Execution transparency.
Four figures matter for execution quality: latency, fill rate, rejection rate and feed uptime. Below is the state of each — published, or not, with the reason.
What a published figure has to carry
A number on this site is not allowed to appear on its own. Before any execution figure is published it must state all four of the following, and the build fails if one is missing:
- 01Endpoints. Where the measurement starts and where it stops.
- 02Statistic. Which statistic the figure is — a median is not a mean and neither is a worst case.
- 03Sample window. The period it covers, with dates.
- 04Sample size. How many observations it rests on.
Three of the four is not a measurement. We would rather show you a gap than a figure you cannot check.
Not published yet4
Each of these is registered, defined and deliberately unpublished. The definition is written first so that the figure cannot be reverse-engineered from a flattering measurement later.
- Price feed latency
- Time from a liquidity provider tick arriving to that price being renderable in the terminal.
Not published yet — awaiting a measured value with its statistic type and sample window.
- Fill rate
- Share of market orders filled at or better than the quoted price.
Not published yet — awaiting a measured value with its statistic type and sample window.
- Rejection rate
- Share of orders rejected by the pricing engine or the liquidity provider.
Not published yet — awaiting a measured value with its statistic type and sample window.
- Feed uptime
- Share of trading hours with a live, non-stale price feed.
Not published yet — awaiting a measured value with its statistic type and sample window.
Ask us directly
If execution quality is what your decision turns on, the demo environment is the place to test it — with your own instruments and your own order sizes rather than our numbers.