How do you keep a crypto trading journal that actually improves your results?
A crypto trading journal is a record of every trade you take, with the plan, the result in R (multiples of the amount you risked) and what you did right or wrong. It turns a pile of wins and losses into numbers you can act on: win rate, expectancy and average R. Below is what to record, a template table, a worked review of ten trades and a weekly routine.
What to record on every trade
Write the first half before you click and the second half after you exit. A journal filled in from memory a week later records the story you prefer, not what happened.
| Before entry | After exit |
|---|---|
| Market, side (long or short) and product (spot, perp, option) | Exit price and how the trade closed (stop, target, manual, liquidation) |
| Setup name: the reason you are entering, in one line | Net P&L after fees, in dollars |
| Entry, stop-loss and target prices | Result in R: net P&L divided by planned risk |
| Planned risk in dollars (1R) and position size | Did you follow the plan? Yes or no, and where you broke it |
| Leverage and liquidation price (perps) | One lesson, or the mistake you made |
The single most useful field is planned risk in dollars. Without it you cannot express results in R, and without R you cannot compare a $3,000 ETH trade with a $150 SOL trade (both example prices). Set it from the stop distance using the position size calculator.
A trading journal template
Three example entries, using illustrative prices and a 1R of $100 (1% of a $10,000 account). Perp fees are the simulator's 0.05% taker rate per fill; spot is 0.1% per buy or sell. Sizes are $100 divided by the stop distance, rounded down to each coin's quantity step (0.001 ETH, 0.0001 BTC, 0.01 SOL).
| # | Trade | Entry / stop / target | Size | Exit | Net P&L | R | Notes |
|---|---|---|---|---|---|---|---|
| 1 | ETH perp long, breakout | 3,000 / 2,940 / 3,120 | 1.666 ETH | 3,120 (target) | +$194.82 | +1.95R | Followed plan |
| 2 | BTC perp short, rejection at resistance | 80,000 / 81,000 / 78,000 | 0.1 BTC | 81,000 (stop) | -$108.05 | -1.08R | Valid setup, stopped out |
| 3 | SOL spot buy, pullback | 150 / 144 / 162 | 16.66 SOL | 153 (manual) | +$44.93 | +0.45R | Closed early on fear |
How row 2 was worked out: 0.1 BTC is $8,000 of notional at entry and $8,100 at exit. The gross loss is 0.1 x $1,000 = $100. Fees are $4.00 + $4.05 = $8.05, so the net loss is $108.05, or 1.08R. Fees turned a clean -1R into something worse, which is exactly why the journal records net, not gross.
Row 3 shows why the notes column matters. The setup was planned for +2R, and the trader took +0.45R because the price wobbled. One such exit means nothing; ten of them in a month is a pattern costing real expectancy.
The metrics that matter: win rate, average R and expectancy
Once you have at least a couple of dozen trades, calculate these from the R column:
Win rate = winning trades / total trades Average win = sum of winning R / number of wins Average loss = sum of losing R / number of losses Expectancy = win rate x average win - loss rate x average loss Average R = sum of all R / total trades Profit factor = sum of winning R / sum of losing R
Expectancy and average R are the same number calculated two ways: the average result per trade, in units of risk. If it is positive after fees, the process makes money on average; if it is negative, no amount of position sizing will rescue it.
Win rate on its own is misleading. A 70% win rate with small wins and occasional large losses can still lose money. Always read it next to the average win and loss, as covered in risk/reward ratio.
