Dashboard guide

Every number, explained

What each statistic on your dashboard measures, exactly how Candela works it out, and what to do about it. No jargon left undefined.

How to read this

Every figure on the dashboard comes from trades you have closed — nothing is estimated, sampled or filled in. Open positions are excluded until they have an outcome, so a number only moves when something actually finished.

Candela is a per-account journal: every figure describes ONE trading account, in that account's own currency, and should match that broker's statement. The picker at the top right chooses the account; there is deliberately no combined view, because money in different currencies doesn't add and blended accounts stop describing your decisions.

Everything also respects the date range beside it. Change either control and every card recomputes. If a figure disagrees with your broker, the range is almost always the reason.

Days are grouped in your own timezone, which you can change in Preferences. Trades are dated by when they closed, not when they opened.

The four headline numbers

The row across the top of your dashboard. Each one compares against the previous period of the same length, so a 30-day view is measured against the 30 days before it.

Exampleinvented figures

Net P&L

Everything the account's closed trades made or lost in the selected period, after fees — shown in the account's own currency.

How it's worked out
The sum of net profit and loss across every trade that closed inside the range. Open positions are excluded entirely — a trade only counts once it has an outcome. Fees and commissions are already deducted.
How to read it
This is the number the rest of the page explains, and it should reconcile with the account's broker statement. If it disagrees, check the date range first: your broker is usually showing account-to-date while Candela is showing the window you selected.
Exampleinvented figures

Trade win %

The share of your closed trades that ended positive.

How it's worked out
Winning trades divided by all closed trades. Break-even trades count in the denominator but not as wins, so they pull the percentage down.
How to read it
Half the story on its own. A 40% win rate with large winners beats 70% with large losers, every time. Always read it next to profit factor — one without the other tells you almost nothing.
Exampleinvented figures

Profit factor

How much you make for every unit you lose.

How it's worked out
Gross profit divided by gross loss. Both figures are absolute totals, so one enormous winner can carry it — which is why consistency is scored separately.
How to read it
Above 1.0 you are net profitable. 1.5 and up is a genuine edge. ∞ means you had no losing trades in the period, which almost always means the sample is too small to trust.
Exampleinvented figures

Max drawdown

The deepest peak-to-valley fall in your equity for the period.

How it's worked out
Peak to valley on your equity curve, with deposits and withdrawals taken out. Each cash flow ends one stretch and starts the next, so paying money into a losing run cannot shrink the number — only trading moves it. The figure in currency is the fall from your high-water mark; the percentage is that fall as a share of the peak it fell from.
How to read it
The most giving-back you actually sat through. Smaller is better. This is the number that decides whether you can hold a position without your judgement going with it — and the one prop firms fail people on.

The equity curve

Your closed trades in the order they happened, totalled as they go.

Exampleinvented figures

Equity curve

Your running profit and loss, trade by trade.

How it's worked out
Every closed trade in the period, sorted by close time, added cumulatively. The x-axis is trade order rather than calendar time, so a busy day takes more width than a quiet week.
How to read it
Look at the shape, not the endpoint. A staircase means a repeatable process. A flat line with one cliff means a single trade is doing the work. A deep valley that recovers tells you what your drawdown tolerance actually is.

Net daily P&L

One bar per day you traded, green above the line and red below.

Exampleinvented figures

Net daily P&L

What each trading day finished at, and how the days compare.

How it's worked out
Trades are grouped by the calendar day they closed, in your timezone, and netted. Days you did not trade are left out rather than drawn flat, so weekends and rest days don't stretch the chart into empty space.
How to read it
The headline is your average trading day, not the period total — the total is already in Net P&L above. Watch the balance of green to red days and the size of the red ones. Frequent small reds among larger greens is a healthy pattern; one red the size of five greens is a risk problem, not a strategy problem.

How every number is made

Every figure on the dashboard, with the arithmetic that produces it and the fields it is read from. The worked examples are not typed out — they are the real functions run on the example book above, so this page cannot quote a number the code has stopped producing.

