CCI explained: the Commodity Channel Index and the ±100 levels

The Commodity Channel Index (CCI) measures how far price has moved away from its recent average, scaled by how much it usually deviates. A reading above +100 means price is unusually high compared with its recent behaviour; below −100, unusually low. What you do with that information depends on whether the market is ranging or trending, and that is where most of the confusion around CCI lives.

How CCI is calculated

Donald Lambert introduced CCI in 1980, originally for commodities, but it works on any price series. It has four steps, with a default period of 20:

1. Typical price (TP) for each candle = (high + low + close) / 3.
2. Average of TP: a simple moving average of typical price over 20 periods.
3. Mean deviation: the average of the absolute distances between each TP and that average.
4. CCI = (current TP − average TP) / (0.015 × mean deviation).

The constant 0.015 was chosen by Lambert so that most readings fall between −100 and +100. Unlike RSI or Stochastic, CCI has no upper or lower bound; in strong moves it can reach ±200, ±300 or more.

A worked example

Suppose on EUR/USD M5 the current typical price is 1.08600, the 20-period average of typical price is 1.08500 and the mean deviation is 0.00050.

CCI = (1.08600 − 1.08500) / (0.015 × 0.00050) = 0.00100 / 0.0000075 ≈ +133.

Price is 10 pips above its average, while it usually strays about 5 pips. CCI expresses that as +133: unusually stretched to the upside. Whether that means "about to pull back" or "strong enough to keep going" is not something the number itself can tell you.

Two opposite ways to use it

Mean reversion

In a range, traders read +100 as stretched and −100 as depressed, expecting price to drift back towards the average. A typical rule: wait for CCI to rise above +100 and then cross back below it, rather than acting the moment it crosses up. This is the mean reversion approach, and it is close to how people use RSI overbought and oversold zones.

Trend

Lambert's original idea was the opposite: a move above +100 indicates unusual strength and possibly the start of a trend; below −100, a downtrend. Trend traders stay with the move while CCI remains beyond the threshold.

Both interpretations are legitimate, and they contradict each other. The same +133 reading is a sell idea for one trader and a buy idea for another. Decide which regime you think you are in first, for example with ADX or market structure, and only then interpret CCI.

Divergence and zero-line crosses

Two other readings are common:

• Zero line. CCI above zero means typical price is above its average; below zero, beneath it. Crosses happen often and are weak signals by themselves.
• Divergence. If price makes a higher high but CCI makes a lower high, the push is losing relative strength. Divergences can persist for a long time in trends, so they are warnings, not timing tools.

False signals

• In a strong trend, CCI can stay above +100 for dozens of candles. Fading every reading above +100 means repeatedly betting against the trend.
• In a quiet market, the mean deviation becomes small, so a modest move can push CCI past ±100. A low-volatility "extreme" may be a two-pip move.
• Short CCI periods produce constant threshold crossings that mean very little.

Short expiries and OTC

On S5–M1 charts, CCI with a 20-period setting reacts to a few seconds or minutes of ticks. It oscillates rapidly, and ±100 crossings become frequent. Treat it as a description of how stretched price is right now, not as a timer for a short expiry. Many traders prefer to check CCI on M5 or above for context.

On Sky Trade's OTC instruments, prices are generated by Sky Trade's server and do not follow any real market. The model is built so the next move is unpredictable; CCI readings there describe the generated series and carry no edge, whether you read them as mean reversion or as trend.

How to use it in the Sky Trade terminal

CCI is in the oscillators group of the indicators list and appears in its own panel with a zero line. Since CCI, RSI and Stochastic measure similar things, choose one of them rather than stacking all three; the article on combining indicators explains why.