High Frequency vs Low Frequency Strategy Choice

Frequency is not a style preference — it is a structural choice that determines your cost structure, infrastructure budget, edge source, and capacity. Mismatch the strategy to the frequency and the edge evaporates into fees.

The frequency spectrum

• HFT (microseconds to seconds): market making, latency arbitrage, queue position. Holds seconds or less.
• Intraday (minutes to hours): momentum, mean reversion, news reactions. Flat by close.
• Swing / medium (days to weeks): trend following, cross-asset, macro. Holds through multiple sessions.

What each frequency demands

HFT rewards infrastructure. Edge lives in microseconds, so co-location ($1k–$10k/month), FPGA, and direct market access are table stakes. Costs: exchange fees, colo, data feeds total $50k–$500k annually. Capacity is tiny — most HFT strategies saturate below $5M. Retail cannot compete here; the edge is pure latency.

Intraday rewards signal quality and execution. Round-trip cost of 0.05% (5 bps) on 20 trades/day = 200% annual turnover drag. If your gross edge is 8 bps per trade, costs eat 60% of it. Capacity is moderate, $1M–$50M per strategy.

Low frequency rewards patience and risk sizing. Costs are negligible (10–50 trades/year), so the edge survives. Capacity is huge — trend funds run billions. The cost is psychological: weeks of drawdown between signals.

Decision framework

Ask four questions:

1. What is your gross edge per trade? Subtract 2× realistic costs. If the result is under 3 bps, only low frequency works.
2. What infrastructure can you fund? Under $10k/year data and compute rules out HFT entirely.
3. What is the edge's decay half-life? Edges that decay in hours cannot be exploited at weekly frequency; edges that decay in months are wasted intraday.
4. What capacity do you need? Target $100k personal capital and HFT is irrelevant; target $50M and intraday starts saturating.

The cost trap

Traders underestimate cost by 2–3×. A 0.1% fee looks small until you trade 300 times a year — that is 30% of capital in fees alone, before slippage. The formula: annual cost drag = round-trip cost × trades per year. Run this number before choosing frequency.

The realistic retail answer

For capital under $1M and infrastructure under $20k/year, intraday (5–20 trades/week) and swing (1–5 trades/week) are the only viable frequencies. HFT is a technology business, not a trading strategy. The lowest-frequency strategy you can stand psychologically usually has the best risk-adjusted return because costs compound against you, not for you.