Two conventions are worth knowing, because tools disagree about both. A trade is one netted position, not one fill, so a position closed in three parts counts once. And the daily series covers every business day since the account was funded, counting an idle day as 0% — a day holding nothing is a real observation of no risk taken, and leaving it out roughly doubles every ratio built on top.

What happened

Trades

count(positions where net_pnl is set)

Reads
positions
Worked on the example book
80 closed

Volume

sum(size)

Reads
size
Worked on the example book
113.14 lots

Time in market

sum(closed_at - opened_at)

Reads
opened_at, closed_at
Worked on the example book
5d 15h

Exposure

time in market / (last close - first open)

Reads
opened_at, closed_at
Worked on the example book
4.3%

Mind Score

weighted blend over last 30 days: consistency 30, sleep 25, calm 20, energy 15, routine 10; unlocks at 10 journaled days

Reads
daily_checkins (sleep, quality, stress, energy, exercise, meals)
Worked on the example book
consistency 80, sleep 74, calm 65, energy 70, routine 50 -> 71

Mind vs market

avg(day net P&L) per bucket; a day joins a bucket by its check-in answer; both buckets need >= 5 days

Reads
net_pnl by local day, daily_checkins (sleep, stress, exercise)
Worked on the example book
8 rested days avg +$120/day vs 6 short-sleep days avg -$45/day

The money

Net P&L

sum(net_pnl)

Reads
net_pnl (profit + commission + swap + fees)
Worked on the example book
5,819.10

Monthly return

prod(1 + daily r in month) - 1; the year is prod(1 + monthly) - 1

Reads
net_pnl, deposits & withdrawals
Worked on the example book
+10.00% then +9.09% -> 1.10 x 1.0909 - 1 = +20.00%

Win rate

winners / closed x 100

Reads
net_pnl
Worked on the example book
38 / 80 x 100 = 47.50%

Profit factor

sum(winners) / |sum(losers)|

Reads
net_pnl
Worked on the example book
17,029.19 / 11,210.09 = 1.52

Expectancy

win rate x avg win + loss rate x avg loss

Reads
net_pnl
Worked on the example book
0.475 x 448.14 + 0.525 x -266.91 = 72.74

Avg win / avg loss

avg winner / |avg loser|

Reads
net_pnl
Worked on the example book
448.14 / 266.91 = 1.68

Recovery factor

net profit / deepest money drawdown

Reads
net_pnl
Worked on the example book
5,819.10 / 3,655.36 = 1.59

Consistency

best green day / sum of all green days x 100

Reads
net_pnl, closed_at, your timezone
Worked on the example book
3,517.01 / 10,684.05 x 100 = 32.92%

Return and risk

Time-weighted return

product(1 + r_i) - 1, cut at every deposit and withdrawal

Reads
net_pnl, closed_at, account ledger
Worked on the example book
29.10% across 106 days

Daily return series

one return per business day from funding, idle days = 0%

Reads
net_pnl, closed_at, account ledger
Worked on the example book
106 days, of which 89 idle

Ann. volatility

stdev(r) x sqrt(252)

Reads
daily return series
Worked on the example book
0.01944 x 15.87 = 30.85%

Sharpe

mean(r) / stdev(r) x sqrt(252)

Reads
daily return series
Worked on the example book
0.00259 / 0.01944 x 15.87 = 2.11

Sortino

mean(r) / stdev(losing days only) x sqrt(252)

Reads
daily return series
Worked on the example book
0.00259 / 0.00789 x 15.87 = 5.21

Smart Sharpe

Sharpe / autocorrelation penalty

Reads
daily return series
Worked on the example book
2.11 -> 2.08 (Sortino 5.21 -> 5.13)

