More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    AI Trading Journal Edge & Mistake Analyzer

    1

    Its purpose is not to report superficial statistics such as: Win Rate = 58% and stop there.

    $9.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    AI Trading Journal Edge & Mistake Analyzer

    AI Trading Journal Edge & Mistake Analyzer

    Example session with this skill installed

    TRADING JOURNAL

    File
    Atlas_Trading_Journal_2026.xlsx

    Account
    Fictional $50,000 Futures Evaluation

    Period
    2026-01-02 through 2026-06-30

    Timezone
    America/New_York

    Trades
    327

    Instruments

    NQ
    MNQ
    ES
    MGC

    STRATEGY

    Name
    Atlas Intraday v4

    Primary Setups

    Trend Pullback
    Breakout Retest
    VWAP Reversion
    Failed Breakout

    Setup Grades

    A
    B
    C

    DAILY RULES

    Maximum Trades per Day
    4

    Daily Loss Stop
    -2.0R

    Cooldown After 2 Consecutive Losses:
    15 minutes

    Allowed Trading Window
    09:30–14:30 New York

    C-Grade Setups
    Not permitted in live trading

    ENTRY RULE

    Trend Pullback requires

    Trend aligned.
    Pullback complete.
    Confirmation candle closed.
    Momentum condition valid.

    A trade taken before confirmation-candle close is defined as:

    Premature Entry

    EXIT RULE

    Do not manually exit a valid winner before:

    Target
    Stop
    Explicit Invalidation

    unless risk policy requires otherwise.

    JOURNAL FIELDS

    trade_id
    date
    entry_time
    exit_time
    instrument
    direction
    setup
    setup_grade
    session
    entry_price
    exit_price
    stop_price
    target_price
    initial_risk
    net_pnl
    r_multiple
    mae_r
    mfe_r
    confirmation_complete
    rule_compliant
    rule_violations
    confluence_tags
    emotion_before
    exit_reason
    minutes_since_previous_trade
    trades_today_before
    loss_streak_before
    win_streak_before
    notes

    SESSION DEFINITIONS

    New York Open
    09:30–10:00

    New York AM
    10:00–11:30

    Midday
    11:30–13:00

    New York PM
    13:00–14:30

    REQUEST

    Analyze the entire journal.

    I specifically want

    Most Profitable Setups
    Weakest Setups
    Best and Worst Sessions
    Long vs Short
    Instrument Performance
    Premature Entries
    Premature Exits
    Overtrading
    Possible Revenge / Loss-Response Patterns
    Rule Violations
    Performance After Winning Streaks
    Performance After Losing Streaks
    Confluence Combinations
    Profit Giveback
    Trade Number in Day
    Performance After 2 Consecutive Losses

    Do not stop at headline metrics.

    I want specific conditional findings.

    For example

    "What percentage of New York AM losses came from premature entries?"

    Also calculate

    Count-Based Loss Attribution
    Dollar / R-Based Loss Attribution

    Build

    Edge Map
    Mistake Map
    Rule Compliance Comparison
    Streak Table
    Confluence Table
    Performance Attribution
    Top Action Priorities
    Retest Plan

    Do not invent emotional or behavioral labels that are not supported by the journal.

    • Read your context and instructions
    • Compiled the ai trading journal

    === AI TRADING JOURNAL EDGE & MISTAKE ANALYSIS ===

    Period
    2026-01-02 through 2026-06-30

    Account
    Fictional $50,000 Futures Evaluation

    Strategy
    Atlas Intraday v4

    Trades Reviewed
    327

    Data Quality
    HIGH

    Key Strengths

    Net P&L is complete.
    R-multiple is available.
    Confirmation state is available.
    Rule compliance is available.
    MAE and MFE are available.
    Streak state is available.
    Confluence tags are available.

    Primary Limitation

    Emotion tags are incomplete and therefore revenge-trading conclusions are treated cautiously.

    === EXECUTIVE SUMMARY ===

    Strongest Historical Edge

    A-Grade Trend Pullback
    +
    Higher-Timeframe Alignment
    +
    New York AM

    Trades
    64

    Expectancy
    +0.27R

    Profit Factor
    1.81

    Confidence
    HIGH

    Largest Performance Leak

    Premature Entry

    Premature entries represented

    21.4% of all trades

    but:

    49.8% of all gross loss R.

    Most Important Conditional Finding

    73.3% of New York AM losing trades were entered before confirmation.

    Those trades generated

    61.2% of New York AM gross loss R.

    Most Important Behavioral Pattern

    After two consecutive losses, rule-violation frequency rose from:

    12.1%

    to:

    38.7%

    and next-trade expectancy fell from:

    +0.08R baseline

    to:

    -0.21R.

    Priority Action

    Strengthen the confirmation gate and test a mandatory enforcement mechanism after two consecutive losses.

    === BASELINE PERFORMANCE ===

    Trades
    327

    Wins
    176

    Losses
    145

    Breakeven
    6

    Win Rate
    53.8%

    Net Result
    +21.6R

    Average Winner
    +0.83R

    Average Loser
    -0.68R

    Expectancy
    +0.066R per trade

    Profit Factor
    1.29

    Maximum Drawdown
    -9.4R

    Maximum Consecutive Losses
    5

    === PROCESS QUALITY MATRIX ===

    GOOD PROCESS / GOOD OUTCOME

    Trades
    151

    Net
    +63.7R

    GOOD PROCESS / BAD OUTCOME

    Trades
    91

    Net
    -56.1R

    BAD PROCESS / GOOD OUTCOME

    Trades
    25

    Net
    +13.2R

    BAD PROCESS / BAD OUTCOME

    Trades
    60

    Net
    -39.2R

    Interpretation

    A significant share of profitable trades came from poor process.

    Those winners should not be interpreted as evidence that rule violations are beneficial.

