RTTADocumentation Built by Goblin Reactor ↗

GuppyMultipleMovingAverage

Incremental, causal technical analysis documentation

Summary

GuppyMultipleMovingAverage is RTTA's compact Guppy MMA: six short EMAs and six long EMAs are updated each bar, then summarized as short-group average, long-group average, and their spread.

Update API

result = rtta.GuppyMultipleMovingAverage(fillna=True).update(price)
# result.short_average, result.long_average, result.spread

Periods are fixed: short \(\{3,5,8,10,12,15\}\), long \(\{30,35,40,45,50,60\}\). With fillna=False, outputs are NaN until 60 samples have been seen.

Theory Of Operation

Daryl Guppy's Multiple Moving Average frames market participation as two populations: short-term traders and longer-term investors. Averaging each population's EMAs yields two summary curves; their separation (spread) is a simple trend-strength / agreement measure. For the full twelve-line ribbon, use GuppyMMARibbon.

Recurrence

Let \(x_t\) be price,

\[\mathcal{S} = \{3,5,8,10,12,15\}, \qquad \mathcal{L} = \{30,35,40,45,50,60\}.\]
\[\begin{aligned} S^{\text{avg}}_t &= \tfrac16 \sum_{p\in\mathcal{S}} \operatorname{EMA}_p(x_t) \\ L^{\text{avg}}_t &= \tfrac16 \sum_{p\in\mathcal{L}} \operatorname{EMA}_p(x_t) \\ \operatorname{spread}_t &= S^{\text{avg}}_t - L^{\text{avg}}_t \end{aligned}\]

Each \(\operatorname{EMA}_p\) uses \(\alpha = 2/(p+1)\) as in RTTA's class EMA. Nested EMAs always fill during warmup; the outer fillna only blanks the first 59 samples when false.

Implementation Notes

The recurrence is implemented in src/rtta/indicator.cpp in class GuppyMultipleMovingAverage. Result fields are short_average, long_average, and spread.

Reference