[R-SIG-Finance] qantstrat same period execution

Ulrich Staudinger ustaudinger at gmail.com
Sun Dec 9 18:04:38 CET 2012


hi there.

long story short, the only right thing is to allow execution at the
close of the current period if the current period is not the final
period in a day.

the really dirty aspects are timestamps- with which timestamp do you
want to tag an execution, for sure not with the beginning timestamp of
a priod, as it is misleading, but so is execution at the close price
of the next period (it is reported with the beginning timestamp of a
period).

clearly, data snooping and look ahead bias are to be ruled impossible,
therefore no execution at current bar's open price.

again. the key argument is that the open price of any bar is not
necessarily the actual market quote at the beginning of the bar.
technically and usually, it is the first new price in a period.

again, we have to have a look at how open/high/low/close bars are
built. this is data source specific. mind that most ohlc data sets are
built from actual trades and not quote midpoints, which makes things
even more dirty, as there might be only a few trades a day, but quotes
all daylong, particularly in options markets.

as i see the technical challenges and the work related to permitting
execution at current bars close, i came up with the work around to
overwrite the open values of the next bar with the former periods
close, thus dealing at the same price. obviously, this is a problem
for certain scenarios and obviously it works for others. i do not
think that the instant-execution-myth argument is valid, as no one
assumes instant execution.

we discussed this indeed, and i think it makes sense to add the
non-default behaviour to execute at current periods close.

best regards,
ulrich





--
Ulrich Staudinger
ActiveQuant GmbH

sent from my tablet.

On 09.12.2012, at 15:52, af2tr <dominykasgrigonis at gmail.com> wrote:

> <http://r.789695.n4.nabble.com/file/n4652604/Screen_Shot_2012-12-09_at_14.42.29.png>
>
> I agree it is not realistic for higher frequencies that daily, but on daily
> data it sometimes is. Lets imagine we go short after downside cross.(vice
> versa in image) you wake up and see a huge gap like this. You don't rlly
> want to wait for the next day. Also in other cases you just wait until 4:45
> and call your broker. Market price you get will be pretty close to same
> day's close...
>
> I'm not intending to persuade you to change it, just making a point. :)
>
>
>
> --
> View this message in context: http://r.789695.n4.nabble.com/qantstrat-same-period-execution-tp4652598p4652604.html
> Sent from the Rmetrics mailing list archive at Nabble.com.
>
> _______________________________________________
> 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.



More information about the R-SIG-Finance mailing list