Calmar

annualised return / max drawdown

Reads
daily return series
Worked on the example book
83.52% / 13.41% = 10.91

VaR (95%)

the worst 5% of days, taken from history — no distribution assumed

Reads
daily return series
Worked on the example book
worst 6 of 106 days -> -0.47%

cVaR (95%)

mean of every day at or beyond the VaR day

Reads
daily return series
Worked on the example book
mean of those 6 = -2.35%

Skew

third standardised moment of r

Reads
daily return series
Worked on the example book
4.47

Kurtosis

fourth standardised moment of r, minus 3

Reads
daily return series
Worked on the example book
34.27

Curve linearity

r^2 of a straight line fitted to the compounding curve

Reads
daily return series
Worked on the example book
0.664

Drawdown

Max drawdown

max over t of (peak - index) / peak

Reads
net_pnl, closed_at, account ledger
Worked on the example book
13.41%

Current drawdown

(peak - index now) / peak

Reads
net_pnl, closed_at, account ledger
Worked on the example book
5.29%

Avg drawdown

mean depth of every fall below the high-water mark

Reads
daily return series
Worked on the example book
2.29%

Avg drawdown length

mean length of those falls, in days

Reads
daily return series
Worked on the example book
9.6 days

The score

The Candela score

sum of (component / 100 x weight), over the six below

Worked on the example book
8.8 + 8.7 + 11.9 + 4.8 + 13.7 + 10.2 = 58

Profit factor

clamp(0, 100, (value - 1) / (2.5 - 1) x 100)

Reads
profitFactor
Worked on the example book
1.5235/100 → 8.8 of 25 points

Win rate

clamp(0, 100, (value - 30) / (60 - 30) x 100)

Reads
winRate
Worked on the example book
48%58/100 → 8.7 of 15 points

Avg win / avg loss

clamp(0, 100, (value - 0.5) / (2 - 0.5) x 100)

Reads
payoff
Worked on the example book
1.6879/100 → 11.9 of 15 points

Recovery factor

clamp(0, 100, (value - 0) / (5 - 0) x 100)

Reads
recovery
Worked on the example book
1.5932/100 → 4.8 of 15 points

Drawdown control

clamp(0, 100, (value - 50) / (10 - 50) x 100)

Reads
drawdown
Worked on the example book
13%91/100 → 13.7 of 15 points

Consistency

clamp(0, 100, (value - 60) / (20 - 60) x 100)

Reads
consistency
Worked on the example book
33%68/100 → 10.2 of 15 points

The Candela score

One 0–100 read on how you trade rather than how much you made. It is deliberately blind to position size, so a careful small account can outscore a reckless large one.

Exampleinvented figures

The six inputs

Each is scored 0–100 against a fixed band, then contributes its weight to the total. The weights add up to 100.

Profit factor

Scored 1.00 → 2.50 · worth 25 of 100 points

Add up every winning trade, add up every losing trade, divide the first by the second. At 1.0 you are exactly break-even before costs. Below 1.0 you are paying the market to trade. It carries the most weight of the six because it is the only one that accounts for both how often you win and how much you win by — the other five mostly explain WHY this number is what it is.

Win rate

Scored 30% → 60% · worth 15 of 100 points

The share of your closed trades that finished green. It is the most quoted number in trading and the most misleading on its own, which is why it is worth 15 points rather than 25. A 40% win rate with winners three times the size of your losers is a strong business; an 80% win rate with one catastrophic loser is not. Read it against avg win / avg loss, never alone.

Avg win / avg loss

Scored 0.50 → 2.00 · worth 15 of 100 points

Your average winning trade divided by your average losing trade, also called the payoff ratio. At 1.0 your winners and losers are the same size, so you need to win more than half the time to make anything. Above 1.0 you can be wrong more often than you are right and still finish ahead. This is the number that moves when you cut losers faster or let winners run.

Recovery factor

Scored 0.00 → 5.00 · worth 15 of 100 points

Net profit divided by your deepest drawdown. It answers a question raw profit cannot: what did that profit cost you in pain? Two traders can both finish the period up 5,000 — one never down more than 500, the other down 4,000 at the worst point. The first has a recovery factor of 10, the second 1.25. The first has a business, the second got lucky.

