Case: compare two trade lists

A total profit is not a verdict on a plan. To compare two trade lists fairly, put the same four questions to both, using everything from this module:

1. The average trade after costs (Expectancy: the average trade). Does a trade from this plan make money on average?
2. The worst drawdown (Drawdown: why losses hit harder than gains). How deep did the worst stretch go, in percent of the peak?
3. The number of trades. Is it at least 30, or still too few to tell?
4. A steady risk per trade. Did each trade risk about the same amount, so the results describe one plan rather than a mix of small and big bets?

Each question covers a blind spot of the others. A high average can hide a deep drawdown; a mild drawdown can rest on a dozen trades; and any average is meaningless if the bet size kept changing. In this lesson, Nora's 12 trades (journal A) and Marc's 45 trades (journal B) on the Pebble practice account go through all four.

Totals from lists of different lengths can't be compared directly: 1,104.27 from 12 trades and 391.86 from 45 trades are different kinds of numbers. Divide each by its number of trades to get the expectancy, the average result of one trade after costs.

• Journal A (Nora): 1,104.27 ÷ 12 = +92.02 a trade.
• Journal B (Marc): 391.86 ÷ 45 = +8.71 a trade, after 253.16 of fees.

On the surface A is ten times better per trade, and B's average is thin: before costs it would be (391.86 + 253.16) ÷ 45 = 14.33, so fees take almost 40% of it.

This is the first question, not the last. An average only means something if it comes from enough trades and from trades of a similar size; both conditions still have to be checked before A's 92.02 can be called better than anything.

A journal shows the risk of each trade even when nobody wrote it down: planned risk = distance from the entry to the stop × size, the same arithmetic as in Position size from risk and stop.

In journal A, trade 1 risked (37.88 − 36.61) × 118 = 149.86. Trade 8, on NWF, risked (48.34 − 47.05) × 620 = 799.80, more than five times as much: from about 1.5% of a 10,000 account to 8%. Trade 8 was also marked plan_followed = no. It exited at 46.53, below its stop, and lost 1,149.00, about 1.4 times what was planned.

Journal B never risked more than 100 on any of its 45 trades, and every trade followed the plan.

Checking a few rows this way takes minutes and changes how the whole list reads. A journal whose risk swings fivefold isn't testing one plan; it is testing the trader's moods.

An expectancy describes a plan only if every trade in it was the same plan at roughly the same risk. Mix small and big bets, or trades where the plan was followed with trades where it wasn't, and the average describes the mix.

In journal A the planned risk ranges from 149.86 to 799.80, so the dollar result of each trade depends mostly on how big the bet was. Trade 8, the biggest, broke the plan and lost 1,149.00. Its timing matters too: it came straight after a 413.38 loss on trade 7, with the risk doubled. That is chasing a loss, not sizing up.

So A's +92.02 a trade isn't a property of a repeatable plan. With the same entries and exits at a fixed 1%, the average would be a different number, and the trades where the plan broke wouldn't belong in it at all.

The fix is boring and effective: a fixed risk per trade (module 3) and the plan followed on every trade, so the next journal measures one thing.

The second question is what living with a plan felt like at its worst: the deepest fall from a peak, as in Drawdown: why losses hit harder than gains.

Journal A peaked at 11,788.22 after trade 6. Then trade 7 lost 413.38 and trade 8 lost 1,149.00. Together that is 1,562.38, and 1,562.38 ÷ 11,788.22 = 13.3% of the peak, gone in two trades.

Journal B's worst drawdown over 45 trades was 539.92, or 5.4%, built up over a run of five losses in a row. B had more losing runs simply because it had more trades, and each of them stayed shallow because each loss was capped at about 100.

Two things to notice. First, A's drawdown is more than twice B's despite four times fewer trades, so with more trades A would likely meet deeper holes. Second, the depth came mostly from one oversized trade that broke the plan, which is exactly what the risk question flagged.

The third question checks whether there is enough evidence to judge at all. The rule from Expectancy: the average trade applies: below 30 trades the honest label is "too few to tell", and even well above 30 a result is a first reading, not proof.

Journal A has 12 trades. Nine wins in twelve feel convincing, but a dozen trades can come out that way by luck, as Marc's own first ten trades (two wins, then a positive plan over forty) showed in the opposite direction. A's results simply can't be judged yet, for or against.

Journal B has 45 trades at a steady risk, past the 30-trade floor. That supports "worth continuing on the practice account", not "this plan works". The next 45 trades may tell a somewhat different story.

Length also affects the other questions: a longer list has more chances to meet a bad stretch, so A's drawdown on 12 trades probably understates what it would face over 45.

Put the four answers side by side:

• Average after costs: A +92.02 a trade, B +8.71.
• Worst drawdown: A 13.3% in two trades, B 5.4% over five losses.
• Number of trades: A 12 (too few to tell), B 45 (a first reading).
• Risk per trade: A from 149.86 to 799.80 with one plan break, B at most 100 with the plan followed every time.

Only the first answer favours A, and it is the one the other three undermine. B is the plan with evidence behind it, so B is the one worth continuing on the practice account. A is not shown to be bad; it is shown to be unjudgeable until its risk is fixed and its trades are counted.

A higher total and a higher win rate (75% for A against 42% for B) don't make a better plan. Win rate and reward-to-risk showed why a win rate alone means little; this comparison shows the same for a total.

The last step is the one people skip: checking what could be wrong with the plan you just chose. Being hard on the list you dislike is easy; the skill is being equally hard on the one you like.

For journal B, three doubts are fair. Its edge is thin: 8.71 a trade after 253.16 of fees, so a slightly higher cost per trade could erase it. All 45 trades come from one stretch of the market, and a calm period may not repeat. And 45 trades is a first reading, not proof. Two things are not doubts: a 42% win rate is fine when wins outsize losses, and a 5.4% drawdown needs only about +5.7% to recover.

So the honest verdict is "B is worth another 45 trades on the practice account", not "B works".

How it fits together: the module's four numbers (average after costs, worst drawdown, number of trades, steady risk) judge any list of trades. Module 5, What is a price chart telling me?, starts reading the prices behind the trades in Candles and timeframes.