Chapter 19: Execution, Order Management & Tech: Get Paid What You Deserve

Execution is not a cost you accept — it’s a skill you can improve. Better fills and smarter order management directly raise net returns. Everything after your strategy and sizing is execution: the order type you use, how you place it, the broker routing, and the tech stack that runs it. Small improvements — 0.5 pip average slippage saved, fewer requotes, tighter fills — compound into meaningful P&L gains. This chapter is a hands-on playbook: how to measure execution, what order types to use, how to test brokers, and the tech that reliably improves fills.

1 — The execution mindset

Treat execution like a production process with metrics and continuous improvement. Your trading edge is partly strategy, partly how well you can capture that edge. If you run a scalping system with a 0.5-pip expected edge and your execution costs eat 0.6 pips on average, you’re losing money even though the strategy looks profitable on paper.

Key metric to own: All-in cost per trade = spread + commission + slippage + implicit impact.

2 — Order types & when to use them (practical)

Know these orders and use them deliberately:

• Market order: immediate execution at the best available price. Use when speed beats price. Good for stop-loss fills during fast moves.

• Limit order: executes at your price or better. Use when you want price control and can wait (scalpers using limit entries to capture maker rebates).

• Stop / Stop-loss: market order triggered at a price — use for protection but expect slippage in volatile moves.

• Stop-limit: stop triggers a limit order (no slippage guarantee). Use when you refuse to accept worse than X price.

• OCO / OCA (One Cancels Other / One-Acts-All): pair your TP & SL or bracket orders — reduces manual bookkeeping and emotional errors.

• IOC (Immediate or Cancel): execute immediately any available quantity; cancel the rest. Good for partial fills where you only want instant liquidity.

• FOK (Fill or Kill): require full immediate fill or cancel — useful for block orders where partial fills are unacceptable.

• Iceberg / Hidden orders: split a large order into visible chunks — reduces market impact (primarily available via advanced platforms / FIX).

• VWAP / TWAP algos: slice large orders over time to hit a benchmark price — used by larger accounts to reduce slippage/impact.

Practical rule: use limit orders for entries where possible; use market orders for exits during confirmed breakouts or to avoid emotional losses.

3 — Measure slippage properly (simple, repeatable test)

You must measure execution yourself. Here’s a 30-trade slippage test you can run in one afternoon.

Slippage test protocol

1. Choose 1–2 liquid pairs (EUR/USD, USD/JPY) and the session you trade (London/NY overlap).

2. Place 30 identical trades with the same size using your normal order type (market or limit). Record requested price and executed price for each trade.

3. Compute slippage per trade = (Executed Price − Requested Price) for buys, reverse for sells (pips). Record spread at time of order too.

CSV / log header (use in Excel): trade_id, timestamp, pair, side, requested_price, executed_price, spread_at_order (pips), slippage (pips), order_type, notes

Example (3 sample trades) — arithmetic step-by-step

• Trade 1: requested 1.10000, executed 1.10012 → slippage = 1.10012 − 1.10000 = 0.00012 → 1.2 pips (since 1 pip = 0.0001 for most majors).

• Trade 2: requested 1.10000, executed 1.10003 → slippage = 0.00003 → 0.3 pips.

• Trade 3: requested 1.10000, executed 1.09995 → slippage = 1.09995 − 1.10000 = −0.00005 → −0.5 pips (negative slippage = positive execution for you).

After 30 trades compute:

• Average slippage (pips) = sum(all slippages) ÷ 30 (compute digit-by-digit).

• Std dev of slippage to see dispersion.

• % positive fills (executed equal/better than requested).

If average slippage > 0.5 pips on EUR/USD during your session with your size, you can likely reduce costs by switching account type, broker, order method, or improving tech.

4 — Execution KPIs to track (minimum set)

• Average slippage (pips) by instrument & session.

• Fill rate (percent of limit orders filled).

• Requotes / refused orders per 1000 orders.

• Latency (ms) — order send to ack.

• Market impact (for size tests): price movement caused by your order for X seconds.

Measure weekly. Treat KPI deterioration as a trigger to investigate brokers, routes, or internet issues.

5 — Broker testing, the roadside inspection

Before trusting a broker with real capital, run this checklist:

Broker Execution Test (do the tests in order):

1. Spread snapshot: capture live spreads for 8 hours across your session.

2. Slippage test: 30-trade protocol (see above) on your usual lot size.

3. Latency & re-quotes: place 100 small market orders and count requotes / rejections.

4. Withdrawal test: deposit small amount; withdraw via your intended method to verify timing/fees.

5. Order types & partial fills: test IOC/FOK and large limit orders to see how partial fills are handled.

6. Trade during news: simulate (or paper-trade) around a small scheduled release to observe spread blowout and stop execution.

7. API/FIX test (if relevant): request demo FIX or API creds and run simple round-trip latency & fill tests.

If the broker fails any critical test (very high average slippage, many requotes, or withdrawal problems), don’t scale with them.

6 — Tech stack that matters (practical tiers)

Basic retail (what most discretionary traders need)

• Reliable broker with raw/ECN pricing.

• Desktop client (MT4/MT5) or TradingView + broker plugin.

• Stable internet + laptop/desktop.

• Cloud-backup for logs.

Intermediate (serious discretionary & small EAs)

• VPS near broker’s servers (recommended if running EAs or 24/7 strategies).

• Auto-reconnect rules and heartbeat monitoring.

• Order router or smart order management scripts (auto OCO, partial limit logic).

