# Increasing number of observations

have a data structured like below

data1

  time y
1   10 2
2   20 3
3   30 4
4   40 5
5   50 6


of data1 I make data2

  time y time2 y2
1   10 2    20  3
2   20 3    30  4
3   30 4    40  5
4   40 5    50  6


For data2 I fit some non linear model y2 = f(time, time2, y).

However, I wonder what happens if I will use all possible time and y combinations instead of just shortest one (these which area measured after each other <- compare data2 and data3

data3

    time y time2 y2
1   10 2    20  3
2   20 3    30  4
3   30 4    40  5
4   40 5    50  6
5   10 2    30  4
6   10 2    40  5
7   10 2    50  6
8   20 3    40  5
9   20 3    50  6
10   30 4    50  6


Can you answer why this approach is wrong and what is wrong with that?

• I don't understand what it is you want to do. Do you want to calculate the difference in time for all combinations of y? – JonB Sep 4 '15 at 8:48
• I meant rather general. I modified the question. – Mateusz1981 Sep 4 '15 at 9:17
• I'm sorry, but I still don't understand. What does the data represent and what is it that you want to do with it? What kind of analysis do you want to do you this data? – JonB Sep 4 '15 at 9:20
• the df1 data set contains the measured values. Later I create sequences t1, t2, y1, y2 to fit this to so called "difference equations" y2 ~ f(y1, t1, t2). The question is if I can use all possible combination of measures or just closest to each other (comapre data sets with just 4 observations and 7 observations) – Mateusz1981 Sep 4 '15 at 9:27
• Ok. Sorry, I haven't heard about difference equations so I can't help. – JonB Sep 4 '15 at 9:31