|
PureBytes Links
Trading Reference Links
|
On Tuesday, November 13, 2001, 2:09:54 PM, Bob Fulks wrote:
BF> The rounding should not cause that much of a difference.
True - I was assuming it would be essentially identical. That was the
surprise.
BF> With such a long average, you might not be giving the two cases the
BF> same number of bars of data. Make sure the charts start on the same
BF> date/time and that the MaxMarsBack setting are the same.
I was careful to make both charts identical.
BF> Also, I have seen cases where the 1-minute bars in TS2000i do not
BF> correspond to the tick data. (Your 45 minute chart will be built
BF> from the 1-minute bars, not the tick data.)
Ahhhh - good idea! I usually import a 'chunk' of data that gets added
to the end of the existing data, and always use "Automatically Rebuild
All History Settings..." I guess I assume this results in a good data
base on TS2k. Guess I gotta remember this is TS. :-)
Anyway, I re-imported all the data, and the inds are now within about
1.1 on SPX data. There was apparently some data problem (not visible
to the eye) that caused the large difference. I am still puzzled that
there is a 1.1 difference, though. I would not expect it to be even
that big, given the rounding is done correctly. Do you think the
rounding could have that big an effect?
In any case, why does TS2k round the data? I'd prefer to keep as much
precision as possible.
Many thanks for the reminder to question the data.
ztrader
|