Commission Models by Account Type

The Two Account Shapes

Spread-only (standard) accounts roll all cost into the spread: EUR/USD at ~1.0–1.2 pips, nothing else charged. Simple, and what most beginners trade.

Raw-spread / ECN accounts pass through near-interbank spreads (EUR/USD often 0.0–0.3 pips) and charge an explicit commission, typically $3.00–3.50 per standard lot per side — about $6–7 per round trip, which on EUR/USD converts to roughly 0.6–0.7 pips. Total cost ≈ 0.7–1.0 pips, i.e. usually slightly cheaper than standard accounts, with the difference growing during calm, tight-spread hours.

Stock and index CFDs vary more: index CFDs are usually spread-only; share CFDs often charge commission as a percentage or per-share fee with a minimum. Check the instrument specification — then model exactly that.

Converting Commission to Pips (So Everything Adds Up)

Backtests are easiest to audit when every cost lives in one unit. For forex:

commission in pips = round-trip commission ÷ pip value per lot e.g. $7 ÷ ($10/pip on EUR/USD) = 0.7 pips per trade

Now the spread, commission and swap columns sum into a single per-trade cost, and expectancy math stays clean.

Which Account Wins? It’s a Frequency Question

Strategy profile	Cheaper structure, usually	Why
Scalping (5–15 pip targets, many trades)	Raw + commission	Total ~0.8 pips beats ~1.1; the saving compounds across hundreds of trades
Intraday (15–40 pip targets)	Either — test both	Differences are small vs target size
Swing/position (50+ pips, overnight holds)	Spread-only often fine	Cost difference is noise; swap dominates instead

The honest move in a backtest: compute net expectancy under both structures once. It takes one extra spreadsheet column and occasionally reveals that a marginal scalping strategy is only profitable on raw pricing — which is a real finding about where you can trade it.

Worked Check

The pillar example strategy (100 trades, +1.9 pips/trade gross): on a standard account it pays ~1.0 pip average spread → net +0.9/trade. On raw: 0.2 spread + 0.65 commission = 0.85 → net +1.05/trade. Both live, raw is ~17% better — and a backtest that counted neither believed +1.9, double reality.

Related: the full cost audit · why live results differ

Frequently Asked Questions

Do replay tools model commission?

None of the mainstream ones do — including the free tick-replay tools, which model spread via real bid/ask data but not commissions. Keep commission as a manual journal column: a fixed per-round-trip number in pips, summed with spread and swap before computing expectancy.

Is commission negotiable or volume-based?

Often, yes — many brokers tier commissions down with monthly volume, and some rebate active traders. For backtesting, use the rate you'd actually start at; re-run the analysis at the discounted tier only to see whether scale would change the verdict.

How are share CFD commissions different?

Share CFDs commonly charge a percentage of notional (e.g. ~0.08-0.10%) or cents per share, with a minimum per order. Minimums matter: on small positions the minimum can dwarf the percentage, making small share-CFD trades structurally expensive — a backtest of small-account stock strategies must include it.

Which is better for a beginner's backtest?

Model the spread-only account first — it's the default account type and the simpler cost model. If the strategy survives, re-run with raw-plus-commission to see if the edge improves; if a strategy only survives on raw pricing, treat it as marginal.