Logging trades in R-multiples instead of dollars

A trader who tells you they made $400 last month wants you to be impressed. What they don't tell you is that $300 of it was on a single oversized position, and the rest came from twenty smaller trades that mostly leaked. They don't tell you they were risking $200 per trade on the small ones and $1,000 on the big one. So that "$400 month" is actually a single mediocre win covering the slow drip of seventeen losses, plus one moderately bad day. The dollar number says one thing; the underlying performance says something completely different.

Logging in R-multiples instead of dollars is the fix. R stands for "risk" — specifically, the dollar amount you put on the line at the moment you sized the position. If you risked $200 and made $300 net, that's a +1.5R win. Risked $200 and the stop fired, that's −1R. The scale stays constant across every trade regardless of account size, position size, or instrument. That trader's $400 month becomes something you can actually evaluate: what was the average R, what was the win rate, where's the edge — and is there one at all.

This post is the case for switching, the math behind it, the practical process for converting an existing dollar-only journal to R retrospectively, and the metrics that get cleaner once everything is in the same unit.

What R actually is

R is a ratio, not a unit of currency. Take any closed trade and write down two numbers: the dollar P&L when it closed, and the initial risk you posted when you opened — meaning position size × distance from entry to stop loss. Divide the first by the second. That ratio is the R the trade achieved.

Concrete example. You bought 0.05 BTC at $90,000 with a stop loss at $88,000. Position notional was $4,500; the stop was 2.22% below entry. Initial risk = position size × stop distance = $4,500 × 2.22% = $100. The trade hit your $94,000 take-profit before the stop. P&L = 0.05 × $4,000 = $200. R achieved = $200 / $100 = +2R. The trade was a "two-R winner."

Hit the stop instead? Loss = $100. R = −$100 / $100 = −1R. By construction, every full-stop loss is exactly −1R. Anything between is a partial outcome — closed early at break-even is 0R, scaled out at $92k for half the position is roughly +0.5R on that half.

The mental model that makes R click: every trade you take buys you "one R of variance." A win pays out some multiple of that R; a loss pays back exactly one R. Across enough trades, the average R you collect (positive or negative) is the only number that summarises whether your edge is real.

Why dollars are a mirage

Two traders run the same strategy on the same setups for a year. Trader A starts with $5,000 and risks 1% per trade. Trader B starts with $500,000 and risks 1% per trade. End-of-year, A's net P&L is $1,800. B's is $180,000.

Their dollar journals look completely different. Their R-journals are identical to within a rounding error — same setups, same exits, same risk-as-fraction-of-account, so the same R-per-trade. The "100x performance gap" is account size, not skill. The dollar lens hides that.

The same effect operates inside a single account over time. A successful trader's account compounds; what was a $50 loser at year one is a $500 loser at year three on the same R-percentage of risk. The dollar amount climbs while the underlying performance is constant. A drawdown of $10,000 means something completely different on a $50,000 account versus a $500,000 account; "drawdown in R" — five back-to-back −1R losses — means the same thing in either case.

R also lets you compare across instruments without converting. A BTC trade with a $400 win on $200 risk and a forex trade with a $40 win on $20 risk are both +2R. The dollar values are 10× different and not directly comparable; the R values are identical and tell you the trades performed equivalently. If you trade more than one market, R is the only common unit that doesn't lie about which one is doing the work.

Converting an existing dollar log to R

Most traders keep a dollar journal first and discover R later. Converting retrospectively takes a single afternoon and unlocks every R-based metric for the entire history. The process:

1. For each closed trade, write down the original stop loss distance. In dollars per coin (or per share, or per contract). If your journal didn't capture this, reconstruct from screenshots, exchange order history, or your own discipline rule (e.g. "I always set stops 2% from entry" — defensible if it was actually true).
2. Compute initial risk per trade. Initial risk = position size × stop distance. For a $4,500 BTC position with a 2.22% stop, that's $100. This is your R unit for that specific trade.
3. Compute R achieved. Trade P&L (net of fees if you're being thorough) divided by initial risk. A $250 win on $100 risk = +2.5R. A $80 loss on $100 risk (closed before the stop) = −0.8R.
4. Build a column. Add an "R achieved" column next to your dollar P&L column. From this point on, every new trade gets logged with both.

Common edge cases that come up the first time you do this:

• Scaled-in entries. Average the entries weighted by size, then compute risk against the average entry. Treat the position as a single trade for R purposes.
• Stop loss moved up to break-even. Compute R against the original stop, not the moved one. The decision to move the stop is a separate decision worth tracking but the trade's "R outcome" is best measured against the risk you originally posted. Some traders log both — original-R and effective-R — for that exact reason.
• Trades without an explicit stop. Either estimate where the stop would have been (a 2x ATR distance from entry is a reasonable default), or exclude those trades from R-statistics. "Implicit stop" trades often hide larger risk than the trader thinks; surfacing them is informative either way.
• Partial exits. A trade where you took 50% off at +1R and the rest at +3R is a single +2R trade weighted by size, or two trades for journal purposes. Either choice is fine if applied consistently.

