Reusable prompts, copy and go.

Act as my day-trading preparation assistant. My focus markets are US500, US100, UK100, GER40, JP225, WTI crude, Gold and Silver (all traded as CFDs, 5m–15m timeframes). Today is [DATE].

1. List today's scheduled high-impact economic events for the US, UK, Eurozone/Germany and Japan with times in UTC, and mark which of my markets each one hits hardest.
2. For each market, tell me which session window is likely to offer the best conditions today and why.
3. Flag anything unusual (holidays, half-days, OPEC/EIA/BOJ specials, big earnings).
4. End with a 5-line "watch out for" summary I can pin to my screen.
If you don't have live calendar access, tell me exactly what to look up and give me the template to fill in.

Why it works: forces a structured brief instead of a vague "what's happening today", and degrades gracefully when Claude lacks live data.

Here is yesterday's data for my 8 focus markets — daily OHLC, today's overnight high/low so far, and daily ATR(14):

[PASTE DATA — one line per market, e.g. "GER40: O 18400 H 18490 L 18370 C 18450 | overnight H 18470 L 18430 | ATR 160"]

Rank these markets from best to worst trading candidate for today. For each, state:
- likely day type (trend / range / undecided) and the evidence,
- how much of the typical daily range is already used,
- the single most important level to watch,
- which playbook fits: opening-range breakout, trend pullback, range fade, or stand aside.
Be willing to say "no trade" for markets that look poor. Give me the top 2 with one sentence of reasoning each.

The closest safe equivalent of "best tickers today": you supply the numbers, Claude supplies the triage.

Build a pre-session level map for [MARKET]. Yesterday: high [H], low [L], close [C]. Prior day POC/VAH/VAL if I give them: [OPTIONAL]. Current price: [P].

1. Calculate classic daily pivots (P, R1, R2, S1, S2).
2. List all levels in one sorted table from highest to lowest, tagging each (pivot / yesterday H-L-C / round number / value area).
3. Mark which levels cluster within 0.15 × ATR of each other — those are my A-grade zones.
4. Tell me which side of the map current price sits on and what that implies for opening bias.

Turns five minutes of platform clicking into one paste; the clustering step surfaces confluence automatically.

This headline just crossed: "[PASTE HEADLINE OR SUMMARY]".

For my markets (US500, US100, UK100, GER40, JP225, WTI, Gold, Silver):
1. Which markets does this actually matter for, and in which direction is the knee-jerk move likely?
2. Is this the type of news that typically REPRICES (trend continues all day) or SPIKES-AND-FADES? What would tell me which one it is within the first 15 minutes?
3. What's the trap scenario — the move that punishes the obvious trade?
Keep it under 200 words. No hedging on every side; commit to the most likely read and label your confidence.

The repricing-vs-spike question is the only news question that matters intraday; the trap scenario counters confirmation bias.

[MARKET] opened at [PRICE]. Yesterday's value area was [VAL]–[VAH] with POC at [POC]. The overnight range was [ON-LOW]–[ON-HIGH]. First 15 minutes: [describe briefly — e.g. "drove straight up 40 points on strong volume" or "chopped both sides of the open"].

Using open-location + early behaviour logic:
1. What day type does this evidence point to — trend day, rotation/balanced day, or trap day?
2. What confirming behaviour should I see in the next 30 minutes if that's right?
3. What single observation would falsify it?
4. Which of my playbooks does this activate, and which is OFF today?

Anchors the day-type call to evidence and gives an explicit falsifier — the discipline half of trading.

I'm considering this trade — argue against it before you agree with it.

Market: [MARKET] · Direction: [long/short] · Entry: [PRICE] · Stop: [PRICE] · Target: [PRICE]
Setup: [describe — e.g. "pullback to 20 EMA in uptrend, RSI 44, bullish engulfing just closed, VWAP below at X"]
Regime evidence: [e.g. "ADX 27 rising, price above VWAP all session"]

1. Grade the setup A/B/C against: regime alignment, location quality, trigger quality, R:R to the nearest REAL level.
2. Give me the three strongest reasons this trade fails.
3. State the one condition under which you'd skip it entirely.
4. If it's still a take: confirm size for [RISK AMOUNT] risk with that stop distance.

Ordering matters — asking for the bear case first prevents the assistant from rubber-stamping your excitement.

I'm in a trade and want a structured management decision, not reassurance.

Position: [long/short] [MARKET] from [ENTRY], stop at [STOP], original target [TARGET].
Now: price is [CURRENT], the trade has been open [TIME], and since entry: [describe what price has done — e.g. "hit +1R, stalled below R1 for 6 candles, volume fading"].