    === RULE COMPLIANCE ===

    Rule-Compliant Trades

    242

    Expectancy
    +0.18R

    Profit Factor
    1.53

    Net
    +43.5R

    Rule-Violating Trades

    85

    Expectancy
    -0.26R

    Profit Factor
    0.71

    Net
    -21.9R

    Historical Difference

    +65.4R between the aggregate compliant and violating trade groups.

    This does not represent a causal counterfactual.

    It demonstrates that historically the strategy performed materially better when the defined process was followed.

    === SETUP ANALYSIS ===

    1. A-GRADE TREND PULLBACK

    Trades
    91

    Expectancy
    +0.22R

    Profit Factor
    1.67

    Net
    +20.0R

    Classification

    VERIFIED HISTORICAL EDGE

    1. BREAKOUT RETEST

    Trades
    78

    Expectancy
    +0.14R

    Profit Factor
    1.39

    Net
    +10.9R

    Classification

    CANDIDATE / MODERATE HISTORICAL EDGE

    1. VWAP REVERSION

    Trades
    83

    Expectancy
    +0.02R

    Profit Factor
    1.06

    Net
    +1.7R

    Classification

    FRAGILE EDGE

    1. FAILED BREAKOUT

    Trades
    42

    Expectancy
    -0.12R

    Profit Factor
    0.78

    Net
    -5.0R

    Classification

    NO EDGE DETECTED IN CURRENT SAMPLE

    1. C-GRADE TRADES

    Trades
    33

    These trades were not permitted under the supplied live-trading rules.

    Expectancy
    -0.20R

    Net
    -6.6R

    Rule Status
    PROCESS VIOLATION

    === EDGE MAP ===

    EDGE

    A-Grade Trend Pullback + HTF Alignment / NY AM

    Trades
    64

    Expectancy
    +0.27R

    PF
    1.81

    Confidence
    HIGH

    EDGE

    Breakout Retest / New York Open

    Trades
    39

    Expectancy
    +0.19R

    PF
    1.52

    Confidence
    MEDIUM-HIGH

    EDGE

    NQ Long / Trend Pullback

    Trades
    52

    Expectancy
    +0.24R

    Confidence
    HIGH

    === SESSION ANALYSIS ===

    NEW YORK OPEN

    Trades
    79

    Expectancy
    +0.12R

    Net
    +9.5R

    NEW YORK AM

    Trades
    118

    Expectancy
    +0.15R

    Net
    +17.7R

    Best Session
    NEW YORK AM

    MIDDAY

    Trades
    61

    Expectancy
    -0.04R

    Net
    -2.4R

    NEW YORK PM

    Trades
    69

    Expectancy
    -0.05R

    Net
    -3.2R

    Worst Session
    NEW YORK PM

    === WHY NEW YORK AM IS STRONG ===

    New York AM is not profitable simply because of the session.

    Its strongest performance comes from

    Trend Pullback
    Breakout Retest
    Higher-Timeframe Alignment

    Rule-compliant NY AM trades

    Expectancy
    +0.25R

    Rule-violating NY AM trades

    Expectancy
    -0.31R

    Interpretation

    The underlying session contains historical edge, but poor execution destroys a substantial portion of it.

    === PREMATURE ENTRY ANALYSIS ===

    Premature Entries
    70

    Percentage of All Trades
    21.4%

    Expectancy
    -0.24R

    Net
    -16.8R

    Confirmed Entries
    257

    Expectancy
    +0.15R

    Net
    +38.4R

    Loss Count Attribution

    Premature entries represented

    55.2% of all losing trades.

    Gross Loss R Attribution

    Premature entries represented

    49.8% of total gross loss R.

    === NEW YORK AM PREMATURE ENTRY FINDING ===

    New York AM Losing Trades
    30

    Premature-Entry Losses
    22

    Count-Based Loss Attribution

    22 / 30

    73.3%

    New York AM Gross Loss
    -21.4R

    Premature-Entry Gross Loss
    -13.1R

    R-Based Loss Attribution

    13.1 / 21.4

    61.2%

    Finding

    73.3% of New York AM losing trades were entered before confirmation, accounting for 61.2% of gross loss R in the session.

    Confidence
    HIGH

    Interpretation

    The session itself is not the primary problem.

    The main historical leak inside New York AM is entry timing.

    === MAE COMPARISON ===

    Confirmed Entries

    Average MAE
    0.48R

    Premature Entries

    Average MAE
    0.81R

    Difference

    Premature entries experienced approximately 69% more adverse excursion before exit.

    This supports the interpretation that early execution is materially degrading trade location.

    === PREMATURE EXIT ANALYSIS ===

    Tagged Premature Exits
    37

    Actual Realized Result
    +4.8R

    Rule-Based Historical Counterfactual
    +9.7R

    Estimated Historical Difference
    +4.9R

    Classification

    COUNTERFACTUAL ESTIMATE

    Most Premature Exits Occurred

    After One Prior Loss
    During New York AM Winners
    When MFE Exceeded +1R

    Interpretation

    Exit anxiety is not inferred.

    The supplied data shows early manual exits under these conditions, but emotional cause is not established.

    === LONG VS SHORT ===

    LONG

    Trades
    181

    Expectancy
    +0.10R

    PF
    1.41

    SHORT

    Trades
    146

    Expectancy
    +0.02R

    PF
    1.08

    Initial Interpretation

    Longs appear materially stronger.

    Confounding Check

    72% of long trades occurred in:

    NQ and ES

    while:

    41% of shorts occurred in MGC and lower-performing Failed Breakout setups.

    Adjusted Interpretation

    Direction contributes to the difference, but instrument and setup mix materially confound the raw long-vs-short comparison.

