[R-SIG-Finance] tick data and one minute bar data appear out of line (IBrokers)

[Stephen Choularton]

Hi Joshua

Just trying to follow that comment 'If you look at the tick data versus 
the high-low range, you should see that all the tick prices fall within 
that range for every aggregate interval.'

The summary of the two datasets is:

  summary(bars)
      Index                       CBA.Close
  Min.   :2016-03-03 10:06:00   Min.   :48.31
  1st Qu.:2016-03-03 11:37:00   1st Qu.:48.38
  Median :2016-03-03 13:08:00   Median :48.45
  Mean   :2016-03-03 13:08:00   Mean   :48.46
  3rd Qu.:2016-03-03 14:39:00   3rd Qu.:48.53
  Max.   :2016-03-03 16:10:00   Max.   :48.73


and

summary(ticks)
    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
   47.82   48.41   48.44   48.47   48.49   49.00

Now that min figure is the result of some market closing phenomena and 
it really never went much below 48.3 so yes its true that they have a 
similar range and min and max but ticks gets up into the just under 49.0 
level around 10:15 and bars only gets to that level around 11:30 and 
again at noon while ticks is under 48.6 at that point.

Sounds like you don't think there is anything that unusual in the two 
datasets so I guess I had better take that on board and keep on trying 
to make this work.

Thanks for looking at it

Stephen Choularton PhD, FIoD

On 4/03/2016 2:50 PM, Joshua Ulrich wrote:
> On Thu, Mar 3, 2016 at 9:38 PM, Stephen Choularton
> <stephen at organicfoodmarkets.com.au> wrote:
>> Hi
>>
>> I have been working on the spread between Commonwealth Bank (CBA.ASX) and
>> National Australia Bank (NAB.ASX).
>>
>> I was a bit confused by the performance of my dealing loop so I took out the
>> spread movement from rqHistoricData 1 minute bars using the close and
>> plotted it and also the spread derived from the tick data from reqMktData
>> which is being used to drive my dealing algorithm.
>>
>> They are in the same ballpark but really rather different.  The graph shows
>> the position best and can be viewed at this url:
>> http://www.organicfoodmarkets.com.au/bar-n-tick-plots.pdf
>>
> The 1-minute close price is a (potentially very small) sample of all
> the price changes that occur in a given minute interval.  The fact
> that they're similar is likely because the price wasn't moving much
> during the hours in your plot.  If prices were moving a lot, I
> wouldn't expect them to be very similar at all.
>
> If you look at the tick data versus the high-low range, you should see
> that all the tick prices fall within that range for every aggregate
> interval.
>
>> I am not experienced with this so I wondered if this sort of discrepancy is
>> typical or am I just making a mess of it.  If they are typical so it makes
>> setting trigger levels a bit difficult as mine are derived from bar data but
>> implemented by tick data movement.  I'm surprised the tick data doesn't fall
>> within the one minute bar so any comments would be of great interest.  Am I
>> better to only work with tick data for example?
>>
>> Stephen Choularton PhD, FIoD
>>
>> _______________________________________________
>> R-SIG-Finance at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
>> -- Subscriber-posting only. If you want to post, subscribe first.
>> -- Also note that this is not the r-help list where general R questions
>> should go.
>
>