Trade history review

An AI Trading Coach That Works From Your Actual Trades

TMM's coach reviews the trades already in your journal and helps you understand what shaped your results. Ask about entries, exits, position sizing, risk, timing, setups, or recurring patterns and get answers grounded in your own trading record.

It is built for post-trade review. It does not predict prices, generate trading signals, tell you what to buy, or place orders.

Your trade history. Verified metrics. Read-only connections.

Last updated: August 2026.

How Does AI Help You Review Trades?

A journal-connected coach reviews a trader's recorded activity, explains performance patterns, and helps the trader ask better questions about their process. TMM combines AI conversation with structured journal data, so its answers can refer to actual trades and calculated metrics instead of generic trading advice.

Turn Your Trading History Into a Useful Review

The coach connects conversation to the data and analytics already inside TMM.

Ask

Use plain language to explore your results: which setups work, when losses cluster, how position size changes outcomes, or what deserves a closer review.

Analyze

Compare performance by symbol, setup, time of day, risk, entry reason, exit reason, drawdown, and other metrics calculated from your journal.

Improve

Turn findings into specific review questions, dashboard views, and measurable rules you can evaluate against future results.

How the AI Trading Coach Works

  1. Bring in your trading history

    Connect a supported account using a read-only key. TMM checks the permissions before accepting the key, then imports the trade data available from that connection.

  2. Ask about your performance

    Ask a question in plain language or start from a metric, setup, period, or concern you want to examine.

  3. Review the evidence

    The coach uses your journal data and TMM metrics to explain patterns, build relevant dashboard views, and help you compare possible changes.

What the Coach Can Analyze

The coach can work with the same verified performance data available in your TMM journal and analytics dashboards.

Decisions and execution

  • Entries and exits by the reasons or tags you recorded
  • Results by setup and symbol
  • Favorable movement after entry
  • Partial fills and the executions that make up a trade

Risk and position management

  • Position size and risk per trade
  • Risk-to-reward data recorded in TMM
  • Drawdown and recovery factor
  • Fees and other trading costs recorded in the journal

Timing and consistency

  • Profit and win rate by hour or trading session
  • Trade frequency over time
  • Performance across selected periods
  • Recurring patterns you choose to track

Analysis and dashboards

  • Plain-language explanations of supported metrics
  • Dashboard widget suggestions based on your question
  • What-if analysis using your historical results
  • Context you have asked the coach to remember about your process and goals

Questions You Can Ask About Your Trades

Start with a question you would normally investigate manually.

  • Which setups have produced my most consistent results?
  • How does my performance change when I increase position size?
  • At what times of day do my losses cluster?
  • Which exit reasons are associated with giving back the most profit?
  • How did my drawdown change during the last selected period?
  • Build a dashboard view for setup performance and risk.
  • What would my historical equity curve look like if selected losses were capped at a lower level?

The quality of an answer depends on the data and notes available in your journal. The coach should distinguish calculated facts from interpretation and should not invent missing context.

A real session

Watch the coach read an account

Four questions, four answers, straight from the TMM coach. Scroll to replay the session.

TMM assistant
Illustrative data

Question 1

The honest read

Can you take an honest look at my trading and tell me how I'm really doing? Don't sugarcoat it — I want to know if I actually have an edge or if I've just been getting lucky.

TMM assistant
Thinking · 3s

Alright, you asked for honest — here it is, no sugar.

The short answer: You have a real edge, but it's narrower than your P&L suggests, and you're currently burning a lot of it on setups you shouldn't be trading.


The headline numbers (all-time, 302 trades):

  • Total PnL: +$29,752 · Avg/trade: ~$98.50
  • Win rate: 50.7% (153 wins / 149 losses) — basically a coin flip
  • Payoff ratio: 1.31 — your average winner is only 31% bigger than your average loser
  • Profit factor: 1.35 — you make $1.35 for every $1 you lose
  • Max drawdown: −24.71% on May 25, with a second spike to −14.7% the next day and a recent −9.9% on July 8

A profit factor of 1.35 and a 50% win rate with 1.31 payoff is statistically positive, but it's fragile. A handful of trades carry the whole thing — your single best winner

ShortDOGEUSDT1.41x
+$6,682 (+10.68%)
14d

DOGEUSDT was coiling into a squeeze; entered on the breakout candle with volume expansion.

