Thanks Gary how ever that wont work in this kind of instance where the trend
is higher
10
20
10
30
9
40
9
and also this one where the trend is clearly lower
50
40
50
30
51
20
51
lastly this one where the trend is even
20
19
21
Any other ideas
Thanks
> Consider:
>
[quoted text clipped - 21 lines]
>> all the functions like trend etc and stdev show loss and gain as no
>> difference
Pete_UK - 09 Jan 2006 19:18 GMT
Are these meant to be equally-spaced values (eg at some time interval)
? In other words, does the row number imply an x-value to these
y-values? If so, then you could do a least-squares fit to the data and
derive the best "slope" of the values to indicate whether it is
increasing, decreasing or level.
It's been a few years since I last did this, so I can't help you with
the formulae straightaway, but I hope this helps.
Pete