loops - How to find variance and count periods until condition is met after event in R -

--sorry bad title, suggestions make more clear?--

i have following data frame:

df <- data.frame( day = c(1,2,3,4,5,6,7,8,9,10,11),                   score = c(67,51,52,57,66,63,63,68,64,57,77),                   attempt = c(0,1,0,1,0,0,0,1,0,0,0)) 

i want calculate how many days takes on threshold of >10% in scores each time attempt occurs. threshold should calculated scores occurred after attempt.

the threshold calculated percentage difference attempts until next score >10%. first value in table below 57/51-1=0.12

attempt = 1

as result, table shows me actual percentage deviation on score @ time attempt number of days took.

day score  attempt  threshold   periods 1    67       0      2    51       1        12%         1 3    52       0      4    57       1        16%         0 5    66       0      6    63       0      7    63       0      8    68       1        13%         2 9    64       0      10   57       0      11   77       0      

if don't have threshold yet, can calculate follows. i'll assume have starting point start_score:

start_score  <- 45 later_scores <- df$score[df$attempt == 1]  target       <- c(start_score, later_scores) # 45 51 57 68      these want calculate percentage increase:      # -length(target) remove last value of target denominator pct_increase <- (diff(target) / target[-length(target)]) * 100  df$threshold[df$attempt == 1] <- pct_increase 

with threshold column in place, can proceed: find rows in df$threshold > 10, precede indices 0 , calculate number of rows (i.e. periods) between rows df$threshold > 10:

inds <- c(0, which(df$threshold > 10))  df$periods       <- rep(na, 11) df$periods[inds] <- diff(inds)-1  # day score attempt threshold    periods # 1    67       0        na      na # 2    51       1  13.33333       1 # 3    52       0        na      na # 4    57       1  11.76471       1 # 5    66       0        na      na # 6    63       0        na      na # 7    63       0        na      na # 8    68       1  19.29825       3 # 9    64       0        na      na #10    57       0        na      na #11    77       0        na      na 

edit find sell point 'attempt': first value in 'df$score' troublesome illustration, because if buy @ point, you'll able sell @ +10% in last period. however, if have bought, should've sold in first period immediately. therefore deleted value data frame:

df <- data.frame(day = c(2,3,4,5,6,7,8,9,10,11),                  score = c(51,52,57,66,63,63,68,64,57,77),                  attempt = c(1, rep(na, 9))) 

if understand correctly, sell after score reached 10% higher score in period in bought. parallel this, buy again in period sell, right? wait selling newly bought shares (?) until scores have again risen 10%:

sell_time1 <- 1 repeat{     sell_thres <- df$score[sell_time1] * 1.1   sell_time2 <- min( which( (df$score > sell_thres) & (df$day > df$day[sell_time1]) ))   ifelse(sell_time2 == sell_time1, break, sell_time1 <- sell_time2)   df$attempt[sell_time1] <- 1 } 

this produce warning, because @ point sell_time2 in second line of repeat{} try take minimum value of empty vector. in application, nothing worry about. result in:

#    day  score attempt #    2    51       1 #    3    52      na #    4    57       1 #    5    66       1 #    6    63      na #    7    63      na #    8    68      na #    9    64      na #   10    57      na #   11    77       1