← Skills Library
Content & Retro

Trade Retro

Pull recent trades, calculate P&L by strategy, identify winning/losing patterns, and store a structured retro to your knowledge base.

How It Works

Trade Retro · Workflow
Confirmed fills → P&L → patterns → operationalized rules.
▸TriggerPost-session review · /trade-retro --period 7d / --strategy <name>
1
Pull trade data
Confirmed exchange fills / trades DB — not bot logs
2
Calculate P&L
Per-trade + aggregate: win rate · avg win/loss · expectancy
3
Pattern analysis
What worked · what lost · execution-rule deviations
4
Write the retro
Performance · lessons · rules for next session · confidence
✓Store + operationalize · Akashic (trading/episodic) · make rule changes NOW
  1. Post-session review: /trade-retro --period 7d / --strategy <name>
  2. Pull trade data: Confirmed exchange fills / trades DB — not bot logs
  3. Calculate P&L: Per-trade + aggregate: win rate · avg win/loss · expectancy
  4. Pattern analysis: What worked · what lost · execution-rule deviations
  5. Write the retro: Performance · lessons · rules for next session · confidence
  6. Store + operationalize: Akashic (trading/episodic) · make rule changes NOW
ↆ download card

Invocation Triggers

/trade-retrotrading retrotrade reviewP&L review

Use Cases

  • Post-session trade review across all active strategies
  • Identify patterns in winning vs losing trades
  • Build a compounding knowledge base of trading lessons

The Problem

You close the session up or down and move on, and the lesson evaporates. Win rate looks fine, so you call the strategy working — without ever checking whether a seventy-percent win rate is being eaten alive by a three-to-one loss-to-win size. The same execution mistake shows up retro after retro because it was noted but never fixed. P&L gets eyeballed from bot logs that report intent, not settlement. None of it compounds, so you keep paying the same tuition.

What It Does

  1. 1
    Pull confirmed trade data

    Read trades from confirmed fill data — the exchange or the trades database, not bot log estimates. Bot logs report intent; fills report what actually settled.

  2. 2
    Calculate P&L and expectancy

    Per trade: entry, exit, size, absolute and percent P&L, hold duration, triggering signal. Aggregate to total P&L, win rate, average win versus average loss, and largest win and loss — expectancy, not just win rate.

  3. 3
    Separate winner and loser patterns

    What made the winners work — market conditions, signal strength, hold-duration sweet spot. What caused the losers — entry against trend, ignored stop, oversized position, news event.

  4. 4
    Audit execution honestly

    Did you follow the strategy rules, exit when the signal said to, cut losses fast or hold hoping. The execution-mistakes section is the highest-value one — empty two retros running means the retro is not being honest.

  5. 5
    Write the structured retro

    Performance, what worked, what did not, execution mistakes, market conditions, rules to apply next session, and confidence in the strategy with the reason.

  6. 6
    Store and operationalize now

    Save to the knowledge base as category trading, type episodic, tagged trade-retro. If a lesson implies a rule change, make it now — and any mistake appearing three retros running gets a systemic fix, not another note.

What You Get / What It Doesn't Do

What you get
  • A P&L breakdown from confirmed fills: total, win rate, average win versus average loss, expectancy
  • Winner and loser pattern analysis tied to entry conditions and signal quality
  • An execution-mistakes audit — where you deviated from your own rules
  • A structured retro with specific rules to apply next session and a confidence rating
  • The retro stored to the knowledge base, tagged trading and trade-retro for next-session recall
What it doesn't do
  • Trust bot log P&L — it pulls from confirmed exchange fills or the trades DB, stated explicitly
  • Call a strategy working on win rate alone — it computes expectancy or it is incomplete
  • Let a recurring mistake slide — three retros running demands a systemic fix, not a note
  • Defer the rule change — a lesson that does not change config or behavior this session is decoration

Tips

Expectancy beats win rate

A seventy-percent win rate with a three-to-one loss-to-win size is a losing strategy. Always check average win against average loss before you trust the win rate.

An empty mistakes section is a tell

If you followed the rules perfectly two retros in a row, either the rules need tightening or the retro is not being honest. Audit those entries explicitly.

Change the rule in the same session

A lesson written down and acted on next session never gets acted on. If it implies a config, code, or behavior change, make the change before you close the retro.

Get the Skill

Elite SkillELITE

Unlock the full Trade Retro SKILL.md — drop it into ~/.claude/skills/ and trigger it by name.

What you unlock
  • A P&L breakdown from confirmed fills: total, win rate, average win versus average loss, expectancy
  • Winner and loser pattern analysis tied to entry conditions and signal quality
  • An execution-mistakes audit — where you deviated from your own rules
  • A structured retro with specific rules to apply next session and a confidence rating
...
Skills Library

29 more production skills ready to install.

Browse All Skills