TASC's May 2023 edition of Traders' Tips features an article titled "An Adaptive Moving Average For Swing Trading" by Scott Cong. The article presents a new adaptive moving average (AMA) that adjusts its parameters automatically based on market volatility. The AMA tracks price closely during trending movements and remains flat during congestion areas.
█ CONCEPTS
Conventional moving averages (MAs) use a fixed lookback period, which may lead to limited performance in constantly changing market conditions. Perry Kaufman's adaptive moving average, first described in his 1995 book Smarter Trading, is a great example of how an AMA can self-adjust to adapt to changing environments. Scott Cong draws inspiration from Kaufman's approach and proposes a new way to calculate the AMA smoothing factor.
█ CALCULATIONS
Following Perry Kaufman's approach, Scott Cong's AMA is calculated progressively as: AMA = α * Close + (1 − α) * AMA(1), where:
Close = Close of the current bar
AMA(1) = AMA value of the previous bar
α = Smoothing factor between 0 and 1, defined by the lookback period
The smoothing factor determines the performance of AMA. In Cong's approach, it is calculated as: α = Result / Effort, where:
Result = Highest price of the n period − Lowest price of the n period
Effort = Sum(TR, n), where TR stands for Wilder’s true range values of individual bars of the n period
n = Lookback period
As the price range is always no greater than the total journey, α is ensured to be between 0 and 1.
wTR(...) can be replaced with nz(math.max(...), sHigh - sLow)
thermal_winds
⋅
It actually is really great, first i tried using its phase shift at longer periods to identify trend bias, it was really responsive even at 233, 377 and 610 look back periods. A better way i found is to create a cloud between this moving average and its delayed values. That is so responsive, very less choppy and very much accurate. Thank you so much
@mcbrearty1, Try plotting CAMA and CAMA{delay}, where delay can be any integer value. You can write this line of code in the end - plot(cama{input(5)}. See if it adds some more value - p.s. use square brackets instead of curly brackets, dunno why we can't use square brackets in comments
thermal_winds
⋅
Thank you so much, really great tool to have in arsenal