Page 2 of 2

Re: Measuring slippage by session: London vs NY vs Asia sample

Posted: Mon Sep 28, 2026 10:26 pm
by LondonScalper
The Asia shape you describe is what my histogram looked like too: mostly zeros and a tail that does all the damage. The number that mattered was the size of the worst 5% of fills, which was bigger in Asia than in London despite far fewer trades.

On the news tagging, the Forex Factory weekly feed is handy going forward but only covers the current week, as you say, so run it every Friday and keep the file, or the history is gone. I'd also tag by event rather than by currency. A US release matters to EURUSD and gold but barely registers on EURGBP, and splitting EURUSD into EUR and USD will over-tag the sample.

The broker offset is the part that bit me. My server runs on GMT+3 in summer and GMT+2 in winter, and the week when Europe has changed its clocks and the US hasn't sits awkwardly in every log. I keep everything in UTC now and convert only for display.

Re: Measuring slippage by session: London vs NY vs Asia sample

Posted: Sun Oct 04, 2026 6:10 pm
by LondonNewsTrader
PTScalper wrote: The Overlap presents a different trap.
To make the overlap numbers honest I'd separate out the fills that happen within a few minutes of scheduled releases. Otherwise the 8:30 New York data dominates the tail of the overlap bucket and makes the whole session look worse than it is for ordinary trading. In my own logs, slippage outside release windows in the overlap is close to the London morning, and almost all of the bad fills come from a handful of minutes.

For the same reason I'd report percentiles rather than averages. The median slippage in every session is usually zero or close to it, so the useful figures are the 90th and 99th percentiles, split by side.

The Asia point is interesting and I'd like to see the numbers. A thin book with occasional large misses fits what I'd expect, but the sample size in Asia is usually small, so a few bad fills can change the picture.

Re: Measuring slippage by session: London vs NY vs Asia sample

Posted: Sat Oct 31, 2026 7:21 am
by FXS Prop Desk
Order book density is a better explanation than volume, agreed. The Asia pattern of zero slippage most of the time and a large slip occasionally is exactly the kind of thing an average hides. For anyone running a similar study, it helps to report two figures per session: the share of trades with any slippage, and the average size of slippage when it occurs. A session with rare but large slips needs a different response, such as smaller size or wider stops, than one with frequent small slips, which is more about choosing a better broker.