Stonkalysis: Stonk Valuation

#trading algorithm #1

#library(pracma)
#training_data <- read.csv("GE.csv")
#ma_LONG <- 26 #days for long
#ma_SHORT <- 12 #days for short
#ma_SIGNAL <- 9
#ma_DAYS <- 26 #days to use for average of absolute maximas

#training
#portfolio <- 1000
#portfolio2 <- 1000
#shares <- 0
#shares2 <- 0

#average_price <- c()
#number_of_buys <- 0
#number_of_sells <- 0
#counter <- 0
#count <- 0

#MACD
#long <- movavg(training_data[,5],ma_LONG, type="e")
#short <- movavg(training_data[,5],ma_SHORT, type="e")
#MACD <- short - long
#SIGNAL <- movavg(MACD,ma_SIGNAL, type="e")
#hist_difference <- MACD - SIGNAL
#p_n_avg <- check_sign(hist_difference)
#p_avg <- p_n_avg[1]
#n_avg <- p_n_avg[2]
	
	
#calculates average of historical MACD differences
#check_sign <- function(hist_difference) {
#	pp_avg <- 0
#	nn_avg <- 0
#	for(i in 1:length(hist_difference)) {
#		isS <- sign(hist_difference[i])
#		if(!is.null(isS)) {
#			if(isS) {
#				pp_avg <- c(pp_avg, hist_difference[i])
#			} else if (!isS) {
#				nn_avg <- c(nn_avg, hist_difference[i])
#			}
#		}
#	}
#	if(length(nn_avg) > 1) {
#		#if(length(nn_avg) > 2) {
#		#	nn_avg <- (mean(nn_avg[-1]) - sd(nn_avg[-1]))
#		#} else {
#			nn_avg <- mean(nn_avg[-1])
#		#}
#	} else {
#		nn_avg <- 0
#	}
#	if(length(pp_avg) > 1) {
#	#	if(length(pp_avg) > 2) {
#	#		pp_avg <- (mean(pp_avg[-1]) - sd(pp_avg[-1]))
#	#	} else {
#			pp_avg <- mean(pp_avg[-1])
#	#	}
#	} else {
#		pp_avg <- 0
#	}
#	return(c(pp_avg,nn_avg))
#}

#check if positive, negative, or zero
#sign <- function(difference) {
#	if(difference > 0) {
#		return(TRUE)
#	} else {
#		if(difference == 0) {
#			return(NULL)
#		} else {
#			return(FALSE)
#		}
#	}
#}

#training
#for(i in (ma_DAYS+1):nrow(training_data)) {
#	if(counter >= 30) {
#		portfolio <- portfolio + 1000
#		portfolio2 <- portfolio2 + 1000
#		counter <- 0
#		count <- count + 1
#	}
#	counter <- counter + 1
#	price <- training_data[i,5] #testing on closing price
#	difference <- hist_difference[i]
#	p_n_avg <- check_sign(hist_difference[(i - ma_DAYS):(i - 1)])
#	p_avg <- p_n_avg[1]
#	n_avg <- p_n_avg[2]
#
#	if(i == (ma_DAYS+1)) { #initial investment
#		can_buy <- floor(portfolio / price) 
#		portfolio <- portfolio - (can_buy * price)
#		number_of_buys <- number_of_buys + 1
#		shares <- can_buy
#		average_price <- price
#		
#		
#		
#		can_buy <- floor(portfolio2 / price) 
#		portfolio2 <- portfolio2 - (can_buy * price)
#		shares2 <- can_buy
#		
#	}
#
#	isS <- sign(difference)
#	if(!is.null(isS)) {
#		if(isS) {
#			if(difference > p_avg) {
#				if(shares > 0) {
#					if((mean(average_price) * shares) < (shares * price)) {
#						portfolio <- portfolio + (shares * price)
#						shares = 0
#						number_of_sells <- number_of_sells + 1
#						average_price <- c()
#					}
#				}
#			}
#		} else if (!isS) {
#			if(difference < n_avg) {
#				if(portfolio > price) {
#					can_buy <- floor(portfolio / price) 
#					if(can_buy > 0) { 
#						portfolio <- portfolio - (can_buy * price)
#						number_of_buys <- number_of_buys + 1
#						shares <- shares + can_buy
#						average_price <- c(average_price, price) 
#					}
#				}
#				if(portfolio2 > price) {
#					can_buy <- floor(portfolio2 / price) 
#					if(can_buy > 0) { 
#						portfolio2 <- portfolio2 - (can_buy * price)
#						shares2 <- shares2 + can_buy
#					}
#				}
#				
#			}
#		}
#	}
#	if(i == nrow(training_data)) {
#		if(shares > 0) {
#				portfolio <- (shares * price)
#				shares = 0
#				number_of_sells <- number_of_sells + 1
#				average_price <- 0
#		}
#		investment <- (count * 1000) + 1000
#		print(paste("Investment: ", investment, sep=""))
#		print(paste("Portfolio: ", portfolio, sep=""))
#		print(paste("Return: %", round(((portfolio - investment)/investment)*100, digits=2), sep=""))
#		print(paste("Bought: ", number_of_buys, sep=""))
#		print(paste("Sold: ", number_of_sells, sep=""))
#		
#		print(paste("Portfolio2: ", portfolio2 + (shares2 * price), sep=""))
#		print(paste("Return: %", round((((portfolio2 + (shares2 * price)) - investment)/investment)*100, digits=2), sep=""))
#		
#	}
#}