    === INSTRUMENT PERFORMANCE ===

    NQ

    Trades
    139

    Expectancy
    +0.16R

    PF
    1.48

    Classification
    STRONGEST

    MNQ

    Trades
    81

    Expectancy
    +0.08R

    PF
    1.24

    Classification
    POSITIVE

    ES

    Trades
    63

    Expectancy
    +0.05R

    PF
    1.16

    Classification
    POSITIVE BUT MODEST

    MGC

    Trades
    44

    Expectancy
    -0.09R

    PF
    0.83

    Classification
    WEAK

    === TRADE NUMBER IN DAY ===

    Trade #1

    Expectancy
    +0.15R

    Trade #2

    Expectancy
    +0.13R

    Trade #3

    Expectancy
    +0.03R

    Trade #4+

    Expectancy
    -0.16R

    Rule-Violation Rate on Trade #4+:

    47.9%

    Finding

    Later trades show a substantial deterioration in both expectancy and compliance.

    This is stronger evidence of historical overtrading than raw trade frequency alone.

    === OVERTRADING ===

    Defined by supplied rules as

    More than 4 trades per day
    or
    trading after daily stop
    or
    re-entry without a new valid setup.

    Overtrading Trades
    39

    Expectancy
    -0.28R

    Net
    -10.9R

    Percentage of Total Trades
    11.9%

    Percentage of Gross Loss
    27.6%

    Confidence
    HIGH

    === PERFORMANCE AFTER LOSSES ===

    After 1 Loss:

    Trades
    99

    Expectancy
    +0.03R

    Violation Rate
    18.2%

    After 2 Consecutive Losses:

    Trades
    31

    Expectancy
    -0.21R

    Violation Rate
    38.7%

    Average Re-Entry Delay
    8.4 minutes

    After 3+ Consecutive Losses:

    Trades
    14

    Expectancy
    -0.29R

    Violation Rate
    50.0%

    === TWO-LOSS COOLDOWN ===

    User Rule

    15-minute cooldown after 2 losses.

    Compliant Cooldown Trades

    Trades
    15

    Expectancy
    +0.07R

    Cooldown Violations

    Trades
    16

    Expectancy
    -0.47R

    Finding

    Historical performance after two losses differed sharply depending on whether the cooldown rule was followed.

    Confidence
    MEDIUM

    Reason

    The subgroup contains only 31 trades and should be validated prospectively.

    === POSSIBLE LOSS-RESPONSE PATTERN ===

    The following cluster appears after two or more losses:

    Shorter Re-Entry Delay
    Higher Rule-Violation Rate
    More C-Grade Setups
    Higher Average Risk

    This is classified as

    HIGH-CONFIDENCE LOSS-RESPONSE PATTERN

    not:

    Confirmed Revenge Trading

    because explicit emotional labels are incomplete.

    === PERFORMANCE AFTER WINS ===

    After 1 Win:

    Expectancy
    +0.11R

    After 2 Wins:

    Expectancy
    +0.09R

    After 3+ Wins:

    Expectancy
    -0.04R

    Average Risk Increase
    +21%

    C-Grade Setup Frequency
    2.1x baseline

    Interpretation

    Performance deteriorates after extended winning streaks while risk and lower-grade setup participation increase.

    Possible mechanism

    Reduced selectivity.

    Confidence
    MEDIUM

    === CONFLUENCE ANALYSIS ===

    Trend Pullback Base

    Expectancy
    +0.12R

    Trend Pullback + HTF Alignment:

    Expectancy
    +0.25R

    Incremental Lift
    +0.13R

    Trend Pullback + Volume:

    Expectancy
    +0.16R

    Incremental Lift
    +0.04R

    Trend Pullback + HTF + Volume:

    Expectancy
    +0.27R

    Incremental Lift vs HTF only
    +0.02R

    Finding

    Higher-timeframe alignment provides the strongest incremental historical value.

    Volume adds comparatively little once HTF alignment is already present.

    === PROFIT GIVEBACK ===

    Days Reaching At Least +2R:
    21

    Average Intraday Peak
    +2.46R

    Average Final Result
    +1.18R

    Average Giveback
    1.28R

    Primary Historical Sources

    Trade #4+
    New York PM
    Trades After Earlier Loss
    C-Grade Trades

    === MISTAKE MAP ===

    1. PREMATURE ENTRY

    Trades
    70

    Net
    -16.8R

    Severity
    CRITICAL

    Confidence
    HIGH

    1. OVERTRADING

    Trades
    39

    Net
    -10.9R

    Severity
    HIGH

    Confidence
    HIGH

    1. TWO-LOSS COOLDOWN VIOLATION

    Trades
    16

    Expectancy
    -0.47R

    Severity
    HIGH

    Confidence
    MEDIUM

    1. C-GRADE SETUPS

    Trades
    33

    Net
    -6.6R

    Severity
    HIGH

    Confidence
    HIGH

    1. PREMATURE EXIT

    Historical Estimated Drag
    -4.9R

    Severity
    MEDIUM

    Classification
    COUNTERFACTUAL ESTIMATE

    === PARETO LOSS ANALYSIS ===

    Premature Entries
    +
    Overtrading
    +
    C-Grade Setups

    overlap on some trades.

    After removing duplicate attribution, these behaviors were present in:

    31.8% of unique trades

    and:

    67.4% of gross loss R.

    === STRATEGY VS EXECUTION ATTRIBUTION ===

    Trend Pullback

    Compliant
    Strong positive expectancy

    Violating
    Negative expectancy

    Classification
    STRONG STRATEGY / WEAK EXECUTION IN VIOLATING SUBSET

    Failed Breakout

    Compliant
    Still negative

    Classification
    WEAK STRATEGY / EXECUTION NOT PRIMARY CAUSE

    === TOP ACTION PRIORITIES ===

    PRIORITY 1

    Problem
    Premature Entry

    Why
    Largest recurring loss concentration.

    Action
    Enforce confirmation completion before entry.

    Measure
    Premature-entry rate
    Expectancy
    MAE
    Gross-loss attribution

    Confidence
    HIGH

    PRIORITY 2

    Problem
    Performance deterioration after two consecutive losses.

    Action
    Test strict enforcement of the existing 15-minute cooldown.

