QuanTAlib/docs/trendcomparison.md

Trend Indicators Comparison

Scale 1–10 where 10 = better for every column. Detailed evaluation criteria at the bottom of this doc.

• Accuracy: Preserve true movement structure (major trends and turning points) without distortion or artificial patterns.
• Timeliness: Minimal lag. Fast response to genuine movement changes and reversals.
• Overshoot Control: Remain within min/max of input, avoid generating artificial over-reaching levels and false threshold triggers.
• Smoothness: Noise suppression. Stable output with smooth derivatives (no erratic velocity/acceleration).

Indicator	Accuracy	Timeliness	Overshoot Control	Smoothness	Notes (revised)
ALMA	8	7	10	8	Positive-weight FIR; accurate-ish but still a lag tradeoff.
BESSEL	9	7	9	8	Strong shape/phase preservation; step response is well-behaved.
BILATERAL	7	6	10	8	Edge-preserving; excellent in ranging markets, variable smoothing by design.
BLMA	7	3	10	10	Standard DSP window; superior noise suppression but significant lag.
DEMA	4	9	3	6	Lag-canceling subtraction ⇒ structure distortion + overshoot risk.
DWMA	7	2	10	10	Ultra-smooth, but smears structure heavily (lag dominates).
EMA	8	6	10	8	Convex IIR (monotone) ⇒ faithful & stable, moderate lag.
HMA	6	9	3	7	Very fast but can ring/overshoot; “accurate” depends on regime.
HTIT	7	8	6	8	Trend extraction can be excellent but can distort around turns/cycles.
JMA	8	9	9	9	Great practical trend estimate; adaptive behavior can reshape structure.
KAMA	8	8	10	8	Variable-alpha EMA: stable, good structure, less lag in trends.
LSMA	3	8	5	3	Regression endpoint/projection: can deviate from true path + noisy.
MAMA	6	9	6	3	Phase-adaptive; fast but accuracy varies with cycle model fit.
MGDI	7	7	10	9	Stable “EMA-like” behavior; good smoothing, not especially fast.
PWMA	6	7	10	6	Positive weights (no overshoot) but can be twitchy vs noise.
RMA	8	4	10	9	Slower EMA ⇒ very stable + faithful, but laggier.
SMA	7	3	10	6	Baseline: faithful but slow; smoothness only moderate.
SSF	9	8	8	9	Excellent smoothing with relatively low lag; mild ringing possible.
T3	7	8	5	10	Extremely smooth; overshoot depends on tuning (can behave “too clever”).
TEMA	3	10	3	6	Near-zero lag feel, but structure distortion + overshoot common.
TRIMA	7	2	10	10	Very smooth FIR; structure preserved but delayed a lot.
USF	9	9	8	9	Low-lag smoother; very good overall, slight ringing possible.
VIDYA	7	8	10	7	Variable-alpha EMA: stable, responsive in trends, moderate smoothness.
WMA	7	7	10	5	Faster than SMA; less smooth; still faithful (positive weights).

Evaluation Criteria

Accuracy (preserving large-scale structure)

Moving average should maintain the important underlying structure of price movements (like major trends and cycles) while filtering out all smaller fluctuations; it should faithfully represent the true price trajectory over longer timeframes.

Timeliness (minimal lag)

Most moving averages lag behind price action - they indicate changes way after they've already happened. A good moving average minimizes this lag, responding quickly to genuine price movements without sacrificing other qualities, providing more actionable signals and earlier entries/exits.

Minimal overshoot

Overshoot occurs when a highly reactive moving average extends beyond the actual price extremes, creating false impressions of price levels never reached. TEMA, DEMA and HMA are examples of overshooting moving averages; good moving average should avoid this distortion, particularly during price reversals, preventing false triggers when used with threshold-based systems.

Smoothness (reduced noise)

A quality moving average filters out random price fluctuations (noise) that don't represent meaningful market activity, especially in steady non-volatile periods. This creates a clean, smooth line that clearly shows the underlying price direction without the jagged, erratic movements that could trigger false signals.