forex-scalping.com

Up 1.5R on the month but down $375: why a journal needs risk in dollars per trade

This is a made up month, not mine, to show a trap in journals that only record R. Six trades. Three winners at plus 1.5R, each with $50 at risk, and three losers at minus 1R, each with $200 at risk.

In R the month is 3 x 1.5 minus 3 x 1 = +1.5R. A spreadsheet that sums the R column says the month was positive, and the system has a 50% win rate and 1.5R winners, a profile that should make money.

In dollars: each winner makes 1.5 x $50 = $75, so three winners give $225. Each loser costs $200, so three losers give minus $600. The month is minus $375.

What happened is that the risk per trade wasn't constant. The winners were taken at a quarter of the size of the losers. That can come from sizing by feel, a larger size on trades the trader felt sure of, or a stop moved wider after entry while the lot size stayed the same. The R figures hide it because each trade is divided by its own risk.

The fix is to log the dollars at risk at entry as its own column. Then the month can be summarised as net dollars divided by total risk taken: minus $375 / (3 x $50 + 3 x $200 = $750) = minus 0.5R per unit of risk.

It also gives a quick check of size discipline. If the risk column varies a lot with no reason in the plan, the R-based statistics can't be trusted.

Yeah, you are right.

In reality it is more complicated, spread, slippage on open, slippage on close, different trading pairs means different pip value etc.

To compute a dollar-weighted figure, take the sum of dollar results over the sum of dollar risk. In the example that is minus 375 / 750 = minus 0.5. Compare this with the plain average of R, which is +0.25R per trade (1.5R over six trades). The two numbers differ in sign because the weights differ. When they differ in sign over a larger sample, the trader is making money on small bets and losing it on big ones, and the fix is behaviour, not strategy. It's one of the clearest signals a journal can give.

Read the risk column over a longer period and checking its spread. A simple test is the ratio of the largest risk to the smallest over a month. If you plan to risk 1% a trade and the figures run from 0.3% to 2.5%, that ratio is above 8, which is too much variation for plain R statistics. Write down the cause for each outlier: lot size miscalculated, stop widened, or a deliberate size-up. Counting the deliberate cases separately tells you whether the size-ups worked, and that may be the most useful result a journal produces.

Spreadsheet setup is short. Columns: date, symbol, side, entry, stop, lots, dollar risk at entry (computed as stop distance in pips x lots x pip value), result in dollars, result in R (dollars divided by dollar risk). Then summary lines: sum of dollars, sum of dollar risk, sum of R, and the ratio of the first two. If the broker's statement is the source of the dollar results, you get honest figures without trusting memory. Recording the planned stop at entry matters, because moving the stop later would otherwise change the R of the trade after the fact.

A final point on commissions and swaps. If the dollar result is taken from the platform it includes costs, and the R computed from it does too. If you compute R from price only, with a stop of 10 pips and an exit at 15, you get 1.5R and ignore a $7 commission on a $100 risk, which is 0.07R. Over 100 trades that's 7R of difference. So use the platform's net profit in the dollar column, and let the R follow from it, so costs sit inside the same figures you use for expectancy.