Lay out my three options — hold as planned / take partial and trail / exit now — with the strongest argument FOR each in one sentence. Then recommend one, based only on whether the original setup thesis is still intact. If the thesis is broken, say "exit" plainly.

Frames management as thesis-intact-or-not — the only question that separates managing from meddling.

Do my risk math and sanity-check it.

Account: [SIZE] [CURRENCY] · Max risk per trade: [X]% · Today's realized P&L so far: [AMOUNT] · Trades already taken today: [N]
Proposed trade: [MARKET] CFD, entry [PRICE], stop [PRICE] ([N] points away), instrument point value [VALUE per point per contract/lot].

1. Position size for exactly [X]% risk, shown step by step.
2. If I lose this trade, what's my day's total drawdown in % — and should daily-loss discipline reduce size or stop me trading?
3. Flag anything about this instrument's spread/slippage character (especially Silver or WTI) that should widen my assumed stop cost.

Includes the daily-drawdown context most sizing calculators ignore — the tilt-guard question.

My indicators disagree. Current state on [MARKET] [TIMEFRAME]:

[LIST — e.g. "RSI 71 (overbought), MACD fresh bullish cross above zero, price riding upper Keltner band, ADX 31 rising, VWAP sloping up well below price"]

Resolve this using regime-first logic: which indicator readings are ENTRY-RELEVANT in this regime and which are just describing the trend? What would each school (trend-following vs mean-reversion) do here, and which one has the evidence on its side right now? One-paragraph verdict, then a one-line rule I can reuse for this exact conflict next time.

"Overbought in a trend" kills more accounts than any pattern; this prompt makes the regime hierarchy explicit and extracts a reusable rule.

It's [TIME] UTC. Given normal session behaviour, tell me for each of my markets (US500, US100, UK100, GER40, JP225, WTI, Gold, Silver):
- what phase of its day it's in (pre-open drift, opening drive, midday chop, closing hour, dead),
- whether liquidity/volatility is typically rising or falling from here,
- which ONE strategy type fits this window (breakout / trend pullback / range fade / stand aside).
Present it as a compact table. Then tell me the single best market-window match right now and the one I should absolutely not touch.

Session context is the most under-used filter in day trading — this makes it a 10-second lookup.

Here are the recent swing pivots on [MARKET] [TIMEFRAME], oldest first:

[PASTE — e.g. "L 18280 → H 18410 → L 18342 → H 18560 → L 18488 → H 18590 (current price 18540)"]

1. Give me the TWO most plausible Elliott wave counts for this data (primary + alternate), naming which pivot is which wave.
2. For each count: current position in the structure, what should happen next, and the EXACT price that invalidates it (rule violation).
3. Fibonacci check: do the wave ratios support one count over the other? Show the numbers.
4. One-line trading implication for each count. Do not give me one "confident" count — I want the fork and the invalidation prices.

Demanding two counts plus invalidations mirrors how professionals actually use Elliott — as scenarios, not prophecy.

Compute a Fibonacci confluence map for [MARKET].

Swing 1 (larger): low [PRICE] to high [PRICE]
Swing 2 (recent): low [PRICE] to high [PRICE]
Optional third leg for extensions: [PRICES]

1. Calculate 38.2 / 50 / 61.8 / 78.6% retracements of both swings, and 127.2 / 161.8% extensions of the recent one.
2. List every level in one sorted table.
3. Identify clusters where 2+ levels land within [N] points of each other, and grade them (2 levels = B, 3+ = A).
4. Add current price [PRICE] to the picture: which cluster is the nearest actionable zone above and below?

Exactly the C6 cluster-building process from the combinations page, done in seconds without fat-finger errors.

Check whether these XABCD points form a valid harmonic pattern:

X = [PRICE], A = [PRICE], B = [PRICE], C = [PRICE], D (current/projected) = [PRICE]. Direction: [bullish/bearish].

1. Compute the actual ratios: B as % of XA, C as % of AB, D as % of XA (and of XC), CD as % of BC.
2. Which named pattern fits best (Gartley / Bat / Butterfly / Crab / Cypher / none), and how far do my ratios deviate from the ideal?
3. If valid: give the PRZ zone, the invalidation price, and targets at 38.2% and 61.8% of CD.
4. If nothing fits cleanly, say so plainly — do not force a pattern.

The "do not force a pattern" instruction matters: harmonic tools' biggest failure mode is wishful ratio-rounding.

I'm looking at [MARKET] [TIMEFRAME] and price is doing this:

[DESCRIBE — e.g. "rallied hard for 2 hours, now 5 pushes to marginal new highs, each smaller, converging trendlines, RSI making lower highs each push" — or attach a screenshot]

1. What chart pattern(s) match this description? Rank up to 3 candidates by fit.
2. For the best fit: what it typically resolves into, the published reliability caveats, the trigger that confirms it, and the measured-move target.
3. Elliott translation: where in a wave structure does this pattern usually sit?
4. The failure mode: what does it look like when this pattern traps traders instead?

