This article introduces a quantitative trading strategy that combines Simple Moving Average (SMA) and Adaptive Moving Average (ALMA). This strategy incorporates multiple technical indicators and generates trading signals based on different parameter settings.
I. Strategy Principle
The core of this strategy is the combination of SMA and ALMA with different parameter settings. SMA is a very common trend-following indicator that shows the direction and momentum of the trend by calculating the arithmetic mean of closing prices over a period of time. ALMA is similar to SMA in averaging historical prices, but it adds two adjustable parameters, α and σ, which make it more sensitive to market changes than SMA.
The strategy first calculates three SMAs representing short-term, medium-term, and long-term trends, respectively. At the same time, it calculates three ALMAs to represent the moving averages at different timeframes. The crossovers between SMA and ALMA form multiple sets of indicators. When the short-term SMA crosses over the medium-term SMA, a buy signal is generated. When the short-term SMA crosses below the medium-term SMA, a sell signal is generated. With the adjustable parameters of ALMA, the signals can respond to the market more quickly.
In addition, the Relative Strength Index (RSI) is introduced to assist in identifying overbought and oversold conditions. When RSI is higher than the overbought threshold, the market is considered overbought. In this case, even if the SMA and ALMA generate buy signals, they may be misleading. Similarly, when RSI is lower than the oversold line, sell signals from the indicators may miss rebounds. So the auxiliary judgment of RSI can avoid certain trapping risks.
By comprehensively utilizing the parameter settings of SMA, ALMA, and RSI, as well as the cross combinations between indicators of different parameters, relatively sensitive trading strategy signals can be formed. Also, the overbought and oversold judgments from RSI can further optimize entry timing and reduce trapping probability.
II. Advantages of the Strategy
The biggest advantage of this strategy is the flexible combination and application of indicator parameters. Both SMA and ALMA are flexible in adjusting parameters to represent different types of moving averages. RSI can also control the frequency of signals by adjusting parameters. The combination of these indicators complements each other and forms trading signals, which can optimize the timing of entries.
Compared with a single SMA indicator, ALMA enhances the sensitivity to market changes and can respond to trend reversals faster. Also, the auxiliary RSI judgment further avoids blindly following the signals from the moving averages. Therefore, this strategy overall has relatively strong adaptability and optimization capabilities.
Another advantage is the diversity of signal sources of the strategy. The interactions between SMAs and ALMAs at different timeframes provide multi-layered references for the strategy. This can filter out random market noise to some extent and make the signals more reliable.
In general, this strategy has flexible parameters and generates stable signals, making it suitable for algorithmic trading across different products.
III. Potential Risks
Although this strategy has certain advantages, there are still some risks to note when applying it in practice.
First, overoptimization problems caused by indicator settings. SMA, ALMA, and RSI are freely adjustable, but improper tuning may lead to overoptimization and the inability to adapt to long-term structural changes in the market. This requires cautious parameter settings based on the characteristics of different products, instead of merely pursuing short-term results.
Secondly, the strategy signals may lag. Although ALMA responds faster than SMA, there is still a certain lag. In rapidly changing markets, this may result in missing the optimal entry timing. Here we may consider combining some leading indicators to optimize.
Finally, the contradictory signals from multiple indicators need to be watched out for. At certain times, different indicators may give conflicting indications. This needs clear priority rules based on experience to resolve.
In summary, this strategy is not perfect and still requires continuous adjustment and optimization in practice. But its flexible parameter settings and advantages of multiple indicator combinations make it a viable algorithmic trading system for the long term.
IV. Summary
In this article, we have introduced in detail a quantitative trading strategy that combines SMA, ALMA, and RSI. Through flexible combinations of the indicators, it forms signals that are sensitive to the markets. Compared with single indicators, it has stronger adaptability and noise filtering capabilities. But we also need to pay attention to potential issues like overoptimization, signal lag, and judgment errors. Overall, this strategy is reasonably constructed and can generate stable algorithmic trading signals through continuous optimization.