    Measure
    Violation rate
    Re-entry delay
    Expectancy
    Setup grade

    Confidence
    MEDIUM

    PRIORITY 3

    Problem
    Trade #4+ deterioration.

    Action
    Audit whether trade #4+ should require A-grade setup or no further trading.

    This should initially be treated as a test hypothesis rather than a permanent rule.

    Measure
    Expectancy
    Violation rate
    Daily giveback

    Confidence
    HIGH

    PRIORITY 4

    Problem
    C-Grade setups.

    Action
    Enforce existing prohibition consistently.

    Measure
    C-grade frequency
    Net R
    Compliance

    Confidence
    HIGH

    PRIORITY 5

    Problem
    Premature winner exits.

    Action
    Track exit reason and MFE for the next 30 eligible winners.

    Do not change target logic yet.

    Confidence
    MEDIUM

    === RETEST PLAN ===

    Experiment 1:
    Confirmation Enforcement

    Sample
    Next 30 qualifying Trend Pullback / Breakout trades

    Primary Metric
    Expectancy

    Secondary
    MAE
    Win Rate
    Loss Attribution
    Rule Compliance

    Experiment 2:
    Two-Loss Cooldown

    Sample
    Next 20 qualifying post-two-loss situations

    Compare
    Cooldown respected vs violated

    Primary
    Next-trade expectancy

    Experiment 3:
    Trade #4+ Selectivity

    Test
    Only A-grade trade #4+

    Compare against historical baseline.

    === JOURNAL DATA IMPROVEMENTS ===

    Add

    Explicit reason for every manual exit.

    Why

    Improves premature-exit attribution.

    Add

    Explicit emotion-before field on every trade rather than optional notes.

    Why

    Allows stronger separation between behavioral labels and inferred loss-response patterns.

    Add

    Market regime.

    Why

    Allows determination of whether setup weakness comes from regime mismatch.

    === FINAL CONCLUSION ===

    The journal contains a measurable historical edge.

    The strongest evidence is concentrated in

    A-Grade Trend Pullbacks
    Higher-Timeframe Alignment
    New York AM
    NQ

    The largest performance leakage comes from:

    Premature Entries
    Later-Day Overtrading
    C-Grade Setups
    Rule Degradation After Consecutive Losses

    The most important finding is not the overall win rate.

    It is that a large share of losses is concentrated in specific, identifiable, and potentially controllable behaviors.

    Historical analysis only.

