Broker execution scoreboard
Broker execution scoreboard
Measured execution speed
Execution speed here is the broker's own millisecond timestamps — the gap between the order reaching the server and the fill — so it excludes your internet connection and measures only what the broker controls. Across 185 measured fills on 1 connected account, OANDA Japan filled orders in a median of 168ms, with 95% of fills inside 176ms. The p95 matters more than the median if you trade news: it's the slow tail that costs money.
On XAUUSD, OANDA Japan's measured fill latency ties for 10 of 16 measured brokers.
Median fill latency across all symbols — latency is the one metric that compares across symbols.
median, broker-side (order received → filled)
Median and p95 fill latency per symbol.
| Symbol | Median latency | p95 latency | Fills |
|---|---|---|---|
| XAUUSD | 168 ms | 176 ms | 185 |
Median 168ms from order received to filled, measured on 185 real fills; the slowest 5% took longer than 176ms. The session table shows how that shifts across the trading day.
From the broker's own order history: MT5 records when the server received the order and when it filled, in milliseconds. That isolates the broker's processing from your network latency — a slow VPS can't make a broker look bad here, and a fast one can't make it look good.
For scalping and news trading, yes — hundreds of milliseconds move prices. For swing trades held days, median latency is nearly irrelevant, though a long p95 tail can still hurt stop-loss fills in fast markets.
Read more: What Execution Speed Is Normal in MT5? Read Lifecycle Timing Correctly · Broker Execution Quality: Measure Slippage, Lifecycle Timing and Spread — or connect an account read-only and put your own fills in these numbers.