Fundamental testing

[It's Snake Pliskin]

Hello,

I want to do some fundamental testing using historical data.

What data providers are available for data such as ratios etc.

Are there any sites that may be readable to learn some stuff from?

 

[Sir Burr]

http://www.ascii-data.com/index.html

 

[It's Snake Pliskin]

Thanks for the link SB.

 

[prawn_86]

Snake,

Just be a bit cautious of that data because its isnt updated heaps often as far as i can tell.

I use it just to run quick scans, then drill down myself into companies that look interesting, and occasionly there is a vast difference between the site and the actual co.

EDIT - also note that it takes placements/share issues as EPS figures, so be cautious of those too

 

[reece55]

I would concur with Prawn here.....

I never use these types of providers to perform fundamental analysis, because in my opinion they are regularly wrong with the data/information. So, the reliability of backtesting using this kind of data IMO is not good enough....

Add to the fact that fundamental measurements often need to be adjusted on the basis of the accounting policies of the entity relative to others. It's pretty difficult to backtest fundies IMO.....

Just my though...

Cheers

 

[happyjack]

Huntley's or morningstar have it but I don't think you could scan it on to a spreadsheat and the huntley disk is PDF format the good news is that there is 10 years on each company one company to a page

 

[It's Snake Pliskin]

I would concur with Prawn here.....

I never use these types of providers to perform fundamental analysis, because in my opinion they are regularly wrong with the data/information. So, the reliability of backtesting using this kind of data IMO is not good enough....

Add to the fact that fundamental measurements often need to be adjusted on the basis of the accounting policies of the entity relative to others. It's pretty difficult to backtest fundies IMO.....

Just my though...

Cheers

Thanks for the replies Prawn, Reece and Hapyjack.

All points noted.

 

[Kauri]

Is this the sort of thingiemijig you are looking for Snake?

Cheers
.............Kauri

 

[stevo]

Just watch out if you are doing fundamental testing that the dates are correct. For example if info comes out every 6 months the info might be attributed to 30 June, but it didn't appear until 2 months after this date.

I was also concerned that it's possible that data could also be "adjusted" after being released to the market if something new came to light or someone made a mistake. I would be testing on the corrected data not the originally released data. I don't know if this really is a problem - but not knowing or trusting something is enough to make it difficult to test.

I did some testing some time back using Amibroker and fundamental but data and the time lag mentioned above made me doubt the results - not that I found anything significant. It's not something that I pursued.

stevo

 

[It's Snake Pliskin]

Is this the sort of thingiemijig you are looking for Snake?

Cheers
.............Kauri

Thanks for that Kauri. It may be of use but I am looking at testing assumptions not actual analysis. For example; high P/E stocks give a poorer return statistically than lower ones (with in reason).

Thanks again.

 

[It's Snake Pliskin]

Just watch out if you are doing fundamental testing that the dates are correct. For example if info comes out every 6 months the info might be attributed to 30 June, but it didn't appear until 2 months after this date.

I was also concerned that it's possible that data could also be "adjusted" after being released to the market if something new came to light or someone made a mistake. I would be testing on the corrected data not the originally released data. I don't know if this really is a problem - but not knowing or trusting something is enough to make it difficult to test.

I did some testing some time back using Amibroker and fundamental but data and the time lag mentioned above made me doubt the results - not that I found anything significant. It's not something that I pursued.

stevo

Thanks Stevo for bringing up the date and time lag issue. I am focussing on some of the less obscure information which will hopefully help.

Cheers...

 

[motorway]

http://www.hubb.com.au/sherpa/default.asp

I have never tried it ..

motorway

 

[It's Snake Pliskin]

http://www.hubb.com.au/sherpa/default.asp

I have never tried it ..

motorway

Thanks for the link motorway.