Roundup: Measuring slippage yourself, from spreadsheets to histograms
Posted: Tue Oct 13, 2026 9:51 am
Roundup: Measuring slippage yourself, from spreadsheets to histograms
Slippage threads tend to die at zero replies because nobody wants to post their fills, but there's a solid set of them on the board now and between them they describe a complete way to measure it. Here's the condensed version.
Why small numbers matter
The 0.1 pip thread does the multiplication most of us avoid. A tenth of a pip against you on 100 EURUSD tickets is 10 pips of pure cost before commission, roughly $100 per standard lot. At a few hundred tickets a month, that's real money leaking without anyone noticing, and it can be a meaningful slice of expectancy for a scalper.
The basic sheet
The spreadsheet template thread lays out the columns: time in both server and local, symbol and direction, order type, requested price, filled price, and signed slippage in pips. From there you look at average slip by hour, by order type, the share of positive, zero and negative fills, and news minutes versus normal minutes. Nothing fancy, and MT5 order history gives you most of it.
Look at the shape, not the average
This is where the threads get more interesting. The GBPJPY histogram thread found the centre of the distribution boring and the tail expensive: a few bad fills did most of the damage. A reply there split the same data by order type and found almost all the ugly tail came from stop exits, not market entries. The execution tests thread makes the same point from the other side: worst values tell you more than averages, because a broker can look fine on average and still hurt you during releases.
Is it fair or asymmetric?
The asymmetric slippage thread gives a test. Collect 100 to 200 market and stop fills, count positive versus negative outcomes, and compare their sizes. Some imbalance in fast markets is normal, especially on stops. A complete absence of positive slippage on market orders is the red flag.
Things that change the numbers
Requested price isn't always recorded cleanly, so some of this is imperfect. The 0.1 pip thread tracks a rolling median rather than reacting to one horror story, which seems right. And measure on a small live account. Demo servers often fill instantly at the quoted price, which tells you nothing.
For those who've done this, was your slippage mostly in stop exits too, or did entries surprise you?
Threads this pulls from
Slippage threads tend to die at zero replies because nobody wants to post their fills, but there's a solid set of them on the board now and between them they describe a complete way to measure it. Here's the condensed version.
Why small numbers matter
The 0.1 pip thread does the multiplication most of us avoid. A tenth of a pip against you on 100 EURUSD tickets is 10 pips of pure cost before commission, roughly $100 per standard lot. At a few hundred tickets a month, that's real money leaking without anyone noticing, and it can be a meaningful slice of expectancy for a scalper.
The basic sheet
The spreadsheet template thread lays out the columns: time in both server and local, symbol and direction, order type, requested price, filled price, and signed slippage in pips. From there you look at average slip by hour, by order type, the share of positive, zero and negative fills, and news minutes versus normal minutes. Nothing fancy, and MT5 order history gives you most of it.
Look at the shape, not the average
This is where the threads get more interesting. The GBPJPY histogram thread found the centre of the distribution boring and the tail expensive: a few bad fills did most of the damage. A reply there split the same data by order type and found almost all the ugly tail came from stop exits, not market entries. The execution tests thread makes the same point from the other side: worst values tell you more than averages, because a broker can look fine on average and still hurt you during releases.
Is it fair or asymmetric?
The asymmetric slippage thread gives a test. Collect 100 to 200 market and stop fills, count positive versus negative outcomes, and compare their sizes. Some imbalance in fast markets is normal, especially on stops. A complete absence of positive slippage on market orders is the red flag.
Things that change the numbers
- Exit type. The exits thread found limit orders at target filled at the price or better almost every time, while manual market closes during fast moves were consistently worse. They were also often earlier than planned, which is a different leak.
- The MT5 deviation setting. Too wide and you'll accept nearly any price, too tight and you get rejected in fast markets. One reply notes it does nothing at all on market execution accounts. For EAs, a per-symbol value of one to two times normal spread made more sense than one global number.
- Venue. The NFP prop server thread logged fills relative to the print in seconds and found the prop book fine on quiet days and ugly in the one minute that mattered, even with the same strategy as a personal raw account.
Requested price isn't always recorded cleanly, so some of this is imperfect. The 0.1 pip thread tracks a rolling median rather than reacting to one horror story, which seems right. And measure on a small live account. Demo servers often fill instantly at the quoted price, which tells you nothing.
For those who've done this, was your slippage mostly in stop exits too, or did entries surprise you?
Threads this pulls from
- Asymmetric Slippage: How to Detect It
- How much does 0.1 pip average slippage cost per 100 trades?
- Slippage Spreadsheet Template: Requested vs Filled Price by Hour
- Slippage tails on GBPJPY: histogram from my deals
- Where my exits slip, market vs limit for taking profit
- The MT5 deviation setting and what it does to your fills on fast markets
- Measuring Execution Quality: A Short List of Tests
- Slippage on prop servers during NFP: sample logs