TRIX

TRIX, developed by Jack Hutson in the early 1980s, is a momentum oscillator that measures the one-period percentage change of a triple-exponentially-smoothed moving average of price. The triple smoothing is the whole point: by running the closing price through three successive EMAs, TRIX filters out the cycles and wiggles that are shorter than the chosen period, leaving behind only the momentum of the dominant trend. The result is a line that oscillates around a zero center. A positive TRIX means the smoothed price is accelerating upward; a negative TRIX means it is accelerating downward. Because so much noise has been removed before the rate-of-change is taken, TRIX tends to be far smoother than a raw ROC or even MACD, which makes its zero-line crossings and turns relatively clean signals rather than the constant chatter a single-smoothed oscillator produces. The trade-off for that smoothness is lag. Three layers of exponential smoothing push the signal further behind real price than a single EMA would. TRIX is therefore best understood as a trend-momentum confirmation tool — it tells you the prevailing trend still has fuel, or that its fuel is running out — rather than a precise top-and-bottom timer. In crypto, where intraday charts are noisy and prone to wicks, the noise rejection of TRIX is genuinely useful: a 15-period TRIX on the 4-hour or daily chart cuts through the chop that would flip a faster oscillator back and forth.

TRIX is built in four steps: 1. Compute EMA1 = EMA(close, n). 2. Compute EMA2 = EMA(EMA1, n) — an EMA of the first EMA. 3. Compute EMA3 = EMA(EMA2, n) — an EMA of the second EMA. 4. TRIX = the one-period percentage change of EMA3: (EMA3_today − EMA3_yesterday) / EMA3_yesterday × 100. The default period is 15. An optional signal line — an EMA (commonly 9) of the TRIX line — is sometimes overlaid for crossover entries, the same way the MACD signal line works. Because each EMA needs warm-up bars before it stabilizes, TRIX needs roughly three times the period in history before its values are trustworthy. With the default 15, expect about 45 bars of warm-up before the reading is meaningful — Quantinger returns no value until enough bars exist rather than printing an unstable early number.

**Zero-line crossovers.** TRIX crossing from below zero to above signals that smoothed momentum has turned positive — a trend-following long trigger. Crossing below zero is the bearish equivalent. Because of the triple smoothing, these crosses are infrequent and relatively reliable compared to a raw ROC. **Signal-line crossovers.** If you add the optional EMA(9) signal line, TRIX crossing above its signal can be used as an earlier, more sensitive entry than the zero-line cross — at the cost of more signals. **Trend confirmation filter.** Use TRIX > 0 as a regime gate: only take long setups from other indicators while TRIX is positive, and shorts only while it is negative. This keeps mean-reversion or breakout entries aligned with the dominant trend. **Divergence.** As with most momentum oscillators, price making a new high while TRIX makes a lower high warns that trend momentum is fading. TRIX divergences are cleaner than RSI/MACD divergences precisely because the line itself is smoother.

**Lag from triple smoothing.** Three EMAs in series mean TRIX reacts later than most oscillators. It confirms trends; it does not call exact tops or bottoms. In fast reversals you will give back a meaningful chunk before TRIX turns. **Whipsaws in ranges.** Despite the smoothing, in a tight sideways market TRIX hovers around zero and its crossovers become noise. Pair it with a regime filter (ADX or Choppiness Index) and only trust crossovers when a trend is actually present. **Warm-up sensitivity.** Values from the first ~3× the period bars are unreliable. Backtests that start signals too early on a freshly loaded chart can show phantom TRIX behavior. **Percentage scale is asset-relative.** TRIX is a percentage rate of change, so its typical magnitude depends on the asset's volatility. A TRIX of 1.0 is large on a sleepy pair and unremarkable on a volatile altcoin — read it relative to the instrument's own history, not as a fixed threshold.