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 3,302 measured fills on 11 connected accounts, Startrader filled orders in a median of 135ms, with 95% of fills inside 161ms. The p95 matters more than the median if you trade news: it's the slow tail that costs money.
On XAUUSD, Startrader's measured fill latency ties for 5 of 14 measured brokers.
Weekly median fill latency.
Median fill latency across all symbols — latency is the one metric that compares across symbols.
median, broker-side (order received → filled)
Median 135ms from order received to filled, measured on 3,302 real fills; the slowest 5% took longer than 161ms. 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.
Server load shifts across the trading day.
| Session | Median latency | p95 latency | Fills |
|---|---|---|---|
| Asia (00–07 UTC) | 136 ms | 158 ms | 171 |
| London (07–13 UTC) | 135 ms | 157 ms | 112 |
| New York (13–21 UTC) | 139 ms | 168 ms | 245 |
| Late (21–24 UTC) | N/A — insufficient evidence | N/A — insufficient evidence | 22 |
Median and p95 fill latency per symbol.
| Symbol | Median latency | p95 latency | Fills |
|---|---|---|---|
| XAUUSD | 136 ms | 161 ms | 3,296 |
| EURUSD | N/A — insufficient evidence | N/A — insufficient evidence | 6 |