Drawdown control

Scored 50% → 10% given back · worth 15 of 100 points

Your deepest fall from a peak, measured as a share of that peak rather than in currency, so it means the same thing on a 1,000 account and a 100,000 one. If you built up to 4,000 and then fell to 3,000 before recovering, you gave back 25%. This is the metric prop firms fail people on, and the one that decides whether you can hold a position without your judgement going with it. Lower is better, so the band runs downward.

Consistency

Scored 60% → 20% from one day · worth 15 of 100 points

How much of your total profit came from your single best day. This is the prop-firm consistency rule, and it exists to catch a specific illusion: a month that looks profitable but is really one enormous day surrounded by mediocrity. If your best day is 60% of your profit, you do not have a repeatable edge yet — you have one good session and a lot of noise. Lower is better, so the band runs downward.

What the bands mean

  • 85+Elite Every part of this is working. Protect it.
  • 70+Strong A real edge, executed well. Keep the size honest.
  • 55+Consistent The edge is there. Your weakest metric is the one to work on.
  • 40+Developing Something works — it's being given back elsewhere.
  • 0+Fragile The maths isn't paying yet. Start with your lowest metric.

Why it stays hidden at first

The score does not appear until you have 20 closed trades in the selected range. A win rate over five trades is noise, and a confident-looking 90 built on noise is worse than no score at all.

Until then the radar still shows the six inputs so you can see what is being measured while the sample builds. Note that the score follows your date range — a seven-day view will rarely have enough trades to qualify.

What, when and where the money moves

Three cards that break the same P&L down by instrument, by session and by the tags you attach when journaling.

Exampleinvented figures

What you trade

Your trades split by instrument.

How it's worked out
Every trade in the period grouped by symbol, showing count and net P&L. The top five are listed individually and the rest collapse into 'Other'.
How to read it
Concentration is not automatically bad, but it should be deliberate. If most of your volume is in an instrument that isn't in your top earners, that gap is worth explaining.
Exampleinvented figures

When you trade

Your trades split by market session.

How it's worked out
Each trade is assigned to Asia, London, New York or After hours based on its OPEN time in your timezone. A trade opened in London and closed in New York counts as London.
How to read it
Most traders have one session that pays and one that quietly costs. If the busiest session isn't the one that pays best, you have found something to change this week.
Exampleinvented figures

Edges & leaks

Which setups pay you, and which mistakes cost you.

How it's worked out
Setups come from the setup tag on each trade; leaks come from the mistake tags. Both are summed by net P&L, so a tag on many small trades can outrank a tag on one big one. Untagged trades appear in neither.
How to read it
This card is only as good as your tagging. If it looks empty or wrong, the fix is in how you journal, not in the numbers.

The journal itself

Two cards that measure the habit rather than the trading, and price what the habit is worth.

Exampleinvented figures

Journal every trade

The share of your trades in the period that have been journaled.

How it's worked out
Trades with a journal entry divided by all trades in the range. The target is 100% — this is a habit meter, not a performance one.
How to read it
Every other insight on this page degrades when this number drops. An unjournaled trade still counts in your P&L but contributes nothing to understanding why.
Exampleinvented figures

What your journal is worth

What following your plan, and your mood going in, are worth in money.

How it's worked out
Trades where you marked 'followed the plan' are netted against those where you didn't. The same is done for trades entered calm or steady versus anxious or tilted. Both comparisons need trades on each side to appear at all.
How to read it
This is the closest thing to a direct price on your discipline. If off-plan trades are net negative and on-plan trades are net positive, the gap between them is what your process is worth per period.

Execution

The numbers under the numbers — position sizing, holding times, streaks, and your two extreme days.

Exampleinvented figures

The execution grid

Position sizing, holding times, streaks, your average win and loss, and your two extreme days.