    These associations do not guarantee future performance or prove psychological causality.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    AI Trading Journal Edge & Mistake Analyzer is a premium post-trade analytics and performance-attribution agent designed to identify exactly where a trader's historical edge is created, weakened, or destroyed. Its purpose is not to report superficial statistics such as: Win Rate = 58% and stop there. Its purpose is to discover deeper conditional relationships such as: 73% of New York AM losing trades were entered before the strategy's confirmation condition was complete, representing 61% of gross loss dollars in that session. This difference makes the agent useful for serious traders who want to improve process quality rather than simply monitor P&L. The agent can analyze: CSV Trade Histories Excel Trading Journals Broker Exports TradingView Trade Reports Prop-Firm Trading Logs Manually Maintained Journals Structured Trade Notes Daily Review Notes Setup Labels Strategy Versions Rule-Compliance Checklists Confluence Tags Market-Regime Tags Session Tags Emotion Tags Entry Reasons Exit Reasons Screenshots Referenced by the User MAE / MFE Data Trade Sequence Data The analytical process is: Inspect Data → Normalize Schema → Validate Trade Integrity → Normalize Time and Sessions → Normalize P&L and R-Multiples → Map Setups → Map Rules → Map Behavioral Tags → Establish Baseline Performance → Segment Results → Detect Edge Concentration → Detect Mistake Concentration → Analyze Entry Timing → Analyze Exit Timing → Analyze Overtrading → Analyze Loss-Response Patterns → Analyze Rule Violations → Analyze Streak Effects → Analyze Confluence Combinations → Quantify Historical Impact → Assess Confidence → Build Edge Map → Build Mistake Map → Rank Action Priorities → Define Retest Plan The agent treats each trade as a combination of: Market Context Strategy Setup Execution Decision Risk Decision Behavioral State Outcome This prevents outcome bias. A winning trade can contain poor process. A losing trade can still represent excellent execution. The skill therefore distinguishes: GOOD PROCESS / GOOD OUTCOME GOOD PROCESS / BAD OUTCOME BAD PROCESS / GOOD OUTCOME BAD PROCESS / BAD OUTCOME This distinction is critical because profitable rule violations can reinforce bad habits while valid losing trades can incorrectly be interpreted as mistakes. The agent can calculate baseline metrics such as: Total Trades Net P&L Gross P&L Fees Win Rate Average Winner Average Loser Payoff Ratio Expectancy Expectancy in R Profit Factor Median Trade Maximum Winner Maximum Loser Maximum Drawdown Drawdown Duration Maximum Consecutive Wins Maximum Consecutive Losses Average Holding Time Trade Frequency Average Trades per Day Profitable Days Losing Days However, baseline statistics are only the starting point. The primary objective is conditional performance attribution. The agent analyzes trading setups individually. For each setup it can calculate: Trade Count Win Rate Net P&L Average R Median R Profit Factor Drawdown Contribution Average MAE Average MFE Rule-Violation Rate Premature-Entry Rate Premature-Exit Rate Session Distribution Instrument Distribution It does not rank setups using win rate alone. A setup with: 55% Win Rate 0.7 Payoff Ratio may be structurally weaker than one with: 42% Win Rate 2.1 Payoff Ratio The agent therefore evaluates: Expectancy Profit Factor Average R Sample Size Drawdown Stability Outlier Dependence Rule Compliance before classifying a setup. Possible setup classifications include: VERIFIED HISTORICAL EDGE CANDIDATE EDGE FRAGILE EDGE NO EDGE DETECTED INSUFFICIENT EVIDENCE "Verified historical edge" means supported by the supplied historical sample. It does not guarantee future profitability. Session analysis is another major capability. The agent can compare: Asia London New York AM New York PM Regular Trading Hours Overnight Pre-Market Power Hour Custom User-Defined Sessions Session analysis requires correct timezone normalization. It can calculate: Trades Net P&L Expectancy Win Rate Profit Factor Average R Rule-Violation Rate Premature-Entry Rate Premature-Exit Rate Overtrading Rate Drawdown Contribution Setup Distribution Instrument Distribution Instead of saying: New York AM is your best session. the agent can produce a more actionable conclusion such as: New York AM generated 62% of total net profit while representing only 38% of trades. Most of the advantage came from Trend Pullback and Breakout Retest setups, while 73% of losing NY AM trades were associated with premature confirmation. This identifies both the edge and the leak inside the same segment. The agent performs long-versus-short analysis. It compares: Trade Count Net P&L Expectancy Profit Factor Average R Win Rate Drawdown Violation Rate Setup Mix Session Mix Instrument Mix It also checks for confounding. For example: Long trades appear significantly stronger than short trades. However: 78% of long trades occurred in NQ during bullish sessions, while short trades were concentrated in Gold and low-performing sessions. The correct conclusion is therefore not simply: Longs are better. The agent identifies when direction is confounded by instrument, setup, session, or regime. Instrument analysis supports: Futures Forex Stocks ETFs Crypto Indices Commodities For each instrument it can calculate: Trades Net P&L Expectancy Expectancy in R Profit Factor Win Rate Drawdown Contribution Average MAE Average MFE Rule-Violation Rate Setup Distribution Session Distribution Fees as Percentage of Gross Edge When instruments have different tick values, contract multipliers, lot sizes, or account sizes, the agent avoids naive raw-dollar comparisons. When initial risk is available, it prefers normalized measures such as: R-Multiple Average R Expectancy in R Return per Unit Risk The skill performs detailed premature-entry analysis. Premature entry can be identified only when supported by: Explicit Confirmation State Strategy Rules Journal Tag Signal-Bar Requirement Breakout Confirmation Rule Retest Requirement Moving-Average Alignment Session Trigger User Note The agent never invents premature entries. If confirmation cannot be reconstructed, it states: Premature entry cannot be measured reliably from the available fields. When available, it compares: Premature Entries Confirmed Entries using: Trade Count Win Rate Expectancy Average R Net P&L Loss Contribution MAE MFE Session Distribution Setup Distribution It can calculate both: Count-Based Loss Attribution and: Dollar-Based Loss Attribution. For example: 22 of 30 losing New York AM trades were premature. Count-Based Loss Attribution: 73% Those trades produced $3,410 of the session's $5,590 gross losses. Dollar-Based Loss Attribution: 61% This level of attribution is a primary differentiator of the skill. Premature-exit analysis can use: Exit Reason Journal Tags Rule-Based Exit Requirements MFE Target Data Post-Exit Movement Manual Close Notes Possible metrics include: Actual Realized R Maximum Favorable Excursion MFE Capture Ratio Rule-Compliant Exit Estimate Premature-Exit Historical Drag Counterfactual estimates are explicitly labeled. Example: Actual tagged early-exit performance: +4.2R Historical rule-compliant counterfactual: +9.8R Estimated historical drag: -5.6R Classification: COUNTERFACTUAL ESTIMATE The agent never reports estimated P&L as if it were actual realized P&L. The skill performs overtrading analysis. It does not equate high trade frequency with overtrading. A legitimate high-frequency strategy may trade often. Overtrading means: Trading beyond the strategy or risk process. Possible evidence includes: Trades Above Daily Maximum Trades After Session Cutoff Trades After Daily Stop Repeated Re-Entries Without a New Setup Trades Outside Approved Setups Rapid-Fire Trades Low-Grade Setups After Earlier Losses User-Tagged Overtrading The agent can analyze trade sequence: Trade #1 Trade #2 Trade #3 Trade #4+ Example: Trades #1–2: +0.14R expectancy Trade #3: +0.03R Trades #4+: -0.12R Trades #4+ also contained: 64% of all rule violations. This helps distinguish whether poor performance occurs because the market changed or because trader selectivity deteriorated. The agent can analyze: Minutes Since Previous Trade to identify rapid re-entry patterns. Possible transparent buckets include: Under 5 Minutes 5–15 Minutes 15–30 Minutes 30+ Minutes The agent performs cautious revenge-trading and loss-response analysis. It does not diagnose psychological conditions. It does not label every trade after a loss as revenge trading. A stronger revenge-pattern candidate may require multiple pieces of evidence: Previous Trade Was a Material Loss Next Trade Occurred Unusually Quickly Position Size Increased Setup Grade Declined Rule Was Violated Journal Note Recorded Frustration / Revenge / FOMO If evidence is weaker, the agent uses neutral language such as: POSSIBLE LOSS-RESPONSE PATTERN or: LOSS-FOLLOWING RULE-DEVIATION PATTERN instead of overclaiming revenge trading. The agent can calculate: Count Net P&L Average R Win Rate Position-Size Change Time Since Prior Loss Violation Rate Session Loss-Streak Position Rule-violation analysis is another central module. Possible violations include: Premature Entry Oversizing Moved Stop Farther Removed Stop Traded Outside Session Traded Restricted News Exceeded Daily Trade Limit Re-Entered Without New Setup Ignored Regime Filter Premature Exit Held Past Cutoff Ignored Cooldown Prop-Firm Rule Violation For each violation type the skill calculates: Trades Frequency Net P&L Average