Exit: Trailing stop triggered on DOGEUSDT after a strong run.

Trend continuationRange bounceFunding rate skewStop-loss hitNews-driven exitEnd-of-session close

is 22% of your all-time PnL, and your worst loser

LongBTCUSDT1.14x
−$5,628 (−13.34%)
9d

Copied a setup similar to a prior winner on BTCUSDT, same structure and volume profile.

Exit: Stop moved to breakeven then hit on BTCUSDT.

Liquidity sweepVolume confirmationScalp on volatilityReversal signal on lower timeframeNews-driven exit

is 19%. Those two alone account for 41% of your net outcome. That's not a diversified edge — that's concentrated exposure.


Here's what your edge looks like broken down by entry reason:

Profit by entry reasons
  • Moving average cross23673.44042392(36)
  • Funding rate skew16848.6589422(30)
  • Correlated asset lead15606.63781573(34)
  • Range bounce14290.00682968(29)

Your PnL isn't spread evenly across your setups — it's radically polarized:

Your real moneymakers (the setups that actually have edge):

SetupTradesTotal PnLAvg/Trade
Moving average cross36+$23,673$657
Funding rate skew30+$16,849$562
Correlated asset lead34+$15,607$459
Range bounce29+$14,290$493
Reversal signal30+$11,321$377

These 5 setups = +$81,740 on 159 trades.

Your money drainers (setups losing you thousands):

SetupTradesTotal PnLAvg/Trade
Volume confirmation33−$12,126−$367
Fibonacci retracement30−$9,193−$306
Scalp on volatility26−$9,177−$353
Liquidity sweep36−$8,962−$249
Session open momentum31−$8,284−$267
Squeeze breakout32−$7,756−$242

These 6 setups = −$55,498 on 188 trades.

The math is brutal: Your winning setups average +$514/trade; your losing ones average −$295/trade. The losing setups aren't just less profitable — they're consistently negative and you trade more of them. Your "retest of breakout level" had 41 trades (most of any setup) and still lost −$2,398.


The verdict:

You clearly have an edge in trend/momentum setups (moving average crosses, range bounces, reversal signals, funding rate plays, correlated movers). Those work — the data is unambiguous.

But you're also trading a laundry list of setups — volume confirmation, Fibonacci scalps, liquidity sweeps, session opens, squeeze breakouts — that are reliably losing money. If you cut just those six losers, your net PnL would jump from ~$29.8k to ~$85k.

So to answer your question directly: yes, you have a real edge — but you're currently paying for it by trading 5+ setups you have no business being in. That's not luck; that's self-sabotage. The question is whether you can stop doing the things that don't work.

Question 2

The one lever

That's a lot to take in. If I could only fix one thing this month, what's the single biggest lever — and what would it realistically do to my bottom line?

TMM assistant
Thinking · 1s

Here's your single biggest lever, quantified.


The one thing to cut this month: Squeeze breakout.

In the last 30 days, you took just 9 squeeze breakout trades and lost −$6,582 — that's −$731 per trade, the worst hit rate of any setup you run. Cutting those 9 trades alone would boost your monthly PnL from +$11,495 to +$18,077 — a 57% improvement with zero other changes.

But here's the real punchline: squeeze breakout is just the worst of a trio. Add volume confirmation (−$6,332 on 14 trades) and scalp on volatility (−$6,108 on 14 trades), and those three setups alone burned −$19,022 in 30 days — that's 165% of your entire month's profit. Your winning setups had to work twice as hard just to dig you out of that hole.

Realistic impact if you cut just squeeze breakout for the next 30 days: assuming your other setups perform similarly, you'd net roughly $18k instead of $11.5k. If you cut all three losers, you'd be looking at ~$30k/month — more than double your current result.

Start with squeeze breakout. It's only 9 trades a month, it's the worst performer, and it's the easiest to consciously skip.

