Fees are recorded as a share of 1R, and the unit is the whole point.

Why currency fails

The same cost figure describes a different situation every month.

A trade costing eleven dollars in spread, commission and swap tells the operator nothing on its own. On a small account risking twenty dollars it is a material tax; six months into compounding, risking two hundred, it is noise. Recorded in currency, that entire change is invisible in the record — the number stayed similar while its meaning inverted. Any attempt to judge friction from a currency column therefore requires mentally re-deflating every row by the account size at the time, which nobody does reliably and the sheet should never have asked for. The column would still be present, still be accurate, and still answer no question anybody has — which is the most expensive kind of field, because it looks like coverage.

The subtraction that has to work

Expectancy is quoted in R, so friction must be quoted in R.

The reason for the unit is arithmetic before it is philosophical. Branch expectancy is an R figure, the scorecard blends R figures, and the whole EV apparatus operates in risk units. A friction cost expressed in currency cannot be subtracted from any of it without a conversion the sheet would have to guess at. Recorded as a percentage of 1R, it subtracts directly: an edge of 0.31R paying 9% of 1R in friction is an edge of roughly 0.22R, and that is the number the operator is actually trading. Net-of-friction expectancy is not a refinement of the gross figure — it is the only version that exists. A gross EV is a description of a trade nobody took, and quoting one alongside the net figure invites the operator to think of the difference as recoverable rather than as structural.

Where the cost concentrates

Friction scales with the inverse of the stop, which is counterintuitive.

The unit reveals a relationship currency hides. Spread and commission are broadly fixed per trade while 1R is set by stop distance, so the tighter the stop the larger the friction as a share of risk. An operator who tightens stops to improve risk-reward is simultaneously and invisibly increasing the proportion of every risk unit consumed by cost, and the two effects can cancel. The column makes that visible per trade. It is also why the tight-stop rows above sit where they do, and why holding a tight-stop position over a weekend is the quiet killer rather than an obvious one.

Why per trade and not per month

A monthly total cannot be attributed to anything.

Recording friction once a month as a summary figure would be less work and would answer no question. Per trade, the cost carries the row's branch, session, pair and duration with it, which means the aggregate can be cut by any of them — and the cuts are where the finding lives. A friction problem is almost never uniform; it concentrates in one pair, one session, or one holding pattern. A monthly total says costs are high and offers nowhere to go, while the same data at row level says which subset is paying for it.

• Currency figures need re-deflating by account size; nobody does it reliably.
• R-denominated friction subtracts directly from R-denominated expectancy.
• Fixed costs against a variable 1R means tighter stops raise proportional friction.

The threshold that matters

There is no universal acceptable number, only a ratio to the edge.

It is tempting to ask what level of friction is too high, and the question has no free-standing answer. Nine per cent of 1R is comfortable against an edge of 0.4R and fatal against an edge of 0.08R, because what matters is the share of expectancy consumed rather than the share of risk. That is the calculation the column enables and the reason it feeds net expectancy rather than sitting as a cost report. An operator whose edge is thin has a friction problem at levels that would be unremarkable for someone else. Stated as a ratio it becomes a governable target: the question is what share of expectancy is being surrendered to cost, and that share is the thing to drive down.

The key idea

The unit choice is what makes friction a governable quantity.

Retail edges die quietly and friction is the most common cause, precisely because currency accounting keeps the cost legible and its significance hidden. Denominating in R converts a bookkeeping entry into a comparable, subtractable, attributable quantity that sits in the same units as everything it competes with. That is a small formatting decision with an outsized consequence: it turns a cost the operator can see into a cost the operator can act on.