R Gross Loss Loss Contribution Setup Distribution Session Distribution Instrument Distribution Streak Context It can compare: RULE-COMPLIANT TRADES against: RULE-VIOLATING TRADES Example: Rule-Compliant: 212 trades +16.8R Rule-Violating: 71 trades -8.9R Interpretation: The historical strategy edge remained materially stronger when the trader followed the defined process. The agent can also identify violation clustering. Example: Premature Entry + New York PM + After Two Losses may be materially worse than premature entries generally. Streak analysis examines behavior and performance after: One Win Two Wins Three or More Wins One Loss Two Losses Three or More Losses Metrics include: Next-Trade Expectancy Position Size Risk Amount Setup Grade Rule Compliance Time to Next Trade Session Instrument Example: After a winning trade: +0.12R expectancy After a losing trade: -0.04R After two consecutive losses: -0.19R Rule-violation rate after two losses: 37% Baseline violation rate: 11% This reveals that the performance deterioration may come from trader behavior rather than strategy mechanics. The agent can analyze risk escalation after wins and losses. Possible observations include: Average risk rises 27% after three consecutive wins. Average risk rises 19% after two consecutive losses. These are historical associations. The skill does not automatically diagnose: Overconfidence Revenge Martingale unless evidence supports those labels. Drawdown-state behavior can also be analyzed. Possible states include: At Equity High Mild Drawdown Moderate Drawdown Deep Drawdown The agent can compare: Trade Frequency Risk Size Setup Quality Rule Compliance Instrument Switching Session Switching Expectancy This can reveal whether the trader's process deteriorates during account drawdown. The agent can also analyze current daily P&L state before each trade: Positive Day Flat Day Negative Day Near Daily Stop This enables profit-giveback analysis. When trade sequencing allows it: Daily Peak P&L Final Daily P&L Giveback can be calculated. The agent can then attribute giveback to: Late-Day Trades Trade Number Setup Violation Loss Streak Overtrading Session Example: Days that reached +2R: 18 Average final result: +0.8R Average giveback: 1.2R Primary historical source: Trades #4+ after 13:30. The skill performs confluence analysis. Possible confluence tags include: Trend Alignment Higher-Timeframe Alignment VWAP Volume Support / Resistance Market Structure Momentum Order Flow Session Regime News Liquidity Pullback Quality The agent does not only evaluate single tags. It can evaluate combinations such as: Trend + Higher-Timeframe Alignment Trend + Volume Trend + HTF + Volume Range + VWAP Breakout + Volume + Retest However, it protects against combinatorial overfitting. When many tags exist: Analyze Pairwise Combinations First Require Adequate Sample Use Selected Three-Way Combinations Apply False-Discovery Caution Prefer Economically Plausible Combinations The skill can calculate incremental confluence value. Example: Trend Pullback: +0.08R Trend Pullback + HTF Alignment: +0.17R Trend Pullback + HTF + Volume: +0.18R Interpretation: Higher-timeframe alignment historically added material value. Volume added little additional value once HTF alignment was already present. This can help simplify strategies by identifying redundant confluences. The skill also detects confluence redundancy. For example: EMA Alignment Trend Alignment may represent substantially overlapping information. It avoids counting correlated signals as independent evidence. Day-of-week and time-of-day analysis can be included when samples are sufficient. Possible finding: 58% of all rule-violating losses occurred after 11:30 even though only 29% of trades were taken during that period. The agent can analyze: First Trade Effect Example: After a first-trade loss: Average daily trades increased from 2.6 to 4.1. Rule-violation rate doubled. This can reveal behavioral sequences that basic performance dashboards miss. The skill uses MAE and MFE where available. Maximum Adverse Excursion can help investigate: Entry Timing Stop Placement Premature Entries Maximum Favorable Excursion can help investigate: Exit Timing Winner Giveback Target Efficiency Premature Exits Possible entry-quality result: Premature entries showed: 1.7x greater MAE and: 31% lower MFE than confirmed entries. Possible exit-quality result: Actual realized winners captured only 42% of recorded MFE in a specific setup. The agent does not assume that capturing 100% of MFE is realistic or desirable. The skill can identify late-entry or chasing patterns when evidence supports them. Possible evidence: Large favorable move occurred before entry. Remaining reward/risk deteriorated. MFE after entry was materially smaller. Journal records chase/FOMO behavior. The agent creates a structured Mistake Taxonomy. SETUP MISTAKES Wrong Setup Low-Quality Setup Incomplete Confirmation Regime Mismatch EXECUTION MISTAKES Premature Entry Late Entry Premature Exit Stop Movement Target Deviation RISK MISTAKES Oversizing Inconsistent Risk Adding Outside Plan Trading After Daily Stop BEHAVIORAL MISTAKES Overtrading Loss-Chasing Revenge-Pattern Candidate Impulsive Re-Entry Late-Session Deterioration PROCESS MISTAKES Skipped Checklist Missing Plan Missing Screenshot Missing Review Incomplete Journal Fields Mistakes can be ranked: CRITICAL HIGH MEDIUM LOW based on: Frequency Monetary Cost R Cost Drawdown Contribution Recurrence Effect on Historical Edge The skill builds a required Edge Map. The Edge Map can contain: Edge Source Segment Trades Expectancy Profit Factor Net P&L Stability Confidence Why It Matters Possible edges include: Trend Pullback in NY AM Short Breakout on Gold Long Setup with HTF Alignment First Two Trades of Day Rule-Compliant Setup A High-Volatility Breakout The skill also builds a required Mistake Map. The Mistake Map contains: Mistake Trade Count Frequency Net P&L Loss Contribution Average R Most Common Context Severity Action Priority It can reveal: A small number of recurring mistakes may account for a disproportionate share of total historical losses. Example: Premature Entries + Trades After Daily Trade #3 represented: 29% of trades but: 68% of gross loss. This supports a Pareto-style improvement process. The skill separates strategy weakness from trader execution weakness. Example 1: Setup B remains negative even when fully rule-compliant. Interpretation: STRATEGY WEAKNESS Example 2: Setup A is strongly positive when rules are followed and negative when the trader violates entry rules. Interpretation: EXECUTION WEAKNESS The agent can summarize this through: STRONG STRATEGY / STRONG EXECUTION STRONG STRATEGY / WEAK EXECUTION WEAK STRATEGY / STRONG EXECUTION WEAK STRATEGY / WEAK EXECUTION The skill can perform outlier-dependence analysis. It can calculate contribution from: Largest Winner Top 3 Winners Top 5 Winners If removing a few trades destroys profitability: flag: OUTLIER DEPENDENCE It can also analyze: Profit Concentration Loss Concentration Setup Concentration Instrument Concentration The agent can evaluate temporal stability. Possible segmentation: Month Quarter First Half vs Second Half Rolling Windows A finding that repeats across several months is more credible than one driven by a single week. The agent can compare: Full Sample Recent 20 Trades Recent 50 Trades Recent 30 Days to detect: Strategy Drift Trader Drift Compliance Drift Example: Compliance Rate: 92% in Q1 71% in Q2 Expectancy: +0.16R in Q1 +0.02R in Q2 This may justify investigating whether performance deterioration is associated with process deterioration. The agent can perform historical counterfactual scenarios. Examples: Remove Trades After Daily Stop Remove Premature Entries Remove C-Grade Setups Apply Rule-Compliant Exit to Tagged Premature Exits Every simulation is labeled: HISTORICAL COUNTERFACTUAL SCENARIO It is not presented as actual history or guaranteed future performance. The skill includes strict double-counting protection. One trade can contain: Premature Entry Oversizing Loss-Streak Context Overtrading The loss cannot simply be added independently to every category. The agent can therefore distinguish: Primary Mistake Secondary Mistakes Multi-Label Analysis Unique-Trade Attribution It can build an attribution waterfall only when categories are mutually exclusive or appropriately adjusted. The final improvement plan prioritizes controllable, high-impact findings. Each priority can include: Problem Evidence Historical Impact Trigger Rule Pre-Trade Countermeasure In-Trade Countermeasure Post-Trade Review Measurement Retest Period The skill frames changes as measurable experiments. Example: TEST: For the next 30 qualifying New York AM setups, allow entries only after the documented confirmation state is complete. MEASURE: Premature-entry rate Expectancy MAE MFE Rule compliance Loss attribution This is preferable to claiming: This will fix your trading. The skill can produce weekly and monthly reviews. A Weekly Review may include: Strongest Edge Largest Performance Leak Compliance Trend Best Session Worst Session Top Setup Streak Behavior One Primary Focus for Next Week A Monthly Review can add: Strategy Drift Behavioral Drift Regime Changes Setup Stability Action-Plan Progress Recent vs Long-Term Performance The final analysis should prioritize conclusions in the following form: Finding Evidence Magnitude Confidence Likely Interpretation Action How to Validate This turns the journal into a structured improvement system rather than a passive record of wins and losses.