Professional / algorithmic

• Colocated VPS / EC2 in Equinix datacenter near broker/LP (NY4, LD4, TY3).

• FIX API or low-latency REST with DMA.

• Order management system (OMS) with slicing algos (TWAP/VWAP).

• Market data feed (tick-by-tick) and logging to a local DB.

• Monitoring & alerting system (latency > X ms, fill rate drop).

7 — VPS & latency best practices

• Use VPS providers with nodes in the same datacenter as broker (ask broker for server location).

• Minimum VPS spec for EA: 1 vCPU, 2GB RAM, SSD — increase depending on EAs and backtesting.

• Monitor ping and packet loss — aim for < 10 ms to broker gateway for low-latency shares.

• Have a failover plan: local laptop and mobile app to close positions manually if VPS fails.

8 — APIs, FIX & automation (what traders must know)

• FIX vs REST vs WebSocket: FIX is industry standard for low latency, REST/WS are fine for retail algos. Use FIX for production algos requiring sub-50ms round-trip times.

• Idempotency & order state: your system must handle duplicate messages, partial fills and rejections safely.

• Sequence numbers & recovery: implement message recovery — missed messages must be detected and resynced.

• Testing: never go live without a month of paper/FIX sandbox testing with simulated load.

9 — Practical scripts & small automations you can use now

• Auto-OCO script: when entry triggers, automatically place TP and SL as OCO to eliminate manual errors.

• Slippage logger: a tiny script (platform EA or API client) that writes requested_price, executed_price, timestamp, spread to a CSV for every trade — use this for ongoing KPI tracking.

• Partial fill handler: set logic to cancel remaining quantity after N seconds or re-issue IOC for remainder if required.

(If you want, I can generate a sample MT4 EA snippet that logs requested vs executed prices into a CSV.)

10 — Market impact & capacity testing (how to scale without breaking price)

If you increase size, test impact:

Incremental order test

1. Choose instrument and session.

2. Submit 10 test orders in increasing lot sizes (e.g., 0.1, 0.2, 0.5, 1.0, 2.0) with limits at mid-spread.

3. Measure average slippage / fill % for each size.

4. Calculate Impact per standard lot = (avg slippage pips) ÷ (lot size in standard lots).

If impact per lot increases non-linearly, that signals limited liquidity and warns against linear scaling.

11 — Practical execution recipes (examples)

Scalper (retail, low latency)

• Use: Raw/Razor account, MT5 or cTrader, VPS colocated, limit entries where possible, IOC for exits when you need instant execution.

• Monitor: average slippage < 0.3 pips on main pair during your session.

Swing trader (discretionary)

• Use: Limit entries on pullbacks, bracket orders (OCO), and larger stop distances to avoid being clipped by spread.

• Monitor: fill rate for limits > 60% and average entry slippage < 1.0 pip.

EA / algorithm

• Use: FIX or low-latency REST, VPS in same datacenter, robust reconnect and order reconciliation.

• Test: 10k simulated orders in sandbox to validate state machine and partial fills.

12 — Execution audit checklist (run monthly)

•  Export last month’s trade log (all instruments).

•  Compute average slippage by instrument & session.

•  Compute fill rate for all limit orders.

•  Check requote / rejections > baseline (investigate if increased).

•  Run 30-trade slippage mini-test monthly (same session/pair) and compare to baseline.

•  Re-test withdrawal & deposit process (quarterly).

•  Re-verify VPS latency and packet loss.

•  If KPIs worsen by >20%, run broker retest (section 5).

13 — Common execution mistakes & fixes

• Mistake: trusting advertised spreads instead of live measures.Fix: measure real spreads over 2 weeks and compute average, not "from" numbers.

• Mistake: crossing liquidity with oversized orders.Fix: capacity tests and slicing orders (TWAP/VWAP).

• Mistake: no logging for fills and slippage.Fix: implement a simple slippage logger today (EA or API client).

• Mistake: single broker dependency.Fix: keep at least one backup broker and test porting procedures quarterly.

14 — Small action plan (do these in the next 48 hours)

1. Run the 30-trade slippage test for your primary pair & session. Log results.

2. If average slippage > 0.5 pips (majors), contact broker support with your CSV and ask for explanation — treat it as data, not a complaint.

3. Set up a basic VPS (1 vCPU, 2GB) and run your EA/demo for 24 hours to verify connectivity.

4. Add an auto-OCO rule in your platform or place bracket orders manually for each trade for one week.

5. Book a monthly “Execution Audit” calendar reminder.

15 — What “better execution” is worth (small math)

If your strategy trades 2 lots/day on EUR/USD (1 standard lot = $10/pip), and you can save 0.3 pips average slippage:

• Pips saved per day = 0.3 pips × 2 lots = 0.6 pip-equivalents (be careful: 2 lots = 2 standard lots = $10/pip × 2 = $20 per pip total).Compute daily dollar improvement step-by-step:

• Value per pip for 2 lots = $10 × 2 = $20 per pip.

• Dollar saved per day = 0.3 pips × $20/pip = step-by-step: 0.3 × 20 = $6.00.

• Trading 250 days/year → annual improvement = $6 × 250 = step-by-step: 6 × 250 = 1500 → $1,500 per year.

Small execution gains scale meaningfully with volume and time.

Execution is an operational edge. Measure, iterate, and treat fills as a KPI you can improve. Start by logging every trade’s requested vs executed price — data will expose avoidable costs.