Technical Indicators for CMT Candidates: Formulas, Signals, and Limits

Read the construction before the signal

The 2026 CMT Program Guide lists moving averages, indicator categories, directional movement, and Bollinger Bands in Level I objectives. Level II objectives include momentum interpretation, VWAP, and Bollinger applications. The public objectives do not establish which named indicator will appear on a particular paper or how often. This guide is a worked reference; Technical Indicators Explained teaches how to select and audit a tool.

Indicators transform underlying data. The transformation can smooth noise, normalize a range, or compare prices, but it cannot remove uncertainty. For every signal, identify input interval, lookback, calculation, market context, action trigger, and invalidation.

Moving averages

A simple moving average (SMA) is the arithmetic mean of the last n observations. For closes 20, 22, 24, and 26, the four-period SMA is (20 + 22 + 24 + 26)/4 = 23. A linearly weighted average puts explicit larger weights on recent observations. If weights 1, 2, 3, and 4 correspond to those four prices, the weighted mean is (1×20 + 2×22 + 3×24 + 4×26)/(1+2+3+4) = 240/10 = 24.

An exponential moving average (EMA) updates recursively. With alpha = 2/(n+1), the new EMA is alpha × current close + (1−alpha) × previous EMA. For n = 4, alpha = 0.4. If the prior EMA is 23 and the new close is 26, the updated EMA is 0.4×26 + 0.6×23 = 24.2. The starting value and platform convention matter. A faster average has less lag but can produce more crossings in a range.

A crossover is an event to investigate, not an automatic trade. If a short average crosses a long one after a large price jump, entry may be far from a sensible invalidation level. Compare the signal with trend structure and price risk.

RSI and momentum

RSI is a bounded momentum measure built from smoothed average gains and losses. Under the conventional formula, relative strength RS = average gain/average loss, and RSI = 100 − 100/(1 + RS). If average gain is 1.5 and average loss is 0.5 under the same smoothing window, RS = 3 and RSI = 100 − 100/4 = 75. This describes stronger recent gains than losses. It does not mean price must fall. If average loss is zero, an implementation needs a defined edge-case convention; it should not divide by zero.

“Overbought” and “oversold” are descriptions relative to an oscillator's chosen reference values. Strong trends can remain near extremes. A divergence occurs when price makes a new high while momentum fails to make a corresponding high, or vice versa. A divergence is an alert that acceleration has changed; it does not itself identify the exact price turning point. Require an independent price trigger if the decision is about a reversal.

MACD and stochastic

A conventional MACD line is a short EMA minus a longer EMA, often displayed with a signal-line EMA of the difference and a histogram equal to MACD minus signal. If the two EMAs are 105 and 102, MACD is +3. If its signal line is +2, the histogram is +1. The positive difference says the shorter smoothed price is above the longer; the positive histogram says MACD is above its own signal. Neither number says whether the asset is cheap or likely to rise tomorrow. Parameter choices affect lag and false-cross frequency.

A stochastic oscillator locates the close within a recent high-low range. A basic %K is 100 × (close − lowest low)/(highest high − lowest low). If the range is 40 to 60 and the close is 55, %K = 100 × 15/20 = 75. When high equals low, the denominator is zero; charting software must handle that case. A close near the top of the range can persist in an uptrend. Smoothed %D and crossovers introduce further delay.

Bollinger Bands and VWAP

Bollinger Bands conventionally place upper and lower lines around a moving average at a chosen standard-deviation multiple. For a teaching example, closes 8, 10, 12, and 10 have mean 10 and population variance (4 + 0 + 4 + 0)/4 = 2. Population standard deviation is about 1.414; bands at two standard deviations are roughly 7.17 and 12.83. A close above 12.83 is outside this prior window's upper band. It is a location fact, not a short-sale instruction. The software's window, standard-deviation convention, and whether the newest close enters the rolling calculation must be stated.

VWAP is the sum of traded price times volume divided by total volume over a specified period. If 100 shares trade at 10 and 300 at 12, VWAP is (100×10 + 300×12)/(100+300) = 4,600/400 = 11.50. The simple average of the two prices is 11, which ignores the larger volume at 12. Intraday VWAP often resets each session; an anchored VWAP starts at a selected event. State the anchor and data granularity. Execution relative to VWAP needs trade-side context: a buyer generally prefers paying less than a suitable benchmark, but fill quality also depends on urgency and market movement.

Build conditional rules, then try to break them

An example rule might use a rising average to define an uptrend, a momentum pullback to define a setup, and a close over a recent swing high for entry. The invalidation is a close beneath the pullback low or a fixed account-risk limit. This rule is illustrative, not an official CMT rule or a return claim. Test it in trends, ranges, and volatile gaps. Indicators derived from the same closes are not independent confirmation.

Common mistakes include comparing different bar intervals, failing to adjust split data, assuming a fixed RSI threshold works in every regime, and optimizing parameters after observing outcomes. An indicator should be judged by its construction and how it changes a decision, not by its visual appeal.

Self-check

A three-period SMA currently includes 9, 12, and 15. The next close is 18. What is the new SMA? If a short EMA is 14 and a long EMA is 11, what is the MACD line? Answer: Remove 9 and include 18: (12 + 15 + 18)/3 = 15. The MACD line is 14 − 11 = +3. Neither calculation establishes a profitable trade.

Next, hand-calculate one indicator from a short price series, then compare with your platform using identical settings and investigate any difference.