Works with screenshots too (paste the image) — Claude reads charts well when the pattern is visible.

Give me the current correlation context that matters for [MARKET I'M TRADING].

Right now: [PASTE the related quotes you can see — e.g. "USD/JPY 158.2 up 0.6%, DXY 105.1 up 0.3%, US 10y yield 4.42% up 5bp, Gold 2365 down 0.4%"]

1. Which of these relationships historically drives my market, in which direction, and how reliably (strong/loose)?
2. Given the numbers I pasted, are the related markets CONFIRMING or FIGHTING the move in my market right now?
3. What divergence between them would be an early warning that my market's move is about to fail?
Keep it practical — no academic correlation lecture.

Especially valuable for JP225 (yen), UK100 (commodities/GBP) and the metals (yields/DXY) where the driver is off-screen.

Autopsy this trade with total honesty. Separate PROCESS from OUTCOME — a loss can be a good trade and a win can be a bad one.

[PASTE journal entry: market, direction, entry/stop/target, what the setup was, what happened, what I felt, final P&L]

1. Was the setup valid per a regime→location→timing checklist? Which step was weakest?
2. Was the loss/win caused by: bad analysis, bad execution, bad management, or plain variance? Pick ONE primary cause.
3. What would the ideal version of this exact trade have looked like?
4. Give me one specific, checkable rule to add to my playbook so this lesson sticks. Not advice — a rule I can grade myself against.

The one-primary-cause constraint prevents the "bit of everything" non-conclusion that makes journaling useless.

Analyze my trade log like a QA engineer analyzing test results. Here it is (CSV or table):

[PASTE — columns like: date, market, session-time, setup-tag, direction, R-result, notes]

1. Stats per setup-tag AND per market: count, win rate, average R, expectancy. Flag anything with n < 10 as "insufficient sample" rather than drawing conclusions.
2. Time analysis: are there session windows or weekdays where I reliably lose?
3. Find my single biggest leak (the setup/market/time combination costing the most R) and my strongest edge.
4. Give me a one-line "stop doing / keep doing / start testing" verdict, each backed by a number from my own data.

The n<10 guard keeps Claude from over-fitting stories to noise — the same discipline you'd demand of any eval.

Design a rigorous backtest for this trading rule before I trust it with money:

Rule: [DESCRIBE — e.g. "on GER40 5m, buy the first pullback to the 20 EMA after ADX crosses above 25, stop 1×ATR under the swing low, target 2R"]

1. Specify it as unambiguous pseudocode — every threshold, every timing detail (signal bar vs entry bar).
2. List the specific look-ahead and survivorship traps this rule is prone to, and how to avoid each.
3. Define the test protocol: data needed, in-sample/out-of-sample split, regime segmentation (trend vs range months), realistic CFD costs, and minimum sample size before any conclusion.
4. Tell me what result would DISPROVE the rule — the falsification criterion comes before the test.
5. Sketch the Python structure (pandas) — functions and flow, not full code yet.

Falsification-first is what separates evaluation from confirmation — same methodology as model evals.

Synthesize my trading week from these notes and numbers:

[PASTE — daily notes, P&L by day, trade count, anything you logged]

1. The week in 3 sentences — what actually happened, no spin.
2. Pattern detection: what repeated across the week (good and bad)? Include behavioural patterns (revenge trades, hesitation, oversizing after wins), not just setups.
3. Verdict lists: 3 KEEPS (working — protect them), 3 STOPS (leaks — kill them), 1 EXPERIMENT for next week with a success metric.
4. One question I should be able to answer by next Friday that I can't answer today.

The final question forces the week to produce a research agenda, not just a mood summary.

Quiz me to sharpen my trading decisions. Topic: [PICK ONE — e.g. "Elliott wave counting", "reading VWAP day types", "harmonic PRZ trades", "managing breakout failures"].

Rules:
1. Give me realistic scenarios one at a time: describe price action, indicator readings and context on one of my markets (US500, US100, UK100, GER40, JP225, WTI, Gold, Silver), then ask what I do — including the option that the answer is "nothing".
2. Wait for my answer before continuing. Grade it, explain what the best-practice answer was and why, then escalate difficulty.
3. Mix in at least 2 trap scenarios where the correct answer is to NOT trade.
4. After 10 scenarios, summarize my weak spots and what to study.

Spaced scenario practice with traps included — the cheapest deliberate practice a trader can get.