/*backtest
start: 2023-09-06 00:00:00
end: 2023-09-13 00:00:00
period: 5m
basePeriod: 1m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/
//The plotchar UP/DOWN Arrows is the crossover of the fastest MA and fastest IIR MAs
//
//The dots at the bottom are the two simple averages crossing over
//
//The count over/under the candles is the count of bars that the SMAs on their
//respective resolution are fanning out.
//
//The colored background indicates a squeeze, lime=kinda tight : green=very tight squeeze. based on the 3 IIRs
//
//To answer my own question in a forum, looking at the code, i couldn't figure out how to get it from another timeframe
//and run the same calculations with the same results. My answer in the end was to scale the chosen MA length
//in the corresponding CurrentPeriod/ChosenMAPeriod proportion. This results in the same line in the same place when browsing through the
//different time resolutions. Somebody might find this invaluable
//
//The counts are for MA's fanning out, or going parabolic. Theres IIRs, Almas, one done of the other. A lot.
//The arrows above and below bars are from standard RSI numbers for OB/OS
//
//The IIRs changes color depending on their slope, which can be referenced easily with a variable.
//
//The backgrond on a bar-by-bar basis is colored when 2 sets of moving averages are in a squeeze, aka
//when price is consolidating.
//
//This aims to help the trader combine conditions and entry criteria of the trade and explore these options visually.
//They detail things from all time-frames on the current one. I prefer it because of the fractal nature of price-action, both large and small,
//either yesterday or last year. For best results, go long in short-term trades when the long-term trend is also up.
//and other profitable insights. This is also a great example of an automation algorith.
//
//The pretty ribbon is my script called 'Trading With Colors'. Use them together for fanciest results. 55/233 is my Fib Cross (golden/death) Compare it to the classic 50/200 if
//you get bored. I believe it simply works better, at least for Crypto.
//
//Evidently, I am a day-trader. But this yields higher profits on larger time-frames anyways, so do play around with it. Find what works for you.
//Thanks and credit for code snippets goes to:
//matryskowal
//ChrisMoody, probably twice
//Alex Orekhov (everget)
//author=LucF and midtownsk8rguy, for PineCoders
//If you use code from this, real quick search for perhaps the original and give them a shoutout too. I may have missed something
//Author: Sean Duffy
//@version=4
strategy(title = "Combination Parabolic MA/IIR/ALMA Strategy",
shorttitle = "MA-QuickE",
overlay = true,
backtest_fill_limits_assumption = 0,
default_qty_type = strategy.cash,
default_qty_value = 1000,
initial_capital = 1000,
currency = currency.USD,
linktoseries = true)
// calc_on_order_fills = true,
// calc_on_every_tick = true,
// Input Variables
showFIBMAs = input(false, type=input.bool, title="═══════════════ Show Fibby MAs ═══════════════")
maRes = input(960, type=input.integer, title="MA-Cross Resolution")
mal1 = input(8, type=input.integer, title="MA#1 Length")
mal2 = input(13, type=input.integer, title="MA#2 Length")
mal3 = input(34, type=input.integer, title="MA#3 Length")
loosePercentClose = input(1.1, type=input.float, title="SMA LooseSqueeze Percent")
showIIRs = input(false, type=input.bool, title="═══════════════════ Show IIRs ═══════════════════")
iirRes = input(60, type=input.integer, title="IIR Resolution")
percentClose = input(title="IIR Squeeze PercentClose", type=input.float, defval=.8)
iirlength1 = input(title="IIR Length 1", type=input.integer, defval=34)
iirlength2 = input(title="IIR Length 2", type=input.integer, defval=144)//input(title="ATR Period", type=input.integer, defval=1)
iirlength3 = input(title="IIR Length 3", type=input.integer, defval=720)//input(title="ATR Period", type=input.integer, defval=1)
showIIR1 = input(true, type=input.bool, title="Show IIR1")
showIIR2 = input(true, type=input.bool, title="Show IIR2")
showIIR3 = input(true, type=input.bool, title="Show IIR3")