The retrospective conversion is the one-time tax. Going forward, you log entry, stop, size, exit — R falls out automatically.

The metrics R unlocks

Once every trade is in R, several useful numbers become directly computable. The headline four:

Expectancy in R. Mean R per trade across the journal. The single most informative number about whether a strategy has edge. Positive expectancy = you're being paid to take the trade. Strategies typically run between +0.1R and +0.5R per trade in retail-accessible markets after fees; anything claiming much higher than that consistently is either a small sample or selling something. Plug your numbers into the trade outcomes calculator for the inverse view at any specific position size.

Profit factor in R. Gross winning R divided by gross losing R (absolute value). Above 1.0 means winners outsize losers in aggregate; the conventional bar for a "trustworthy" strategy is profit factor above 1.5 across at least 100 trades. The shape is the same as dollar-profit-factor but the number is scale-invariant — your $5k-account profit factor will match your $500k-account profit factor on identical strategy execution. The full story on this metric in profit factor vs expectancy.

R-Sharpe (risk-adjusted return per trade). Mean R divided by standard deviation of R across the journal. Captures both edge and consistency in one number. A strategy with +0.4R expectancy but extremely volatile per-trade outcomes (occasional +5R wins surrounded by clusters of losses) gets a lower R-Sharpe than a +0.3R strategy with tight, repeatable per-trade variance. Most retail traders prefer the second.

Max drawdown in R. The longest stretch of consecutive losing R during the journal, in absolute value. Independent of account size and instrument, so directly comparable across strategies and time periods. The losing-streak probability calculator shows how long a "normal" worst streak should be at any given win rate — drawdowns longer than that suggest the edge is shifting.

There are second-order metrics too — R-MAR ratio (annualised R return divided by R-drawdown), R-Sortino (uses only downside deviation), R-Kelly (the Kelly-optimal risk fraction expressed in R) — but the four above cover most decisions. If your average R is positive, your profit factor is above 1.5, your R-Sharpe is above 0.5, and your worst R-drawdown sits inside the band the streak math predicts, you have a strategy. None of those numbers can be computed cleanly from a dollar-only journal.

Where R doesn't work cleanly

R is a tool, not a religion, and it has edges where the lens breaks down.

Trades without an explicit stop loss. Some swing-trading approaches deliberately don't set hard stops; the exit is judgmental, based on changing fundamentals or technical invalidation that develops over time. R is undefined for those trades unless you assign a hypothetical stop. The honest move is to keep them in a separate journal and accept that R-statistics describe only your stop-defined trades.

Strategies with deliberately variable risk. A discretionary trader who sizes "based on conviction" — small risk on average setups, larger risk on A+ setups — has by design a non-uniform R. The R lens still works (each trade still has its own initial risk), but you lose some of the comparative power. The unstated assumption that R-based statistics carry is risk-per-trade is approximately constant, and the more it varies, the more you'll need to weight or normalise.

Funding and holding costs on perpetual futures. A short BTC perp held across a positive funding cycle pays funding to longs every eight hours. That's a real cost the trade absorbs but R doesn't natively capture (since it's not part of the entry-to-stop distance). If you hold perps, log funding cost in a separate column and either deduct it from R achieved or report it alongside.

Tax filing. Tax authorities want dollars. R is for performance evaluation, not for filing. Keep both; never replace.

What R does for the trader, beyond the math

Switching to R-multiples is one of the small process changes that produces an outsized behavioural shift, separate from the metric improvement. Two specific effects worth naming.

The first is that R disconnects emotional intensity from dollar amount. A trader who reads a $1,500 loss on a Monday and feels like they took the elevator down twelve floors is reading the dollar number, not the underlying performance — that loss may have been a routine −1R inside a strategy with +0.4R expectancy. Reading "−1R" instead carries the right weight. The math hasn't changed; the lens has. Many traders find their own post-loss recovery time shrinks meaningfully once they stop logging in dollars.

The second: R forces honesty about the role of position sizing. A trader who increased their risk-per-trade from 0.5% to 2% to "see what the strategy can really do" hasn't found a better strategy — they've quadrupled their R-volatility on the same edge. The dollar log will scream at them about "amazing wins" when it's running hot and "catastrophic losses" when it's not. The R log will show flat performance with quadrupled noise — accurate in a way the dollar log can't be.

R-multiple journaling is one of those small invariants that, once added, you don't want back. The conversion takes a Saturday afternoon. The clarity it gives back lasts the rest of your trading career.

If you're starting fresh, the R-multiple calculator computes R for any single trade you punch in. The trade outcomes analyser handles the position-sizing side of the same coin. The 60-trade equity curve shows what R-based metrics look like across an actual run. And once R is the unit, the next layer is what to tag each trade with — the trade-tagging taxonomy covers the four labels that turn an R log into a post-mortem instead of a spreadsheet.

FAQ