Hello, here is the code of this strategy. I like mean reversion, this is my favourite concept in trading, I’m not so in love with trend following I must admit.

Your idea to count how much bars were above or below the auto-adjusted bands which act as dynamic overbought/oversold levels is quiet interesting. Please find below and attached the code. I let you test and propose improvement if any.

defparam cumulateorders = false

//parameters
P = 14 //RSI period
dev = 1.8 //deviation factor
minbarcount = 5 //quantity of bars that the RSI needs to be above or below its bands

//RSI SAB 
rsiplus = RSI[P](close)
bsup = 50+ (dev*STD(rsiplus))
binf = 50 - (dev*STD(rsiplus))

//RSI above buylevel and selllevel count
countB = 0
countS = 0
for i = 0 to minbarcount do 
  if rsiplus[i]>bsup[i] then 
   countB=countB+1
  elsif rsiplus[i]<binf[i] then
   countS=countS+1
  endif
next

//case BUY
if rsiplus crosses over binf and countS>=minbarcount then 
 BUY 1 SHARE AT MARKET
endif

if longonmarket and rsiplus>=70 then 
 SELL AT MARKET 
endif

//case SELL
if rsiplus crosses under bsup and countB>=minbarcount then 
 SELLSHORT 1 SHARE AT MARKET 
endif

if shortonmarket and rsiplus<=30 then 
 EXITSHORT AT MARKET 
endif

rsi-self-adjusting-bands-trading-strategy.png RSI-SAB-strategy.itf