An RSI pattern must be logical with price

Core idea

An RSI pattern (a cycle of expansion → peak → contraction relative to the two moving averages) can only be counted as a real pattern when it corresponds to a real wave on the price chart. If RSI looks pretty but price doesn't follow through → that's not a pattern, just a noise zone, and it must never be used as grounds for a decision.

Details

The reasoning

• An uptrend in RSI = buying pressure increasing continuously.
• If buying pressure keeps increasing, price is forced to rise → there must be a corresponding upwave.
• So: if RSI draws an uptrend but price fails to produce a matching upwave, the premise was wrong — there is no dominance by the buy side there.

The same logic applies on the downside.

The procedure for checking a pattern (2 steps)

1. Turn off the price chart, look only at RSI: is there enough slope and spread between the two moving averages? (If not enough → discard immediately.)
2. Turn the price back on: does that stretch have a real wave moving in the same direction?
   • Yes → the pattern is valid, usable for decisions.
   • No → it is not a pattern, no matter how good RSI looks.

An example the teacher dissects (H8 timeframe)

A student asked whether H8's downward pattern looked good. Looking at RSI: there was slope, there was spread → at first glance it looked like a pattern. But turning price back on showed: price bounced up (from about 25–26k to nearly 28k) then folded back down to exactly where it started.

Read in terms of force: the upward leg did have real buying pressure (otherwise what pushed price up near 28k), but the selling pressure that came in afterward completely cancelled out that buying pressure. The two sides neutralized each other → neither side gained the upper hand → price went into a range. → Conclusion: this is not an RSI pattern, because it doesn't correspond to a prior established price trend. It cannot be used to make a decision.

If there had really been selling pressure in that stretch, we would have been able to sell and made a profit — in reality nobody could sell there, and price just wandered into a noise zone.

Other out-of-phase cases

• RSI falling while price goes sideways / holds support: buying pressure is resting, selling pressure hasn't come in yet. Verification: RSI doesn't break below 40 and support holds. This is still a valid pattern, but it must be read correctly for what it is (a rest, not a reversal).
• RSI correcting while price not only doesn't fall but pulls the higher timeframe up: this is an abnormal wave. In the correction leg of a strong downtrend, price should only go sideways (selling pressure resting temporarily), not pull the whole higher timeframe up. An abnormal wave is a signal to re-examine the entire set of assumptions.
• Out-of-phase patterns between timeframes: whichever timeframe has a trending pattern still decides the market — but it will not be strong if it's out of phase with the higher timeframe. Prior low not lower, prior high not higher, plus being out of phase → you can tell from the start that this wave is weak and prone to falling into a range.
• A timeframe with no pattern = a timeframe with no opinion. If H4 is sideways, H4 contributes nothing to the decision, no matter how pretty the pattern on 2D/3D is.

Putting it into practice

1. Scan in order from higher timeframe down to lower timeframe (1M → 1W → 1D → H16 → H12 → H8 → H4 → H1…), marking which timeframes currently have a trending pattern.
2. For each pattern: run the 2-step procedure above. Keep only the patterns that match up with a price wave.
3. Make the decision based on the remaining timeframes, prioritizing the one with the clearest pattern (best slope + best spread).
4. Time your entry around point 1 / point 2 / point 3 (the three times the two moving averages contract) of the dominant timeframe — see rsi trend and the two moving averages.
5. Enter by zone, not by point, so there's still room to calmly watch what the market suggests next.

What the teacher stresses / common traps

The mechanical trap. Reading RSI as "cross up = buy, cross down = sell" gets you killed. The small, fidgety crosses of RSI against the two moving averages while everything is loitering around the 40–60 zone are worth nothing. You have to look at the whole shape of the movement.

The "I can explain it either way" trap. In a range, buying pressure and selling pressure are balanced → every trading system is noisy there, because every system is built on a buy/sell imbalance. At that point you can argue for up just as easily as for down — and that's exactly when it's easiest to rationalize the urge to enter a trade. The teacher's closing line: it's not that the market is hard to analyze — it's that there's nothing there to analyze. Force an analysis anyway and the one who dies is you.

The "getting ready to rise / getting ready to fall" trap. The only question you're allowed to ask is: is it currently rising or currently falling. The current state is a fact; "getting ready" is an expectation.

The trap of guessing weak/strong momentum in advance. At the moment momentum forms, all you're allowed to know is "there is momentum." Only once there's real evidence (e.g. the H12 timeframe gets pulled up into an uptrend state) are you allowed to conclude the downward momentum ahead has weakened and the chance of it failing is high — only then do you enter against it.

The trap of missing an entry and jumping to a higher timeframe. If you missed H4's entry point, you are not allowed to use "entering on the higher timeframe instead" as an excuse, because the stop would have to sit at the higher timeframe's high/low → too much risk. The two correct choices: (a) wait for H4 to pull back and then enter on H4, or (b) drop to a lower timeframe (M15) to find an entry, but still manage the trade by the parent timeframe — once the parent timeframe starts rolling back over, you exit.

The trap of a range inside a range. There are stretches where several timeframes are sideways at once and a higher timeframe is the one deciding — at that point the timeframes below it have almost no information left. An example the teacher gives: 11/2018 → 4–5/2019, price spun inside one band for nearly half a year. In that stretch, the faster you trade the faster you die; one trade a whole week is completely normal.

Three entry states: buy, sell, and DO NOTHING. Doing nothing is also a decision — it preserves what you already have. "Not losing money is already the same as having made money." Everyone sees the obvious opportunities, everyone avoids the obvious risks; the hard part of this profession is handling the 50-50 opportunities.

On timing an exit: don't close out inside a 50-50 zone. Lay the facts out on a table: once the table is tilted in your favor and starts shifting toward the opposite side, and there's one more confirming factor (e.g. price breaks up through a resistance zone that had been holding) → that's when you close.

You still have to stay with the market even during stretches with no open trade — so you don't get lulled to sleep and miss a wave — but stay with it in a state of alertness, without getting pulled into opportunities that aren't clear.

Related lessons