"Price improvement" is a phrase worth testing
Posted: Fri Oct 30, 2026 12:49 pm
"Price improvement" is a phrase worth testing
Some brokers advertise price improvement, meaning fills at a better price than the one requested. It sounds like a free gift. It can also be a statistic built on a flattering baseline.
The question is always: better than what? If the comparison is the price on the screen two seconds earlier, a fill that lands a tenth of a pip better is trivial to produce, because the screen price was stale. If the comparison is the best available bid or offer at the exact millisecond of execution, the claim means a lot more.
A fair audit lines up each fill against a reference tick stream at the timestamp of order arrival. Then the question becomes whether the improvements outnumber and outsize the deteriorations. Some brokers show a rosy headline number while the negative slippage lives in a different column, labeled differently.
Another thing to look for: improvement that only shows up on small orders. A 0.01 lot trade has no market impact, so it can be filled almost anywhere inside the spread. Larger sizes tell the real story.
Does the broker publish execution statistics, and are they broken down by order size?
Some brokers advertise price improvement, meaning fills at a better price than the one requested. It sounds like a free gift. It can also be a statistic built on a flattering baseline.
The question is always: better than what? If the comparison is the price on the screen two seconds earlier, a fill that lands a tenth of a pip better is trivial to produce, because the screen price was stale. If the comparison is the best available bid or offer at the exact millisecond of execution, the claim means a lot more.
A fair audit lines up each fill against a reference tick stream at the timestamp of order arrival. Then the question becomes whether the improvements outnumber and outsize the deteriorations. Some brokers show a rosy headline number while the negative slippage lives in a different column, labeled differently.
Another thing to look for: improvement that only shows up on small orders. A 0.01 lot trade has no market impact, so it can be filled almost anywhere inside the spread. Larger sizes tell the real story.
Does the broker publish execution statistics, and are they broken down by order size?