Worked example: reviewing ten trades
Say a week of trading produced these results in R: +2.0, -1.0, +1.5, -1.0, -1.0, +3.0, -0.5, -1.0, +2.0, -1.0.
| Metric | Calculation | Result |
|---|---|---|
| Win rate | 4 wins / 10 trades | 40% |
| Average win | (2.0 + 1.5 + 3.0 + 2.0) / 4 = 8.5 / 4 | +2.125R |
| Average loss | (1.0 + 1.0 + 1.0 + 0.5 + 1.0 + 1.0) / 6 = 5.5 / 6 | -0.917R |
| Expectancy | 0.4 x 2.125 - 0.6 x 0.917 | +0.30R |
| Average R | (8.5 - 5.5) / 10 | +0.30R |
| Profit factor | 8.5 / 5.5 | 1.55 |
With 1R = $100, +0.30R per trade is about $30 a trade, or $300 over the ten. Losing more often than winning was not a problem here because the wins were more than twice the size of the losses.
Ten trades is far too few to trust. Treat early numbers as a rough guide and keep collecting; the averages settle down as the sample grows.
Put fees into the R column
Small stops make fees look large in R terms. Take the ETH long from the template: entry notional is 1.666 x $3,000 = $4,998, so the taker fee is $2.50. If it had hit the stop, exit notional would be 1.666 x $2,940 = $4,898.04 and the fee $2.45.
Gross loss at stop = 1.666 x 60 = $99.96 Fees = 2.50 + 2.45 = $4.95 Net loss = $104.91 -> -1.05R
About 0.05R per round trip disappears to fees on that trade. Halve the stop distance and size up to keep the same $100 risk, and fees roughly double in R terms. Resting limit orders that fill as maker cost 0.02% instead of 0.05%; see maker vs taker fees. Check any single trade with the crypto P&L calculator.
A weekly review routine
- Export or copy the week's closed trades. The trade history on CryptoPaperTrading lists entry, exit, fees, net P&L and how each trade closed, and can be downloaded as CSV.
- Add the R value for each trade: net P&L divided by the risk you planned.
- Recalculate win rate, average win, average loss and expectancy for the week and for all trades so far.
- Group by setup. If one setup has negative expectancy over a meaningful sample, stop trading it or change its rules.
- Count plan breaks: moved stops, early exits, trades with no stop, size above 1R. Note the R they cost.
- Check the losers for stops placed too close or near the liquidation price; see stop-loss.
- Write one rule change for next week, and only one, so you can tell whether it helped.
Once a month, look at the largest losing streak in R and ask whether your account could survive one twice as long at your current risk per trade. If not, cut the risk, as explained in risk management.
Common journal mistakes
- Recording only P&L in dollars. Dollar results change with position size; R results show whether the decisions were good.
- Skipping the losing trades. A journal missing its worst trades will always show a positive expectancy.
- Changing the stop after entry and logging the new one. Record the original stop; that is the risk you actually took.
- Too many fields. If filling it in takes ten minutes, you will stop doing it. The template above is enough.
- Never reviewing. A journal that is written but not read changes nothing.
Paper trading is the ideal place to build the habit. Sign up, take trades on live prices with virtual money, and keep the journal from the first trade so the numbers are there when they matter.
Frequently asked questions
What should a crypto trading journal include?
At minimum: market and side, setup, entry, stop and target, planned risk in dollars, exit and how it closed, net P&L after fees, the result in R and one note on whether you followed the plan.
What is a good expectancy?
Any expectancy that stays positive after fees over a large sample of trades means the process makes money on average. The size you can expect depends on your strategy, so compare your own numbers over time rather than chasing a target.
Is a high win rate enough to be profitable?
No. Profitability depends on win rate and the size of wins versus losses together. A 40% win rate with wins twice the size of losses is profitable before fees; a 70% win rate with losses three times the size of wins is not.
Does CryptoPaperTrading keep a trade history I can use for a journal?
Yes. Closed trades are listed with entry, exit, fees, net P&L and how each one closed, along with your win rate, and the history can be downloaded as CSV. You add the plan and R values yourself.
Should I journal paper trades?
Yes. Journaling with virtual money builds the habit and gives you a sample of trades to measure before any real money is involved.
Practise this: Practise in the crypto trading simulator with $10,000 in virtual USD.