showCounts = input(true, type=input.bool, title="═════════════ Show Parabolic MA Counts ════════════")
showSignals = input(true, type=input.bool, title="══════════════ Show Buy/Sell Signals ══════════════")
showBackground = input(true, type=input.bool, title="══════════════ Show Background Colors ══════════════")
//runStrategy = input(true, type=input.bool, title="══════════════ Run Strategy ══════════════")
debug = input(false, type=input.bool, title="══════════════ Show Debug ══════════════")
barLookbackPeriod = input(title="══ Bar Lookback Period ══", type=input.integer, defval=5)
percentageLookbackPeriod = input(title="══ Percentage Lookback Period ══", type=input.integer, defval=1)
bullcolor = color.green
bearcolor = color.red
color bgcolor = na
var bool slope1Green = na
var bool slope2Green = na
var bool slope3Green = na
var bool buySignal = na
var bool sellSignal = na
var bool bigbuySignal = na
var bool bigsellSignal = na
bool smbuySignal = false
bool smsellSignal = false
var bool insqueeze = na
var bool intightsqueeze = na
var bool infastsqueeze = na
var bool awaitingEntryIn = false
// My counting variables
var int count1 = 0
var float madist1 = 0
var int count2 = 0
var float madist2 = 0
var int sinceSmSignal = 0
var entryPrice = 0.0
var entryBarIndex = 0
var stopLossPrice = 0.0
// var updatedEntryPrice = 0.0
// var alertOpenPosition = false
// var alertClosePosition = false
// var label stopLossPriceLabel = na
// var line stopLossPriceLine = na
positionType = "LONG" // Strategy type, and the only current option
hasOpenPosition = strategy.opentrades != 0
hasNoOpenPosition = strategy.opentrades == 0
strategyClose() =>
if (hasOpenPosition)
if positionType == "LONG"
strategy.close("LONG", when=true)
else
strategy.close("SHORT", when=true)
strategyOpen() =>
if (hasNoOpenPosition)
if positionType == "LONG"
strategy.entry("LONG", strategy.long, when=true)
else
strategy.entry("SHORT", strategy.short, when=true)
checkEntry() =>
buysignal = false
if (hasNoOpenPosition)
strategyOpen()
buysignal := true
// if (slope1Green and (trend1Green or trend2Green) and awaitingEntryIn and hasNoOpenPosition)
// strategyOpen()
// buysignal := true
buysignal
checkExit() =>
sellsignal = false
// if (trend1Green == false and trend2Green == false) // to later have quicker exit strategy
// sellsignal := true
// strategyClose()
if (hasOpenPosition)
sellsignal := true
strategyClose()
sellsignal
multiplier(_adjRes, _adjLength) => // returns adjusted length
multiplier = _adjRes/timeframe.multiplier
round(_adjLength*multiplier)
//reset the var variables before new calculations
buySignal := false
sellSignal := false
smbuySignal := false
smsellSignal := false
bigbuySignal := false
bigsellSignal := false
ma1 = sma(close, multiplier(maRes, mal1))
ma2 = sma(close, multiplier(maRes, mal2))
ma3 = sma(close, multiplier(maRes, mal3))
madist1 := abs(ma1 - ma2)
madist2 := abs(ma1 - ma3) // check if MA's are fanning/going parabolic
if (ma1 >= ma2 and ma2 >= ma3 and madist1[0] > madist1[1]) //and abs(dataB - dataC >= madist2) // dataA must be higher than b, and distance between gaining, same with C
count1 := count1 + 1
else
count1 := 0
if (ma1 <= ma2 and ma2 <= ma3 and madist1[0] > madist1[1]) //<= madist2 and dataB <= dataC) //and abs(dataB - dataC >= madist2) // dataA must be higher than b, and distance between gaining, same with C
count2 := count2 + 1
else
count2 := 0
crossoverAB = crossover(ma1, ma2)
crossunderAB = crossunder(ma1, ma2)
plot(showFIBMAs ? ma1 : na, linewidth=3)
plot(showFIBMAs ? ma2 : na)
plot(showFIBMAs ? ma3 : na)
// Fast Squeese Check WORK IN PROGRESS
//
float singlePercent = close / 100
if max(madist1, madist2) <= singlePercent*loosePercentClose
bgcolor := color.yellow
infastsqueeze := true
else
infastsqueeze := false
// IIR MOVING AVERAGE
f(a) => a[0] // fixes mutable error
iirma(iirlength, iirsrc) =>
cf = 2*tan(2*3.14159*(1/iirlength)/2)
a0 = 8 + 8*cf + 4*pow(cf,2) + pow(cf,3)
a1 = -24 - 8*cf + 4*pow(cf,2) + 3*pow(cf,3)
a2 = 24 - 8*cf - 4*pow(cf,2) + 3*pow(cf,3)
a3 = -8 + 8*cf - 4*pow(cf,2) + pow(cf,3)
//----
c = pow(cf,3)/a0
d0 = -a1/a0
d1 = -a2/a0
d2 = -a3/a0
//----
out = 0.
