Temporal Performance Tab Guide

~6 min – the time-of-day & day-of-week X-ray that finally reveals whether your edge is real… or just a Monday illusion that dies every Tuesday

You’ve dissected exits in PEE Analytics. You’ve ranked setups in Performance Breakdown. You’ve stress-tested ratios in Performance Ratios.

Now open Temporal Performance (fourth tab in CharterElite, LayoutDashboard icon).

This tab doesn’t care about setups or psychology in isolation. It slices your performance purely by time: hour of entry, day of week, week of month, month of year—with moving averages to cut through noise and autocorrelation to tell you if your returns are trending (momentum), bouncing (mean-reversion), or random.

It’s the tab that answers the question most traders never ask: “When do I actually make money… and when do I just bleed quietly at the same time every week?”

Where to Find It

• Sidebar → CharterElite
• Fourth tab → Temporal Performance (LayoutDashboard icon)
• Expand tab help accordion (top) for definitions + pro tips (focus on moving averages, avoid bad hours/days, spot momentum vs reversion)

Filters – Time Is the Only Variable

Top card:

• Portfolio — All or single
• Setup — Multi-select (or none = all)
• Side — LONG / SHORT / Both
• Period — Daily / Weekly / Monthly (how to bucket the main time-series)
• From / To — Date range (exit_date / date / exit_time / entry_time)

Only completed trades count. Filters recompute everything below instantly.

Metrics Cards (Top 6 – Instant Temporal Snapshot)

When ≥1 period has trades:

• Total Periods — Days/weeks/months with ≥1 trade
• Best Period PnL — Highest single-period net
• Worst Period PnL — Lowest single-period net
• Average Win Rate — Mean of per-period win rates
• Total PnL — Cumulative across all periods
• Avg Period PnL — Total PnL ÷ period count

PnL & Win Rate Over Time Charts

Two dual-line charts:

• PnL Over Time

X: period label (day/week/month) Y: $ PnL Lines: raw per-period PnL + 5-period moving average (dashed) Tooltip: period + exact PnL

• Win Rate Over Time

X: same periods Y: 0–100% Lines: raw per-period win % + 5-period moving average (dashed) + 50% baseline (grey) Tooltip: period + win rate

Moving averages smooth noise—upward trend = improving consistency; downward = decay.

Best & Worst Trading Times (Hour-of-Day Cards)

Four small cards (requires entry_time on trades):

• Best hour by win rate
• Worst hour by win rate
• Best hour by avg PnL
• Worst hour by avg PnL

Each: hour (e.g. 14:00), metric value, trade count.

Performance by Hour of Day

Bar chart (dual Y-axis):

• Bars per hour (0–23)
• Left: Win Rate % (green/orange/red bands)
• Right: Avg PnL $ (green/red)

Optional table (if exit_quality_score exists): Hour | Avg Quality | Count — sorted by quality descending.

No entry_time? → Message: “Entry time tracking required for hour analysis.”

Autocorrelation Analysis (ACF/PACF) – Momentum or Mean-Reversion?

Shown only if ≥20 completed trades (ordered by exit date/time):

• ACF chart — Autocorrelation at lags 1–20 (bars + 95% confidence bands)

Green positive = momentum (winners follow winners) Red negative = mean-reversion (winners follow losers) Inside bands = random walk

• PACF chart — Partial autocorrelation (removes earlier lags)

• Pattern Alert — One line:

“Momentum (Trend-Following)” / “Mean-Reversion” / “Random Walk” + strength % + short explanation

Use this to validate strategy fit: trend-following needs positive ACF; counter-trend needs negative.

Behavioral Temporal Patterns

(Requires taxonomy tags on planning/entry/execution + entry_time/date)

Table: behaviors that appear ≥2× average frequency in specific hour ranges (Early Morning 6–9, etc.) or days of week.

• Behavior (tag name)
• Time Period (hour range or weekday)
• Frequency
• Multiplier (e.g. 3.2×)

Top 10 by multiplier. High multiplier = “you do this stupid thing every Tuesday at 10 a.m.”

No patterns? → Tag harder.

Performance by Market Regime & Day of Week

(Requires market_regime + entry_time/date)

Table: Regime × Day of Week cells (≥3 trades):

• Regime (bull/bear/sideways/etc.)
• Day (Monday–Sunday)
• Count | Win Rate | Avg PnL

Colored green/red for win rate & PnL. Sorted by win rate descending.

No regime data? → Message: “Market regime classification required.”

Quick Workflow – Time-Based Leak Hunting

1. Open Temporal Performance → set filters + Period (start with Weekly)
2. Read Metrics Cards + PnL / Win Rate charts → moving average trend up or down?
3. Check Best & Worst Trading Times + Hour of Day bars → avoid worst hours?
4. ≥20 trades? → Read Autocorrelation → momentum or reversion? Align strategy.
5. Tagged? → Behavioral Temporal Patterns → kill the Tuesday 10 a.m. FOMO habit
6. Regime-tagged? → Regime × Day table → avoid bear Fridays?
7. Action: add time filter to setups (“No entries 14:00–16:00”) or tag bad hours → re-run Council

Quick Reality Checks

• Moving average flat/down? → Consistency dying—check behavioral patterns
• Worst hour consistent? → Session fatigue/news overlap—avoid it
• Autocorrelation random? → No serial dependence—random entries, not strategy
• No patterns? → Tag planning/entry/execution harder
• All green hours? → Small sample—wait for more data

Next: Episode 27 – Full Flywheel Mastery: Integrating Weekly Alpha → Calendar Analytics → CharterElite (Temporal Performance, Performance Ratios, PEE, Exit DNA) → Council (Kill List, Reality Check, Setup DNA) → Evolution into the self-reinforcing compounding loop that turns monthly audits into weekly micro-kills and quarterly A-grade upgrades—without ever increasing risk % or chasing new shiny setups.

Or open CharterElite → Temporal Performance right now. Set Weekly period + last 6 months.

Look at the Win Rate Over Time moving average.

Flat or down?

That’s your edge decaying one week at a time.

Now check Behavioral Temporal Patterns.

See the multiplier on “FOMO Entry” at 14:00?

That’s your Tuesday killer.

Tag it. Filter it. Kill it.

The calendar doesn’t lie. It just waits for you to notice.

Notice it now.

Make next week different.

One hour. One pattern. One kill.

Your grade is watching. 😈