How it's worked out
All computed across closed trades in the period. Average hold is the mean time between open and close. Win streak / skid is the longest consecutive run of each, in close order. The two days at the foot of the card are the highest and lowest daily nets, with the setups and mistakes recorded on that day's trades.
How to read it
Look for mismatches. A largest position several times your average usually marks the trade that also produced your largest loss. A payoff ratio below 1.0 means you need to win more than half the time just to stand still. And compare the two days: if the best is far larger than everything around it, your consistency score will reflect that.

Risk

Everything here is built from one series: your return each day on the capital that was actually in the account that day. A deposit ends one measurement and starts the next, so paying money in cannot flatter a single figure below it.

Exampleinvented figures

Return per unit of risk

Annualised volatility, Sharpe, Sortino, their Smart variants and Calmar — five ways of asking what your profit cost you in risk.

How it's worked out
Volatility is the standard deviation of your daily returns, annualised over 252 trading days. Sharpe divides your mean daily return by it. Sortino divides by the deviation of the losing days only, on the grounds that upside movement is not risk. The Smart variants widen the denominator when your returns are autocorrelated — a run of days that follow one another is less independent evidence than its count suggests, so the ratio honestly shrinks. Calmar divides annualised return by your deepest drawdown.
How to read it
Sortino sitting above Sharpe means most of your volatility is upside, which is a good sign rather than a rounding error. A Smart figure well below its plain version means your results arrive in streaks, so the plain one is flattering you. Calmar answers the question a prop firm asks: what did you make relative to the worst hole you sat in? Under 60 trading days none of these appear, because a ratio built from a handful of days is a number rather than information. For scale: a Sharpe above 2 held over years is rare, and anything far higher usually means the sample is short rather than the trading exceptional.

Tails: VaR, cVaR, skew and kurtosis

How bad your bad days get, and whether the shape of your returns is hiding a rare disaster.

How it's worked out
VaR is the fifth-percentile day of your own history — no distribution is assumed, these are days that actually happened. cVaR is the mean of every day at or beyond it, which is what the tail costs once it arrives. Skew measures whether your returns lean positive or negative. Kurtosis measures how fat the tails are, where zero is a normal distribution and positive means extremes are more common than normal would predict.
How to read it
cVaR is always worse than VaR, and the gap between them is the point: VaR gives you the threshold, cVaR tells you what is waiting behind it. Negative skew with high kurtosis is the classic dangerous profile — many small wins and a rare large loss, the shape that makes an equity curve look wonderful right up until it does not. If that describes yours, the answer is in your stop discipline rather than your entries.

The shape of the curve

How deep your drawdowns usually run, how long they last, and how far below your high-water mark you are right now, and how straight your equity curve is.

How it's worked out
Every stretch where the curve sits below its own high-water mark counts as one drawdown episode. Average drawdown is the mean depth of those episodes; average length is their mean duration in trading days. Curve linearity is the r-squared of a straight line fitted to your compounding equity curve, from 0 to 1.
How to read it
Max drawdown tells you the worst thing that happened once; average drawdown tells you what normal feels like. The long episodes are the ones that break people — a deep fall recovered inside two days is far easier to sit through than a shallow one that grinds on for six weeks. Linearity near 1 means steady compounding; a low figure alongside good profit usually means a single trade carried the period.

The calendar

The month at a glance, coloured by each day's net result.

Exampleinvented figures

Calendar

Every day of the month with its net P&L and trade count.

How it's worked out
Days are grouped in your timezone by trade close time. Green and red shading is by direction only, not magnitude, so a small green day and a large one look the same. The month total sits above the grid.
How to read it
Use it for rhythm rather than size. Clusters of red often line up with a particular weekday or the days right after a big win — patterns that are invisible in a single total.

These statistics describe trades you have already closed. They are a record and a teaching tool, not a prediction and not advice — Candela does not recommend trades, forecast markets or tell you what to do next. Past performance says nothing about future results.

Something here still unclear? Visit the Help Centre.