out := nz(c*(iirsrc + iirsrc[3]) + 3*c*(iirsrc[1] + iirsrc[2]) + d0*out[1] + d1*out[2] + d2*out[3],iirsrc)
f(out)
iirma1 = iirma(multiplier(iirRes, iirlength1), close)
iirma2 = iirma(multiplier(iirRes, iirlength2), close)
iirma3 = iirma(multiplier(iirRes, iirlength3), close)
// adjusts length for current resolution now, length is lengthened/shortened accordingly, upholding exact placement of lines
// iirmaD1 = security(syminfo.tickerid, tostring(iirRes), iirma1, barmerge.gaps_on, barmerge.lookahead_on)
// iirmaD2 = security(syminfo.tickerid, tostring(iirRes), iirma2, barmerge.gaps_on, barmerge.lookahead_on)
// iirmaD3 = security(syminfo.tickerid, tostring(iirRes), iirma3, barmerge.gaps_on, barmerge.lookahead_on)
slope1color = slope1Green ? color.lime : color.blue
slope2color = slope2Green ? color.lime : color.blue
slope3color = slope3Green ? color.lime : color.blue
plot(showIIR1 and showIIRs ? iirma1 : na, title="IIR1", color=slope1color, linewidth=2, transp=30)
plot(showIIR2 and showIIRs ? iirma2 : na, title="IIR2", color=slope2color, linewidth=3, transp=30)
plot(showIIR3 and showIIRs ? iirma3 : na, title="IIR3", color=slope3color, linewidth=4, transp=30)
// checks slope of IIRs to create a boolean variable and and color it differently
if (iirma1[0] >= iirma1[1])
slope1Green := true
else
slope1Green := false
if (iirma2[0] >= iirma2[1])
slope2Green := true
else
slope2Green := false
if (iirma3[0] >= iirma3[1])
slope3Green := true
else
slope3Green := false
// calculate space between IIRs and then if the price jumps above both
//float singlePercent = close / 100 // = a single percent
var float distIIR1 = na
var float distIIR2 = na
distIIR1 := abs(iirma1 - iirma2)
distIIR2 := abs(iirma1 - iirma3)
if (distIIR1[0] < percentClose*singlePercent and close[0] >= iirma1[0])
if close[0] >= iirma2[0] and close[0] >= iirma3[0]
bgcolor := color.green
insqueeze := true
intightsqueeze := true
else
bgcolor := color.lime
insqueeze := true
intightsqueeze := false
else
insqueeze := false
intightsqueeze := false
// if (true)//sinceSmSignal > 0) // cutting down on fastest MAs noise
// sinceSmSignal := sinceSmSignal + 1
// if (crossoverAB)
// //checkEntry()
// //smbuySignal := true
// sinceSmSignal := 0
// if (crossunderAB) // and all NOT greennot (slope1Green and slope2Green and slope3Green)
// //checkExit()
// //smsellSignal := true
// sinceSmSignal := 0
// else
// sinceSmSignal := sinceSmSignal + 1
f_draw_infopanel(_x, _y, _line, _text, _color)=>
_rep_text = ""
for _l = 0 to _line
_rep_text := _rep_text + "\n"
_rep_text := _rep_text + _text
var label _la = na
label.delete(_la)
_la := label.new(
x=_x, y=_y,
text=_rep_text, xloc=xloc.bar_time, yloc=yloc.price,
color=color.black, style=label.style_labelup, textcolor=_color, size=size.normal)
posx = timenow + round(change(time)*60)
posy = highest(50)
// CONSTRUCTION ZONE
// TODO: program way to eliminate noise and false signals
// MAYBEDO: program it to differentiate between a moving average bump and a cross