Question 3

When you bleed

Are there certain times of day or days of the week when I trade worse? I feel like I make dumb trades at certain points and I want to know if the data backs that up.

TMM assistant
Thinking · 2s

Your gut is right — the data screams it. Let me pull some specific trades to show you what's going on in your worst stretches.

Profit by time of day
  • 1418999.852174(17)
  • 2112453.20063866(11)
  • 1111791.38796755(12)
  • 23-8031.15956899(13)
Win percentage by time of day
  • 1470.5882(17)
  • 1964.2857(14)
  • 2163.6364(11)
  • 1560(5)

The most dangerous hours, ranked:

HourTradesPnLWin RateVerdict
Noon (12 PM)9−$6,14422%🚩 Your worst hour by a mile
11 PM (23:00)13−$8,03139%🚩 Bleeding out at day's end
Midnight (0:00)10−$2,45320%🚩 1-in-5 win rate
5 AM23−$4,49044%Most-traded losing hour
1 PM13−$3,86746%Late lunch wreckage

Compare that to your peak — 2 PM (14:00) where you're +$19,000 with a 71% win rate over 17 trades, or 9 PM (21:00) at +$12,453 with 64%. The swing between your best and worst hours is $27k — that's your entire lifetime PnL.

Weekdays: Wednesday is the anchor (−$5,145 on 46 trades) and Thursday is a disaster zone at 38% win rate — the worst single day of the week. Friday (+$12k), Saturday (+$10.6k), and Sunday (+$11.8k) are your heavy lifters, though their win rates are only 45–53% — meaning large winners, not consistency, carry those days.


The pattern: You crash at noon (22% WR), midnight (20% WR), and late night 11 PM (39% WR). These aren't normal variance — those are statistical black holes. Meanwhile, 2 PM, 9 PM, and 6 PM are your sweet spots.

Want me to dig into a few trades from noon or midnight to see what exactly goes wrong there — wrong setups, revenge trades, holding too long?

Question 4

The 30-day plan

Okay. Give me a simple, realistic 30-day plan to actually improve — nothing I'll need a spreadsheet to follow.

TMM assistant
Thinking · 2s

Here's your 30-day plan — three rules, no spreadsheet.


Rule 1: No trades at noon, midnight, or 11 PM.

Your win rate at those hours is 20–39% and you've lost −$16,628 combined. That's more than half your all-time PnB, gone in three time slots. Your sweet spots are 2 PM, 6 PM, and 9 PM — trade those instead. When noon or midnight rolls around, close the charts. Full stop.

Rule 2: If you take a loss, walk away for 2 hours.

Trade

ShortSOLUSDT0.45x
−$518 (−2.89%)
1d

Copied a setup similar to a prior winner on SOLUSDT, same structure and volume profile.

Exit: Booked gains on SOLUSDT at the prior high.

Reversal signalLiquidity sweepBreakeven stop hitEnd-of-session closeRevenge trade

tells the story: a SOLUSDT short tagged revenge trade — you lost −$518 on a setup you yourself described as "copied a similar prior winner." The order timeline shows you scaled in 7 times over 15 hours, then dumped the whole position in one shot. That's not a trade — that's a tilt. Your journal shows you've been fighting this pattern for months. Make it automatic: loss → 2-hour break. No exceptions.

Rule 3: Only trade your top 5 setups. Skip the other 8.

Your winners:

  • Moving average cross
  • Funding rate skew
  • Correlated asset lead
  • Range bounce
  • Reversal signal

Everything else — squeeze breakout, volume confirmation, scalp on volatility, Fibonacci retracement, liquidity sweep, session open momentum — is burning you. In the last 30 days those 6 losers cost you −$19,022. If a trade doesn't fit one of your 5 winning patterns, don't take it.


What this realistically does to your bottom line:

Based on your last 30 days (117 trades, +$11,495):

  • Rule 1 removes ~15–20 trades in your dead hours. Most of those are losers, so your win rate climbs into the high 50s.
  • Rule 2 prevents the revenge spiral that turns a small loss into a −$500+ disaster.
  • Rule 3 cuts your trade count by about half but keeps only the +$514/trade setups instead of the −$295/trade ones.