    AI Trading Journal Edge & Mistake Analyzer is only as reliable as the underlying journal and trade-history data.

    Broker exports generally contain executed trades only.

    They do not reveal:

    Valid Setups the Trader Correctly Skipped
    Valid Setups the Trader Missed
    Invalid Setups the Trader Avoided

    unless those events are separately journaled.

    This creates selection bias.

    Premature-entry analysis requires:

    A Defined Confirmation Rule
    Explicit Confirmation State
    Journal Tag
    or Equivalent Evidence

    Without these inputs, the agent cannot reliably determine whether an entry was early.

    Premature-exit analysis may require:

    Exit Reason
    Strategy Exit Rule
    Target
    MAE / MFE
    Post-Exit Data

    Without this evidence, exit-quality conclusions should remain limited.

    Revenge trading cannot be diagnosed from P&L alone.

    A quick trade after a loss may be perfectly valid.

    Stronger inference requires multiple signals such as:

    Short Re-Entry Delay
    Rule Violation
    Position-Size Increase
    Lower Setup Quality
    Journal Notes

    Overtrading is strategy-dependent.

    A high-frequency strategy can legitimately produce many trades.

    Trade count alone is not proof of overtrading.

    MAE and MFE data can be distorted if calculated inconsistently across platforms.

    Historical strategy performance can change over time.

    A setup that was historically profitable may deteriorate because of:

    Market Regime Changes
    Strategy Changes
    Execution Changes
    Instrument Changes
    Trader Behavior

    Historical associations do not prove causality.

    Example:

    Trades after two losses may perform poorly because of rule violations, changing market conditions, or both.

    Small samples can produce unstable conclusions.

    A setup with 8 trades should not receive the same confidence as a setup with 150.

    Confluence analysis creates multiple-comparison risk.

    When many tags and combinations are examined, some apparently strong combinations can emerge by chance.

    Discovery results should therefore be validated on later data where possible.

    Journal labels can be inconsistent.

    Examples:

    NY AM
    New York Morning
    NY Morning

    may refer to the same session.

    Normalization should be documented.

    Timezone errors can materially distort session analysis.

    Daylight-saving changes may also matter.

    Raw dollar P&L can distort comparisons across instruments, account sizes, and contract sizes.

    R-normalization requires reliable initial-risk data.

    If initial risk is missing, R should not be fabricated.

    Strategy-version changes can invalidate aggregated analysis.

    Different versions should be separated when logic materially changed.

    Counterfactual scenarios are hypothetical.

    Removing all premature entries from historical data does not prove that the resulting historical equity curve would have occurred exactly as modeled.

    One trade can contain multiple mistakes.

    Without careful attribution, mistake costs can be double-counted.

    Behavioral notes may be incomplete or subjective.

    The agent should not infer psychological conditions or diagnoses.

    Historical Edge Maps and Mistake Maps describe patterns found in the supplied journal.

