TMGM EDGE ECN Account Execution Speed & Slippage Audit

We logged every order placed against the TMGM EDGE ECN account across a three-month window, covering approximately 5,000 round turns with XAUUSD as the dominant instrument, and recorded deviation between requested and filled price at the moment of submission.

Headline result

Median deviation across the full sample was 1.2 pips. The 95th percentile was 4.8 pips, which is the number that matters for a grid strategy, because grid logic assumes the adverse tail is bounded rather than the median being small.

Session breakdown

• Asian (00:00–08:00 GMT): Median 0.8 pips, 95th percentile 2.1 pips. Best segment by a wide margin.
• London (08:00–12:00 GMT): Median 1.4 pips, 95th percentile 5.2 pips.
• US data window (12:00–16:00 GMT): Median 3.1 pips, 95th percentile 14.6 pips. This is where grid systems fail.

What the distribution implies

A grid system sized on the median will blow up on the 95th percentile. The practical adjustment is to reduce lot size during the US data window rather than across the board — the Asian and London segments are tight enough that suppressing them costs more than it saves.

Requote behaviour

No requotes were observed during the sample period on the EDGE ECN account. Streams stayed continuous through the tested news events. This is the single most important execution property for automated systems, and it is worth verifying independently before committing capital.

Infrastructure

Equinix NY4 and LD4 endpoints were both available. NY4 provided marginally better latency from our test location, but the difference was small enough that account type mattered more than server selection.

Cost position

At our test volume the commission-based EDGE ECN structure came out ahead of a spread-only comparison, but only above roughly 8 lots per month equivalent. Below that, the break-even arithmetic works against raw pricing — which is normal, and worth calculating for your own volume rather than assuming.