. Retired versions now redirect here.
Variant merged: Latency Turns Clicks Into Bets
Why it’s included: Variant emphasis: reactive discretionary clicking turns into a latency lottery. Systems with pre-defined triggers degrade less because they don’t “chase the last tick.”
Truth line: If your edge requires perfect timing, your broker and route are part of the edge.
Mechanism add-on: Orders travel: device → platform → broker → liquidity → back. In fast markets, the best price can disappear while your order is in transit.
Failure add-on: Timing-based entry → latency slip → immediate drawdown → emotional management → early exit → you “prove” your edge doesn’t work and start strategy hopping.
Rule add-on: Avoid strategies that require single-tick precision unless your execution stack supports it. Prefer setups with fat targets relative to typical slippage.
Variant merged: Latency Hurts Discretion More Than Systems
Why it’s included: Variant emphasis: reactive discretionary clicking turns into a latency lottery. Systems with pre-defined triggers degrade less because they don’t “chase the last tick.”
Truth line: If you click on a feeling, latency will price you like a tourist.
Mechanism add-on: Discretion often triggers when the move already looks obvious. Add latency and you’re structurally late.
Failure add-on: You see the move, feel urgency, and click market.Your order arrives after the best quotes moved.You start the trade with instant slippage and a tighter real stop.You compensate by widening stops or adding size, creating a loop of late entries.
Rule add-on: For discretionary entries, prefer limits at pre-planned levels.If you miss it, you miss it. Missed trades are cheaper than bad fills.Use a ‘latency tax’ assumption in your risk: if you need precision, don’t click at market.
Field checklist
- Stress-test the tails. The worst days define survival.
- Use variable spreads and slippage in testing.
- Prefer robust plateaus over optimized peaks.
- Look at drawdown shape, not only profit.
- If a tiny parameter change breaks the model, the model is fragile.