Choosing a scalping pair by feel leads to pairs that look lively but cost too much. One number captures most of what matters: the average range in your trading window divided by the average spread.
How to calculate it:
Take the average M5 high–low range during your session (e.g. 07:00–10:00 London) over the last month.
Take the average spread during the same window, plus commission converted to pips.
Divide range by cost.
Illustration: a pair with an 8-pip average M5 range and 0.8 pips total cost scores 10. A cross with 14 pips range and 3 pips cost scores under 5, despite looking more exciting.
Use the ratio to rank pairs per session, not overall. AUDUSD can rank well in Tokyo and poorly in late New York. Recalculate monthly, because both ranges and spreads shift with volatility.
Which pairs come out on top for your session, and did the ranking surprise you?
FXS Prop Desk wrote: Take the average spread during the same window
One thing worth adding: use the spread at the moments you actually enter, not the average across the window. Average spread looks tight because most ticks happen during calm stretches, but a lot of scalp entries happen right as price is moving fast, which is exactly when spread widens. Say EURUSD averages 0.2 over 07:00 to 10:00 but sits at 0.5 for the first couple of minutes after 07:00 or around data. If that's when you trade, your real ratio is a lot worse than the sheet says. Also worth using median range rather than mean, since one news candle can inflate an M5 average for the whole month. Otherwise it's a really clean way to stop picking pairs by how exciting they look.
I went one step finer and score each 15 minute block rather than the whole session. The ranking inside my 07:00 to 10:00 window changed more than the ranking between pairs did.
EURUSD scored about 9 between 07:15 and 08:30 London on median M5 range against spread plus commission, then fell to around 5 after 09:15. GBPUSD started lower, roughly 7 at 07:00 because of the wider opening spread, but held up better through mid morning.
The ratio also says nothing about whether the range is tradeable. A pair can print 8 pip M5 bars that overlap each other completely, and fading those is a different job from following them.
So I add a second number, the net move over the hour divided by the sum of the M5 ranges. Under 0.2 I treat that pair as a fade for the block, and above 0.4 as one to follow.
Commission needs care when comparing brokers. Some quote 3.5 per side and some 7 round turn, and I have seen the same pair ranked differently purely because one sheet counted only one side.
Did anyone find the ranking stable from month to month? Mine reshuffled noticeably between August and September.