    They do not guarantee that those patterns will persist in future trading.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 2 months ago

    Permissions

    Read Files
    Browser
    Write Files

    File Scopes

    **/*.md
    **/*.txt
    **/*.csv
    **/*.tsv
    **/*.json
    **/*.jsonl
    **/*.yaml
    **/*.yml
    **/*.xml
    **/*.html
    **/*.pdf
    **/*.doc
    **/*.docx
    **/*.xls
    **/*.xlsx
    **/*.ods
    **/*.png
    **/*.jpg
    **/*.jpeg
    **/*.webp
    **/trading/**
    **/trading-journal/**
    **/trading-journals/**
    **/journal/**
    **/journals/**
    **/trade-history/**
    **/trade-histories/**
    **/trades/**
    **/broker/**
    **/broker-exports/**
    **/tradingview/**
    **/prop-firm/**
    **/accounts/**
    **/strategies/**
    **/strategy/**
    **/strategy-versions/**
    **/setups/**
    **/sessions/**
    **/instruments/**
    **/futures/**
    **/forex/**
    **/fx/**
    **/stocks/**
    **/equities/**
    **/etfs/**
    **/crypto/**
    **/entries/**
    **/exits/**
    **/mae/**
    **/mfe/**
    **/risk/**
    **/pnl/**
    **/drawdown/**
    **/streaks/**
    **/winning-streaks/**
    **/losing-streaks/**
    **/overtrading/**
    **/revenge-trading/**
    **/behavior/**
    **/behavioral/**
    **/psychology/**
    **/rule-compliance/**
    **/rule-violations/**
    **/mistakes/**
    **/confluence/**
    **/confluences/**
    **/tags/**
    **/screenshots/**
    **/reviews/**
    **/daily-review/**
    **/weekly-review/**
    **/monthly-review/**
    **/performance/**
    **/analytics/**
    **/reports/**
    **/audit/**
    **/audits/**
    **/experiments/**
    **/retests/**
    **/templates/**
    **/docs/**
    **/output/**

    Use Read Files to inspect authorized trading-history CSV files, Excel journals, broker exports, TradingView reports, trade notes, setup definitions, strategy rules, daily reviews, rule-violation records, MAE/MFE data, confluence tags, and historical performance reports. Use Write Files only when the user requests creation or updating of derived artifacts such as: Trading Journal Audit Reports Edge Maps Mistake Maps Setup Performance Tables Session Performance Reports Instrument Performance Reports Rule-Violation Reports Streak Analysis Behavioral Pattern Reports Confluence Matrices Performance Attribution Reports Historical Counterfactual Reports Weekly Trading Reviews Monthly Trading Reviews Improvement Experiments Retest Plans Journal Schema Recommendations Use Browser only when current authoritative public information is necessary and authorized. The core journal analysis normally depends on the user's own historical data, so public information should not replace missing journal evidence. Always distinguish: VERIFIED TRADE FIELD A value directly present in the user's source data. DERIVED METRIC A calculation based on source fields. USER-SUPPLIED LABEL A setup, mistake, rule, emotion, or confluence label explicitly provided by the user. INFERRED PATTERN A historical association derived from trade behavior. COUNTERFACTUAL ESTIMATE A modeled historical scenario rather than realized P&L. UNKNOWN A value that cannot be established. REQUIRES JOURNAL ANNOTATION A behavior that cannot be measured reliably without additional fields. Do not invent: Setup Labels Confirmation States Entry Reasons Exit Reasons Rule Violations Emotional States Revenge Trading Overtrading MAE MFE Skipped Setups Valid Untaken Trades Market Regimes Do not classify a trade as a premature entry unless a confirmation rule or explicit journal evidence exists. Do not classify an exit as premature unless an exit rule, explicit tag, or relevant excursion data supports the conclusion. Do not classify every quick trade after a loss as revenge trading. Prefer neutral terms such as: Loss-Response Pattern Loss-Following Rule Deviation unless stronger evidence exists. Do not define overtrading solely as high frequency. Overtrading means trading beyond the trader's defined process, strategy, or risk rules. Always distinguish: Good Process / Good Outcome Good Process / Bad Outcome Bad Process / Good Outcome Bad Process / Bad Outcome A winning rule violation is still a process violation. A losing rule-compliant trade is not automatically a mistake. Use net P&L when available. Do not mix gross and net results silently. When comparing instruments with different contract economics, prefer R or another normalized risk metric when available. For major findings, report both: Count-Based Attribution and, when possible: P&L-Based Attribution Example: 73% of NY AM losing trades involved premature entry. Those trades accounted for 61% of NY AM gross loss dollars. Do not claim causality from historical association. Use language such as: Associated With Concentrated In Historically Coincided With Appears Linked To unless causal evidence exists. Do not sum mistake costs naïvely when one trade belongs to multiple mistake categories. Use unique-trade attribution or explicitly disclose overlap. Counterfactual performance must always be labeled as hypothetical historical analysis. Do not present it as actual P&L. This skill provides post-trade performance analytics and process analysis, not personalized investment advice or guarantees of future profitability.

    Listed2 months ago

    What's inside

    Frequently Asked Questions

    It identifies exactly where your profitability comes from (Edge Map) and where you are leaking capital (Mistake Map) by analyzing deeper conditional relationships, such as how specific behaviors like premature entries or session-timing impact your R-multiple.

    The agent is designed for maximum flexibility, supporting CSV and Excel exports from most major brokers, TradingView trade reports, and structured manual notes including custom tags for emotions or market regimes.

    Yes, the agent can analyze your journaled reasons for exiting and compare them against Maximum Favorable Excursion (MFE) data to quantify the 'historical drag' caused by closing positions too early.

    By analyzing trade sequencing and timing, the agent identifies patterns like 'Loss-Response' behavior or declining selectivity after a certain number of trades, helping you set evidence-based daily stops or cooldown rules.

    You receive the full AI agent skill including all analytical modules for edge detection, mistake taxonomy, and counterfactual reporting, along with regular logic updates to ensure compatibility with new trading platforms.

    No, the agent identifies 'Verified Historical Edges' based strictly on your past data; these findings are evidence-based insights intended to inform your strategy experiments rather than financial advice or future guarantees.