For future reference this is the crude approach I took. What I was trying to 
determine is the impact of earnings announcements on my system. My system is a 
2-8 day swing and earnings announcements do cause increased volatility. The 
question is what is the impact to returns over a timeframe.

This is what I did,

1. I ran my system over the timeframe and filtered only the top and bottom 10% 
of trades based on returns. 

2. For each ticker in above list I grabbed the earning release data from 
earnings.com (free)

3. I copied and reformatted the data in a symbol import file with the ticker 
name as symbol_ERN. Set the close price to 1. The import file can have 
different symbols so I have one file.

4. Import file into AB

5. Modified my code to not trade if three is a earning announcement a few days 
before and after my signal

5a. Load earnings data if it exists, Don't fixup the data.
E = Foreign(Name()+"_ERN", "C", 0);

5b. Add the earnings date check to my BUY statement. Since I didn't fixup the 
data I need to check if it is null. This example looks two days back and 8 days 
ahead.
AND (Nz(Ref(E, -2)) + Nz(Ref(E, -1)) + Nz(E) + Nz(Ref(E, 1)) + Nz(Ref(E, 2)) + 
Nz(Ref(E, 3)) + Nz(Ref(E, 4)) + Nz(Ref(E, 5)) + Nz(Ref(E, 6)) + Nz(Ref(E, 7)) + 
Nz(Ref(E, 8))) == 0;

My results were not conclusive and I've added bringing in the full historic 
announcement dates for a more detailed analysis as a medium priority item for 
the future.

Ron
--- In [email protected], "Ron" <rond...@...> wrote:
>
> Anyone have a tip on how to get historical earning announcement dates into AB 
> for use in the backtester? I'd like to exclude trading stocks from my 
> backtest where an earnings announcement is a few days +- from the trade 
> signal. 
> 
> Thanks,
> 
> Ron
>


Reply via email to