forex-scalping.com

Spread as a share of the M5 average range: a 15% cap, worked at 6.0, 4.0 and 2.2 pip ATR

A spread of 0.7 pips sounds the same at any time of day, but what it takes out of a scalp depends on how far price travels in a candle. One way to express it is spread divided by the average M5 range.

Here are three made-up readings of ATR(14) on EURUSD M5, with a spread to go with each. Treat them as a template, not measurements.

London open: ATR 6.0 pips, spread 0.5, share 8.3%.
Midday: ATR 4.0 pips, spread 0.7, share 17.5%.
Late evening: ATR 2.2 pips, spread 1.2, share 54.5%.

Now set a cap of 15%. The largest spread you'd accept is 0.15 x ATR, so 0.90 pips at the open, 0.60 at midday and 0.33 in the evening. Only the first reading passes. The second misses by 0.1 pip and the third is far outside.

The effect on a trade: a 5 pip target at midday gives away 0.7 / 5 = 14% of the target before price has done anything, and with a 4 pip stop the entry costs 17.5% of the risk. Late in the evening a 5 pip target loses 24% to the spread and a 4 pip stop 30%.

In an EA the rule is one line: trade only if spread_in_points <= 0.15 x ATR_in_points. By hand it's two numbers glanced at before entering. The cap itself is a choice, and 10% or 20% would work too. What helps is that it's one number applied the same way in every hour, so the filter doesn't depend on your mood.

The ratio also explains why the same pair feels different across the day without anything being wrong with your broker. When ATR halves from 6.0 to 3.0 and the spread rises from 0.5 to 0.8, the share goes from 8.3% to 26.7%, a threefold increase in cost per unit of movement. Your edge, if it's measured in R, loses ground in those hours without any change in the setup. Some people respond by lowering the size, others by moving to a slower timeframe where ATR is bigger, and the arithmetic supports either choice.

A cap in percent is awkward for a pair like GBPJPY, where both the spread and the ATR are larger. Say the spread is 2.0 pips and ATR on M5 is 11 pips, so the share is 18.2%. EURUSD at 0.7 over 4.0 is 17.5%. They look the same by this measure, even though GBPJPY costs three times the pips. This is the advantage of using a share. It compares the cost to what the instrument actually does in the same window, so you can set one rule and apply it across pairs instead of keeping a list of pip limits.

There's a limit to the idea. ATR is an average over the last 14 candles, so right after a news spike it will read high even though the market has gone quiet again, and the share will look small for a few bars while the spread is still wide. Using the median of recent candle ranges rather than the mean reduces the effect, or you can add a second condition that the spread itself is under some absolute number. The cap catches the slow evening hours well and the spike aftermath poorly, so I'd pair it with a plain maximum spread.

To see how often a 15% cap would have blocked you, pull a day of M5 bars and the spread column for a pair you trade, and compute the share for each bar. Count the number of bars above 15%, and note the hours they fall in. You may find that nearly all of them sit in two or three slots, which turns a daily filter into a simple time window. If the count is more than half your trading hours, the pair or the account type isn't suited to scalping at that target size, and that's the information the exercise was for.