|
|
Rank: Member
Groups: Member
Joined: 5/18/2010
Posts: 11
Points: 33
Location: Germany
|
Dear Team,
I would like to backtest die Piotroski screen in some variations.
For example a monthly instead of yearly adapation for the last 10 years.
There is a login secured link "Backtest Europe" in the headframe for developers only.
Is it possible to get the screening results by date with that?
Did you safe older screening results or do you have any data I could use
for my research?
Thankyou.
|
|
|
|
|
|
Rank: Administration
Groups: Administration
Joined: 12/31/2008
Posts: 84
Points: 28
Location: Europe
|
Dear Mike, The backtest Europe module is only for developpers in our Team, this because of the overhead this module is asking from our server. We only analyse yearly results for our backtests. If you want to backtest yourself, I think you will need to take a licence on Bloomberg or another dataprovider... Kind regards, Ik
The Value Investing Forumhttp://www.value-investing.eu
|
|
Rank: Member
Groups: Member
Joined: 5/18/2010
Posts: 11
Points: 33
Location: Germany
|
I have access to Datastream and could get one to Compustat.
What is the bloomberg tool called and which one would you prefer?
The guys of AAII (aaii.com) did backtest in monthly steps and got
quite better results. At least in the US market.
|
|
Rank: Administration
Groups: Administration
Joined: 12/31/2008
Posts: 84
Points: 28
Location: Europe
|
The results will be probably the same in Europe in theory, but we prefer a buy and hold strategy of one year for the following reasons: 1> It is easier 2> transaction costs. -> With a portfolio of 25 shares, you must therefore make only maximum 50 transactions per year. With a monthly basis strategy there will be much more transactions. 3>It is less time consuming. If you have Datastream, I think you having all you need to backtest. Kind regards, Ik
The Value Investing Forumhttp://www.value-investing.eu
|
|
Rank: Member
Groups: Member
Joined: 5/18/2010
Posts: 11
Points: 33
Location: Germany
|
I am going to start my backtest research now and it would be very good if you can tell me following things to make sure that your an my backtest are consistent: 1. Which Datastream datatypes did you use for die Piotroski-9 score ? You already said that you were using PTBV for die Book Value. I wrote down the ones that could be in bold: Quote:1.positive earnings [F_ROA] WC08326
2.positive cash-flow [F_CFO] WC05501
3.increasing ROA [F_ΔROA]
4.Quality of earnings :Compare Cash flow return on assets (2) to return on assets (1) Score 1 if CFROA>ROA, 0 if CFROA < ROA [F_ACCRUAL]
5.decreasing long-term debt as a proportion of total assets [F_ΔLEVER] WC08216
6.increasing current ratio (indicating an increasing ability to pay off short-term debt)[F_ΔLiquid] WC08106
7.decreasing or stable number of shares outstanding [EQ_OFFER] WC05301
8.increasing asset turnover (indicating an increasing sales as a proportion of total assets) [F_ΔTURN] WC08401
9.increasing gross margin WC18276 [Δ_MARGIN] 2. Where in Datastream did you get the dividends from you used in chapter C of your backtest? WC05376?
|
|
Rank: Administration
Groups: Administration
Joined: 12/31/2008
Posts: 84
Points: 28
Location: Europe
|
Hi, For more information about the datatypes, we can refer to the Datastream website : http://extranet.datastream.com/There are a lot of possible worldscope datatype possible for this analysis Kind regards, Philip ps For the dividend I think we used datatype DD
The Value Investing Forumhttp://www.value-investing.eu
|
|
|
Guest |