Confluence: level, candle and trend

Module 2 ended with an honest table: every candle pattern on its own is weak. Module 3 gave you levels, which on their own are also weak; price crosses most levels eventually. Module 4 gave you trend, which on its own tells you direction but not timing. Confluence is the practice of requiring all three to agree before you call something a setup. It is not a trick; it is the filtering that removes most of the noise from each ingredient.

The three questions

Every setup in this course is a "yes" to all three:

1. Trend: does the context timeframe have a defined direction, and is the trade in that direction (or, in a range, at an edge)?
2. Level: is price at a zone with a nameable trapped group, drawn before price got there?
3. Candle: has a candle formed at that zone, on the trading timeframe, that shows the trapped group being squeezed (engulfing, pin bar, failed break) or a decision pending (inside bar)?

Two out of three is a watchlist entry, not a trade. One out of three is nothing.

Why three independent reasons

The word independent matters. A hammer forming at a swing low in an uptrend is three separate observations that happen to point the same way:

• The trend says sellers on pullbacks have been trapped repeatedly.
• The level says there is a specific price where the last group of sellers was trapped, and buyers defended it.
• The candle says that, in the current period, sellers pushed through and were reversed again.

Each piece raises the probability a little. Together, they describe a place where a large trapped group (the pullback sellers, now including the ones in the hammer's wick) is about to be squeezed by a larger group that has been winning all along (the trend buyers). That is a description of order flow, not a pattern name.

Contrast with fake confluence: a hammer, a 50-period moving average, a Fibonacci retracement and a trendline all at the same price. These are four ways of drawing lines near recent price, and they are not independent; three of them are derived from the same swing points. Counting them as four reasons is counting the same reason four times.

Key idea: Confluence means three independent observations, from trend, level and candle, that describe the same trapped group being squeezed. Stacking indicators derived from the same data is not confluence.

Worked example

Daily chart, uptrend by the swing definition (N = 5), last higher low at 61.80 after a swing high at 66.40. Average daily range 1.30.

Check the three questions:

1. Trend: daily uptrend, HH/HL intact so far. Weekly (context) also HH/HL. Yes.
2. Level: the prior higher low at 61.80, drawn as a zone from roughly 61.50 to 62.00, existed before day 5. Yes.
3. Candle: day 5 has body 0.20, lower wick 1.10 (5.5× the body, 79% of the range), upper wick 0.10. A hammer by the Module 2 thresholds, and its wick pierced the zone and closed above it. Yes.

  66.40 ---+---+---------------------------------------
           |   |
           +---+   |
             |   +---+
                 |###|   |
                 +---+ +---+
                   |   |###|   |
                       +---+ +---+
                         |   |###|
                             +---+     |
  61.80 zone /////////////////|///////+---+////////////
                               |      +---+
                                        |
                                        |     <- hammer wick to 60.90

The trapped group: sellers who sold days 2-5 expecting the pullback to become a breakdown, especially those who sold below 61.80 on day 5 and are now, at the close, underwater. The squeeze: the trend buyers who have defended every higher low, plus the day 5 buyers who defended this one.

This is a setup. Lessons 2-4 cover what to do with it: the entry, the stop and the target, and the arithmetic that decides whether it is worth taking.

The order matters

Draw the level before price arrives. Define the trend before looking for a candle. Then wait for the candle. Traders who see a big candle first and then go looking for a level to justify it will always find one, because there is always some line nearby. That is hindsight dressed up as analysis, and Module 6 gives it a name.

The honest workflow is: levels and trend at the start of the session or week, when nothing is happening; candles during the session, when something is. If you find yourself drawing a new level during a fast move, stop.

What confluence does not do

It does not make the trade a winner. The hammer at the level in the uptrend still fails a meaningful fraction of the time. What it does is two things: it reduces the number of trades you take to the ones with the best-described trapped group, and it gives you an exact place where the story is wrong (below the hammer's low), so that the losses are bounded. The rest of the module is about turning a bounded loss and a plausible target into a number you can compare across setups.

Try it: Take any chart you have already marked with levels and trend from Modules 3 and 4. Look for the last five candles that passed a Module 2 definition (engulfing, pin bar, inside bar). For each, fill in the three-question table. How many were full setups? Most traders find one or two out of five, which is the filtering working.

Recap

• A setup requires three independent yeses: trend on the context timeframe, a pre-drawn level with a trapped group, and a candle at that level showing the squeeze.
• Two of three is a watchlist item; one of three is nothing.
• Indicators derived from the same swing points are not independent evidence; counting them is counting one reason several times.
• Levels and trend are drawn before the candle arrives, never after.
• Confluence does not make trades win; it filters to the best-described trapped groups and gives a precise point where the story fails.