// I think the best way would be to calculate the tangent line... OR
// Take the slope of both going back a couple bars and if it's close enough, its a bounce off
// and an excellent entry signal
// program in quickest exit, 2 bars next to eachother both closing under, as to avoid a single wick from
// prompting to close the trade
// Some other time, have it move SMA up or down depending on whether trending up or down. Then use those MA crosses
//THIS CHECKS THE SLOPE FROM CURRENT PRICE TO BACK 10 BARS
checkSlope(_series) => (_series[0]/_series[10])*100 // it now returns it as a percentage
doNewX = input(true, type=input.bool, title="══════════ Show misc MA Cross Strategy ══════════")
iirX = input(13, title="IIRx Length: ", type=input.integer)
iirXperiod = input(21, title="IIRx Period/TF: ", type=input.integer)
iirX2 = input(144, title="IIRx2 Length: ", type=input.integer)
iirX2period = input(233, title="IIRx2 Period/TF: ", type=input.integer) //15
almaXperiod = input(defval=21, title="Alma of IIR1 Period: ", type=input.integer)
almaXalpha = input(title="Alma Alpha Value: ", defval=.99, maxval=.99, type=input.float)
almaXsigma = input(title="Alma Sigma Value: ", defval=8, type=input.float)
iirmaOTF = iirma(multiplier(iirXperiod, iirX), close)
iirma2OTF = iirma(multiplier(iirX2period, iirX2), close)
smaOTF = alma(iirmaOTF, almaXperiod, almaXalpha, almaXsigma) // maybe dont touch, its precise // I took the ALMA of the IIRMA, and i hope thats not cheating ;)
// I could have removed this. the multiplier function adjusts the length to fit the current timeframe while displaying the same
// smaXOTF = security(syminfo.tickerid, smaXperiod, smaOTF, barmerge.gaps_on, barmerge.lookahead_on)
// iirmaXOTF = security(syminfo.tickerid, iirXperiod, iirmaOTF, barmerge.gaps_on, barmerge.lookahead_on)
// iirmaX2OTF = security(syminfo.tickerid, iirX2period, iirma2OTF, barmerge.gaps_on, barmerge.lookahead_on)
plot(doNewX ? smaOTF : na, title="FastMA X-Over : ", color=color.blue, linewidth=1, transp=40)
plot(doNewX ? iirmaOTF : na, title="IIR MAx : ", color=color.purple, linewidth=1, transp=30)
plot(doNewX ? iirma2OTF : na, title="IIR MAx : ", color=color.purple, linewidth=2, transp=20)
iirma2Up = iirma2OTF[0] > iirma2OTF[1] // just another slope up/down variable.
//calculate spaces between averages
distiiralma = abs(iirmaOTF - smaOTF)
crossoverFast = crossover(iirmaOTF[0], smaOTF[0]) // and (iirmaOTF[1] <= smaOTF[1])
crossunderFast = crossunder(iirmaOTF[0], smaOTF[0]) // and (iirmaOTF[1] >= smaOTF[1])
if (crossoverFast and iirma2Up == true) // and (count1 != 0))// or close[0] < (lowest(barLookbackPeriod) + singlePercent*3))) // must be at most a few percent up from a recent low. Avoid buying highs :P
buySignal := true
strategyOpen()
// if (slope1Green and slope2Green and slope3Green and infastsqueeze == false)
// checkEntry()
if (crossunderFast)
sellSignal := true
checkExit()
// I feel like I didn't cite the OG author for this panel correctly. I hope I did, but there are extentions of his/her work in multiple places.
