Building a Repeatable TA Process

A repeatable technical analysis process is a fixed sequence of questions you answer, in the same order, every time you consider a trade — ending in either a fully specified plan (entry, invalidation, target, size) or a documented "no trade." The individual tools from this path — structure, moving averages, RSI, volume, patterns — only produce consistent results when they're assembled into a checklist that removes in-the-moment improvisation, because improvisation is where discipline actually fails.

Why a Process Beats a Toolbox

By this point in the path you know candlesticks, support and resistance, trendlines, moving averages, RSI, MACD, volume, patterns, fibs, Bollinger Bands, and multi-timeframe analysis. That's the problem. Eleven tools consulted ad hoc will happily justify any trade you already want to take: ignore the bearish MACD, cite the bullish fib, done. This is confirmation shopping, and more tools make it worse, not better.

A process fixes this in three ways:

1. Fixed order. You answer the questions in sequence, so the exciting entry signal comes last, after context has already had the chance to veto it.
2. Fixed toolset. You decide in advance which tools count, so contradicting evidence can't be silently dropped.
3. Written output. Every run produces a record — either a plan or a pass — which is the raw material for improvement later.

The realistic goal isn't predicting markets. It's making your decisions consistent enough that you can find out what actually works, and thin enough in number that only your best setups get your money.

Layer One: Context Before Anything

Start every session top-down, before looking for trades — bias timeframe first, then setup timeframe, as covered in the multi-timeframe guide.

Question 1: What is the higher-timeframe structure? Uptrend (higher highs and higher lows), downtrend, or range. Write it down with the level that would invalidate it: "BTC daily uptrend, intact above $92,000."

Question 2: Where are the levels that matter? Mark the handful of zones — major support/resistance, prior consolidations, the current range's edges. Five levels you believe in beat twenty you don't.

Question 3: Where is price right now relative to them? Three possible answers: at a level of interest, approaching one, or nowhere. "Nowhere" ends the run — no trade, and that's a successful output, not a failure of the process.

Question 4: What's the volatility and event context? Compressed and coiling, or extended after a big move? Any scheduled events — Fed meetings, CPI releases, major token unlocks — inside your intended holding window? A perfect setup an hour before a macro release is a worse setup.

Context is deliberately indicator-free. Structure and location first; indicators never overrule the map.

Layer Two: Evidence at the Level

Only when price is at or approaching a marked level do you gather evidence. Choose two or three tools from this path — not all of them — and use the same ones every time. A reasonable core: one trend tool (moving averages), one momentum tool (RSI or MACD, not both — they're derived from the same prices and mostly double-count), and volume.

For each, record what it says for or against the trade idea:

• Trend tool: Is price above or below the daily 50 EMA? Is the MA rising, falling, or flat?
• Momentum: Is RSI diverging against the move into the level? Coming off an extreme, or mid-range and mute?
• Volume: Did the approach to the level come on expanding volume (conviction) or fading volume (exhaustion of the move into it)?
• Optional structure evidence: a pattern forming at the level, a fib confluence, a Bollinger squeeze — extras that strengthen a case, never substitutes for location.

Then apply a threshold you fixed in advance, for example: take the trade only if the trend tool agrees and at least one of the other two does, with none screaming against it. The exact rule matters less than the fact that it was set before you saw this particular chart.

Layer Three: The Trade Specification

A trade idea becomes a trade plan only when four numbers exist on paper before entry:

Entry trigger. Not "around $2,300" but a defined event: "1-hour close back above $2,320 after tagging the zone." Triggers force the market to confirm before you commit.

Invalidation. The price at which the idea is objectively wrong — beyond the swing point or zone that justified the trade, not at a round number of dollars you're comfortable losing. If invalidation is hit, you exit. No renegotiation, no zooming in to find hope.

Target(s). The next meaningful opposing level, a measured move, or a trailing rule — decided now, because after entry you'll be the least objective person watching this chart.

Size. Derived from invalidation distance and your fixed risk per trade (the 1-2% logic from the trading fundamentals path). Entry $2,320, stop $2,270, $10,000 account risking 1% ($100): size is $100 ÷ $50 of risk per unit ≈ 2 ETH-equivalent exposure. Size is an output of the math, never an input from your mood.

Before submitting the order, one last check: is the reward-to-risk acceptable (many traders require at least 2:1), and does this trade conflict with existing positions?

The Checklist and the Review Loop

Condensed into the artifact you actually run — ten lines, five minutes:

1. Higher-timeframe trend and its invalidation level — written.
2. Key zones marked; stale levels deleted.
3. Price at/near a zone? If nowhere → stop, log "no trade."
4. Event risk inside the holding window checked.
5. Trend tool: agrees / disagrees.
6. Momentum tool: agrees / disagrees / diverging.
7. Volume: confirming / fading.
8. Evidence threshold met? If not → stop, log it.
9. Entry trigger, invalidation, target, size — all four written as numbers.
10. Reward-to-risk acceptable and consistent with open positions → place order.

The loop closes with review, which is what makes the process improvable rather than just tidy. Log every run — including the passes — with a chart screenshot and the checklist answers. Then, on a fixed cadence (weekly or monthly, per the trading-plan habit from the fundamentals path), ask the questions only a log can answer: Are my losses actually stopping at invalidation, or am I overriding? Which evidence item, when present, coincides with my winners? Do my "no trade" days outperform my bored days? Expect to discover that one of your chosen indicators adds nothing — cutting a tool is the most common and most profitable revision.

Two rules for changing the process: change it only at review time (never mid-trade, never right after a loss), and change one thing at a time so you can tell what the change did. A process you quietly edit under pressure isn't a process; it's a costume that discretion wears.

Key Takeaways

• A TA process is a fixed question sequence — context, evidence, specification — run identically every time, with "no trade" as a first-class output.
• Structure and location come first and can veto everything; a small fixed set of indicators (one trend, one momentum, plus volume) provides evidence at levels, never trade ideas on its own.
• No entry without four written numbers: trigger, invalidation, target, and size derived from fixed risk per trade.
• Log every run, including passes, and review on a fixed cadence — the log is what turns a checklist into a system that improves.
• Edit the process only at review time and one change at a time; mid-trade edits are discretion in disguise.