Best day and worst day are the highest and lowest cumulative P&L over a single trading day. They are two figures with no nuance, and that is precisely what makes them useful: they point at the days when something went outside the ordinary.
What the ratio compares
Contents
Both figures come from the same calculation: add up the results of every trade in each day, which gives a daily P&L, and keep the maximum and the minimum. It is one of the simplest statistics in a journal, and one of the most revealing.
Their point is not to measure a performance but to spot abnormal days. A good method produces days that resemble each other; a method going off the rails produces extreme days. The two bounds say straight away which case you are in, and the rest of the statistics do not.
The worst day is by far the more instructive of the two. Compared with the daily risk you had planned to take, it directly measures adherence to your own rules: if it is worth three times your limit, it was not the market that was violent, it was the limit that was not held.
Worst day = min of daily P&L — Best day = max of daily P&L
The day is the one the trade closed in, not the one it opened in. A trade held overnight therefore falls in the day it ended.
Fees are included, as everywhere else: a high-trade day can be negative purely because of commissions.
These two figures are not averages. Each describes a single day, so they compare neither across accounts nor across periods without being related to something.
| Worst day against your limit | What it says |
|---|---|
| under 1× | The limit held. That is the normal case, and it is rarer than people think. |
| 1 to 1.5× | Moderate overshoot, often a gap or slippage rather than a decision. |
| 1.5 to 3× | The rule was worked around: extra trades, bigger size, stop pushed. |
| above 3× | A day of lost control. One of these days often wipes out several weeks. |
| no limit defined | The figure compares to nothing, and that is the real problem to fix first. |
The last row is the most frequent and nobody talks about it. Without a daily loss limit written in advance, the worst day is just an anecdote: there is nothing to compare it with, so nothing to conclude. Setting that limit is the prerequisite that makes the statistic readable, and it takes one decision.
The second point is that the best day reads differently from the worst. You relate it to the period's total P&L: if it represents a large share, the performance rests on a single day, which is the same robustness check as removing your biggest winner from a profit factor. A quarter carried by one day is not a quarter.
The worst day is worth 3.85 times the planned daily limit, and five and a half times an ordinary losing day. Neither comparison speaks about the market: they say that on one day, the rule was abandoned. That is the information this term serves to produce, and it requires no interpretation.
This is the main use, and the only one producing an action. Your daily loss limit is a number; your worst day is another. The ratio of the two says in a second whether the rule exists in fact or only in the plan.
The point is that there is nothing to interpret. A threefold overshoot has no acceptable market explanation: it is a day on which you carried on past the moment you had planned to stop.
Relate your best day to the period's total P&L. If it is 40% of it, the method did not produce the result, one day did, and nothing says it will happen again.
It is the same reasoning as profit factor's robustness check, applied to the calendar rather than to trades. Both take a minute and avoid building on an accident.
The first flaw is that these figures are displayed without a reference. A worst day at −$2,310 means nothing on its own: on a $200,000 account it is an ordinary day, on a $15,000 one it is a disaster, and with no written daily limit it is simply a number. Tools display them raw, which makes them decorative.
The second is that they are very sensitive to the chosen period. The worst day of a year and that of a month are not the same information, and comparing two traders looking at different windows compares nothing. It is a statistic that must always carry its period alongside it.
The third is subtler: the worst day does not say what happened. A gap on an overnight position and a day of revenge trading give the same figure, and call for two opposite corrections. This figure signals, it does not diagnose, and believing otherwise leads to fixing the wrong thing.
The classic mistake. Looking at the worst day without having a daily limit. The figure is then comparable to nothing, and you settle for finding it big or small depending on your mood. The limit written in advance is what turns this statistic into a check, and its absence is the real problem the figure reveals.
| Indicator | Formula | What it measures | When to use it |
|---|---|---|---|
| Worst day | min of daily P&L | The worst day, in currency | Checking a daily limit holds |
| Best day | max of daily P&L | The best day, in currency | Seeing whether the period rests on one day |
| Biggest loss | the worst result on a TRADE | A blown stop, not a day | Spotting an outlier trade |
| Maximum drawdown | (peak − trough) ÷ peak | A decline cumulated over several days | Sizing positions |
Without it, your worst day is a number with no reference and the statistic is useless. With it, the ratio of the two becomes an immediate check, and it is the only thing to do to turn a decorative figure into information.
The worst day gives a date. Open that day's journal and look at what it is made of: a gap, twenty trades instead of five, a doubled size. That is where the correction is, and it is never in the figure itself.
If it is more than a quarter of your P&L, the period rests on one day. That is not a reason to throw it away, it is a reason not to conclude anything definitive before living through three more without that day.
They are the highest and lowest cumulative P&L over a single trading day, fees included. You add up the results of every trade in each day, and keep the two extremes of the period.
To check that a daily loss limit is held. Related to that limit, it says immediately whether the rule exists in fact: a worst day worth three times the limit has no market explanation, it is a day on which you carried on past the moment you planned to stop.
That is the problem to fix before looking at the figure. With no limit written in advance, your worst day compares to nothing and produces no conclusion. One sentence is enough: past such an amount lost, the day is over.
The biggest loss concerns a single trade, the worst day a whole day. A worst day made of twenty small losses signals a frequency or discipline problem; one made of a single blown stop signals a risk-per-trade problem. The two figures cross, and it is that crossing which says what to fix.
Not necessarily. Relate it to the period's total P&L: if it represents a large share, your performance rests on a single day, and nothing guarantees it will repeat. It is the same robustness check as removing your biggest winning trade from a profit factor, applied to the calendar.
Best day and worst day are read in the Track record widget on the dashboard, one green and one red, computed on each day's P&L, fees included. They recompute on every sync, with nothing to type in.
The Reports screen supplies the references they usually lack, and that is what changes the reading: it exposes average daily P&L, average P&L of a losing day, daily win rate and max losing-day streak. Comparing the worst day with the average losing day is exactly the check described above, and both figures are in the same table.
To go deeper on this: read the full guide. Every term is gathered in the glossary.