Reference appendix
Moving averages
Different ways of smoothing price into a single trend line — the building blocks most other indicators are made from. The number attached to each one is its lookback period; open any entry for what that setting actually changes.
How these fit together
Every average here does the same thing — collapse a stretch of price into one number — and they differ only in which part of that stretch they weight. That makes the whole family a single sliding scale from slow and stable to fast and jumpy, and picking one is really picking a position on it. The information, though, is rarely in a single line: it is in how two lines sit relative to each other. A fast average is a more recent summary than a slow one, so when the fast crosses above the slow, the recent stretch of trading has simply been stronger than the older stretch. That is the entire content of a crossover signal, and it is why the pairs everyone quotes — 20 with 50, 50 with 200 — are chosen to compare meaningfully different horizons.
- Speed at equal periods: TMA → SMA → WMA → EMA → DEMA → TEMA/HMA
- 20 crosses above 50 → short-term momentum turning up, typically weeks ahead of the long-term trend
- 50 crosses above 200 → the golden cross; 50 below 200 → the death cross
- 20 > 50 > 200 and all rising → a fully aligned uptrend, the configuration trend systems wait for
- The slope matters more than the touch — a flat average is not support, it is a range
- Every crossover needs a filter — ADX above 25, or the same signal in a range whipsaws repeatedly
- Common horizons: 20 ≈ a month · 50 ≈ a quarter · 200 ≈ a trading year
Simple moving average
SMAstudiedThe average closing price over the last n periods, each period weighted equally. The baseline for spotting trend direction and a building block for many other indicators.
Exponential moving average
EMAstudiedLike SMA, but weights recent prices more heavily, so it reacts faster to new price action while still smoothing out noise.
Weighted moving average
WMAstudiedAssigns linearly increasing weight to more recent prices — the newest period counts most, the oldest counts least — landing between SMA and EMA in responsiveness.
Smoothed moving average
SMMAuntestedIncorporates all historical prices with a slowly decaying weight, producing an even smoother, laggier line than SMA or EMA.
Double exponential moving average
DEMAuntestedCombines an EMA with an EMA of that EMA to cancel out much of the lag inherent in a single EMA.
Triple exponential moving average
TEMAuntestedExtends DEMA's lag-cancelling trick with a third layer of EMA smoothing, tracking price even more tightly.
Triangular moving average
TMAuntestedAn SMA of an SMA — double-smoothed, weighting the middle of the lookback period most heavily, for a very smooth trend line.
Hull moving average
HMAuntestedBuilt from weighted moving averages of different lengths combined to cut lag while staying smooth, aiming to solve the classic smoothness-vs-lag tradeoff.
EMA clouds
EMA clouduntestedTwo EMAs of different lengths plotted together with the space between them shaded, so their crossovers and the cloud's width visualize trend direction and strength at a glance.
EMA / SMA / WMA / SMMA envelope
EnvelopeuntestedA pair of bands plotted a fixed percentage above and below the chosen moving average, marking a “normal” trading range; price pushing outside the envelope flags an extended move.
How to read the evidence tag
Every entry here carries one of three tags. The tag rates the evidence behind the claim the entry makes — not how popular the tool is, how old it is, or how confident the people teaching it sound. Close to half of the entries on this site sit on the weakest tier — including every one of the eighteen Smart Money Concepts, which are drawn here in the same ink as Wilder’s ADX and deserve to be told apart from it. That is the honest position rather than an oversight.
- defined33 entries
True by arithmetic. This entry describes a calculation or a piece of terminology, and makes no claim about what price will do next. Nothing here can be wrong except a typo.
- studied56 entries
The predictive claim attached to this entry has been tested in peer-reviewed research. Being on this tier does not mean it works — across the literature these edges are typically small, unstable between markets and periods, and often gone after costs.
- untested87 entries
Widely taught and widely believed, but no published test of the claim was located while building this page. Treat it as a description of what traders do, not as evidence that doing it pays.
The line between studied and untested is drawn by whether a peer-reviewed test of the specific claim could be found, and those tests are listed on the Sources page under “Evidence and base rates”. If you know of a published test of something tagged untested, the tag is wrong and should change — that is the point of labelling it at all.