The lesson on YouTube
Watch. Then work through the idea.
Watch the English lesson first, then use the worked example and quiz below to check the idea. The article is a companion, not a transcript.
Loading connects to YouTube and may transmit device data or use storage. Privacy details. Alternatively, watch on YouTube ↗.
What is actually happening?
An average summarizes past prices. It changes when new observations replace old ones. Its position or slope can classify your setup context, but neither guarantees the next move.
Guided example · 3-bar SMA arithmetic · synthetic closes
See an average update
Use three bars to make the arithmetic visible. The article’s separate setup filter uses 20 bars.
Restart & example assumptions
Authored synthetic observations, shown in sequence. Every comparison uses only the revealed data. This is a worked learning example, not live prices, broker execution or a profitability backtest. Follow the article’s exact definitions; measurements and pattern names alone do not authorize a trade.
For charts: tap a candle or swing to inspect it; drag horizontally to browse; wheel or +/− to zoom; arrow keys to inspect; Home to restore the step view.
Make the conditions explicit.
Fix average type, length, source and timeframe. A 20-bar simple average on five-minute closes is not equivalent to a daily 20-period exponential average.
Example filter: allow an independently defined long setup only when the completed close is above a rising 20-bar simple average. The setup’s own trigger must still occur.
If the filter condition is absent at signal time, skip the setup. Once in a trade, follow the pre-defined exit rule rather than inventing an average-based exit.
The quoted rules are illustrative research specifications. Parameters are not recommendations and have not been presented as profitable backtest results.
What can this cost?
A filter can reduce trade count and can also remove profitable trades. It does not make poor stop placement or excessive size safe.
How the idea can fail.
In a range, repeated crossings can produce frequent changes of classification. A rising average may lag a turning price sequence.
Practical habit
Keep your entry rule unchanged when comparing filtered and unfiltered samples. Otherwise you cannot isolate what the filter changed.
Go one level deeper
Report retained trades, excluded trades, net R after costs and out-of-sample results—not just the best average length.
Check your understanding
What does a moving average directly use?
Further reading & boundaries
These examples are original educational exercises. General background: CME Group: position sizing; Fidelity: technical analysis. Neither endorses these example rules. No hypothetical result establishes future profitability.
Editorial revision: 13 October 2026 · Synthetic data · Educational use