// I could have gotten it confused.
if (debug)
f_draw_infopanel(posx, posy, 18, "distiiralma from IIR: " + tostring(distiiralma), color.lime)
//f_draw_infopanel(posx, posy, 16, "distiirs: " + tostring(distiirX1), color.lime)
f_draw_infopanel(posx, posy, 14, "Value of iirmaOTF: " + tostring(iirmaOTF), color.lime)
f_draw_infopanel(posx, posy, 6, "slope X: " + tostring(abs(100 - checkSlope(iirmaOTF))), color.lime)
f_draw_infopanel(posx, posy, 12, "value of smaOTF: " + tostring(smaOTF), color.lime)
f_draw_infopanel(posx, posy, 6, "slopeAlma: " + tostring(abs(100 - checkSlope(smaOTF))), color.lime)
f_draw_infopanel(posx, posy, 2, "slopeIIR2 " + tostring(abs(100 - checkSlope(iirma2OTF))), color.lime)
f_draw_infopanel(posx, posy, 2, "slopeIIR2 " + tostring(abs(100 - checkSlope(iirma2OTF))), color.lime)
// I kept this separate because it discludes the calculations. Its hard to hold a train of thought while fishing for the right section
bgcolor(showBackground ? bgcolor : na)
plotshape(showSignals ? buySignal : na, location=location.bottom, style=shape.circle, text="", size=size.tiny, color=color.blue, transp=60)
plotshape(showSignals ? sellSignal : na, location=location.bottom, style=shape.circle, text="", size=size.tiny, color=color.red, transp=60)
plotchar(showSignals and smbuySignal, title="smBuy", location=location.belowbar, char='↑', size=size.tiny, color=color.green, transp=0)
plotchar(showSignals and smsellSignal, title="smSell", location=location.abovebar, char='↓', size=size.tiny, color=color.orange, transp=0)
// can not display a variable. Can only match the count to a corresponding plotchar
// to display a non-constant variable, use the debug box, which was so kindly offered up by our community.
plotchar(showCounts and count1==1, title='', char='1', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==2, title='', char='2', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==3, title='', char='3', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==4, title='', char='4', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==5, title='', char='5', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==6, title='', char='6', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==7, title='', char='7', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==8, title='', char='8', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1==9, title='', char='9', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count1>=10, title='', char='$', location=location.belowbar, color=#2c9e2c, transp=0)
plotchar(showCounts and count2==1, title='', char='1', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==2, title='', char='2', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==3, title='', char='3', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==4, title='', char='4', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==5, title='', char='5', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==6, title='', char='6', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==7, title='', char='7', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==8, title='', char='8', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2==9, title='', char='9', location=location.abovebar, color=#e91e63, transp=0)
plotchar(showCounts and count2>=10, title='', char='$', location=location.abovebar, color=#e91e63, transp=0)
showRSIind = input(true, type=input.bool, title="═══════════════════ Show RSI Arrows ═══════════════════")
// Get user input
rsiSource = input(title="RSI Source", type=input.source, defval=close)
rsiLength = input(title="RSI Length", type=input.integer, defval=14)
rsiOverbought = input(title="RSI Overbought Level", type=input.integer, defval=80)
rsiOversold = input(title="RSI Oversold Level", type=input.integer, defval=20)
// Get RSI value
rsiValue = rsi(rsiSource, rsiLength)
isRsiOB = rsiValue >= rsiOverbought
isRsiOS = rsiValue <= rsiOversold
// Plot signals to chart
plotshape(isRsiOB, title="Overbought", location=location.abovebar, color=color.red, transp=0, style=shape.triangledown)
plotshape(isRsiOS, title="Oversold", location=location.belowbar, color=color.green, transp=0, style=shape.triangleup)
//reset the var variables before new calculations
buySignal := false
sellSignal := false
smbuySignal := false
smsellSignal := false
bigbuySignal := false
bigsellSignal := false