Projected result: ~$18k–$22k next month instead of $11.5k — and you'll trade less, not more.

Want me to save these three rules to your memory so I can hold you accountable next time?

Captured 2026-07-11 from a real coach session. The figures are one account's, not a typical result.

Why a Journal-Connected Coach Is Different

A general chatbot only knows the trades and context you paste into the conversation. That can be useful, but the answer depends on whether the summary is complete and whether the numbers were calculated correctly.

TMM's coach works with structured trade records, journal notes, supported metrics, and saved context from your account. It can refer to the same data you use in your dashboards and explain how supported metrics are calculated. This makes the review traceable to a real record rather than a reconstructed prompt.

If you prefer another assistant, the TMM MCP server can provide supported AI clients with read-only access to the same trading history. The built-in coach adds TMM's dashboard tools and saved coaching context.

Your Data Stays Connected to the Review, Not to Trade Execution

Uses
Your imported trading history, calculated performance metrics, journal notes, tags, and the coaching context you choose to save.
Helps with
Post-trade review, performance questions, metric explanations, dashboard analysis, and historical what-if comparisons.
Never does
Predict prices, sell signals, choose assets for you, place orders, or withdraw funds.

TMM accepts only read-only API keys. A key cannot be added until it passes the permission check. If a key can place orders or withdraw funds, TMM rejects it. The service can only retrieve trading data.

Automatic imports currently cover TMM's supported crypto exchanges, including spot and futures data where available. Funding, fees, leverage, liquidations, and partial fills are handled when the connected source provides those fields. Stock, forex, options, and traditional broker connections are not currently available.

Who the AI Trading Coach Is For

The coach is for traders who want to review decisions with evidence instead of relying on memory. It is most useful when you have enough recorded trades to compare setups, risk, timing, and execution—and when you are willing to turn a pattern into a rule you can measure.

It is not for anyone looking for guaranteed returns, live trade calls, autonomous execution, or a prediction engine.

Key Takeaways

  • TMM's AI Trading Coach reviews the trades in your journal; it does not predict the market.
  • Answers can use structured trade records, notes, tags, and verified performance metrics from your account.
  • The coach can explain patterns, suggest dashboard views, and run supported historical what-if comparisons.
  • The coaching workflow focuses on decisions, risk, timing, and consistency rather than on predicting a particular market.
  • Connections are read-only and must pass a permission check before TMM accepts them.

FAQ

Review your trades with AI

Bring your trading history into one review and ask better questions about your performance.

What does an AI trading coach do?
It reviews recorded trades, explains performance patterns, and helps a trader examine decisions, risk, timing, and consistency. TMM's coach uses the data in the trader's journal rather than relying only on a manually written prompt.
Is this an AI trading journal?
The coach is built into TMM's trading journal. The journal stores and calculates the trading record; the coach lets you question that record in plain language, explain supported metrics, and turn findings into dashboard views or further review.
Does the coach provide trading signals or financial advice?
No. It reviews trades you have already made. It does not predict prices, generate buy or sell signals, choose assets, or promise future results. Its output is for performance review and education, not financial advice.
What data does the coach use?
It can use the imported trade data, calculated metrics, journal notes, tags, and saved coaching context available in your TMM account. The exact answer depends on which fields and history are present.
Can the coach analyze different trading styles?
Yes. The coach can compare the setups, symbols, periods, risk levels, and other supported fields recorded in your journal. The available analysis depends on the data and tags in your account, so it should not assume a strategy or trading style that you have not recorded.
How is this different from pasting trades into ChatGPT?
A pasted prompt contains only the information you provide in that conversation. TMM's coach works with the structured trading history and supported metrics already stored in your journal, so answers can be tied to the same record you inspect in the product.
Can the AI Trading Coach place trades or withdraw funds?
No. TMM accepts only read-only API keys, and a key cannot be added until it passes the permission check. If a key can place orders or withdraw funds, TMM rejects it.
How much trading history do I need?
You can ask questions as soon as trades are available, but comparisons become more useful when the selected period contains enough relevant trades. The coach should state when the available sample is too small for a reliable pattern.