Australian (ASX) Stock Market Forum

Best Data for me?

Joined
25 July 2008
Posts
383
Reactions
5
Trying to sort through the mountains of Data providers out there.

I know Nick Radge Recommends Norgate PremiumData for $30/month; but frankly; I subscribed to the trial and it is a little disappointing. The update program etc. looks like it was built for windows 3.1! I know looks aren't everything, but come on.

also, I mucked around with a few charting programs, and the way the data is divided into a folder for each letter isn't the handiest thing.

I had a look at Paritech Data Director. $175 for EOD data after midnight or $352 for EOD data same day plus 2 intraday snap shots!

This is cheaper than premiumdata!

Is there anything I am missing here? Any others I should check out?

Programs I will likely be using will be AmiBroker and possibly MTPredictor.

Cheers
Ryan
 
There are other alternatives.
Not recommendations, just ideas:

JustData (com.au)
Cooltrader (com.au)
datahq (com.au)
 
With Premium Data it is easy to manage custom groups, and has a mountain of custom groups already pre-defined.

Also has excellent GICs per-configured for Amibroker.
 
With Premium Data it is easy to manage custom groups, and has a mountain of custom groups already pre-defined.

Also has excellent GICs per-configured for Amibroker.

Premium Data is a great service.
 
I have used Paritech in the past, I would recommend you avoid Paritech Data and go with premium data, Amibroker works well with it and I agree with comments above the service is excellent.
 
Programs I will likely be using will be AmiBroker and possibly MTPredictor.
Cheers
Ryan

I'm interested in why people choose certain technical and statistical analysis programs. What do you wish to achieve with the program you choose?
 
What is it that makes it so great?

Paritech don't deal with Amibroker it can be imported I know I have done it, I found some data with Paritech has history missing I pretty sure AGK is one check history between Premium and Paritech see what you get.
I have never found a fault with data in Premium I have seen faults in other data providers.

With Premium download the integration scripts everything is done this wont happen with Paritech, When I first set up with Premium I had trouble importing the data I sent a message to Richard and his reply was very fast within the hour found the problem everything sorted, I've also swapped subscriptions a few times and they have been very helpful in this.

Sure the update program etc. looks like it was built for windows 3.1 but to me this does not matter I prefer good service and clean data and Premium have been excellent in this regard.
 
What is it that makes it so great?

Have found Richard Dale who runs PD is very much on the ball , very knowledgeable, very accessible and replies directly and timely to any questions. Also the data quality is excellent. While the comments about the aesthetics of the program, it may not be to everyone's taste, I believe there is a lot going on in the back-end server and client end when makes it very reliable.

Also the comments I made previously.
 
I'm interested in why people choose certain technical and statistical analysis programs. What do you wish to achieve with the program you choose?

Good question, but probably best for a different thread. I can comment about both programs ... although both are very different.
 
I am in the free Symbol, OHLC, Date and Volume camp because the data is generally correct. I mean how difficult is it to collect OHLC date and volume. For pin point accuracy then I would use a paid service.
 
I am in the free Symbol, OHLC, Date and Volume camp because the data is generally correct. I mean how difficult is it to collect OHLC date and volume. For pin point accuracy then I would use a paid service.

If using any type of indicator which is a derivative of price such as moving average, which I don't believe is that shunned upon by even people only using pure price, then stock splits etc ... will cause issues in analyzing data. Even perhaps support and resistance levels , and trend lines ?

Wysiwyg, I haven't followed your particular trading style ... long term trending systems on stocks , futures, fx ? Swing trading , mean reversion ? Options ?

I definitely find managing an existing trade with non-adjusted data very troubling. This is all T/A talk though.
 
If using any type of indicator which is a derivative of price such as moving average, which I don't believe is that shunned upon by even people only using pure price, then stock splits etc ... will cause issues in analyzing data. Even perhaps support and resistance levels , and trend lines ?
Hi Dave, the incidents of stocksplits aren't common but would certainly throw a chart out for that specific time and event.

Wysiwyg, I haven't followed your particular trading style ... long term trending systems on stocks , futures, fx ? Swing trading , mean reversion ? Options ?
You are lucky you don't :p:, On shares I was a buy and hold dummy but severe market corrections have changed that. I now trade with the trend and trail stop or sell target. Duration variable.
With FX and market Indices I did day/night trade but have to review this short term approach because I lost and have stopped.
I definitely find managing an existing trade with non-adjusted data very troubling. This is all T/A talk though.
The data I use to manage a trade is through a recognised broker so all data is true and correct every day.
 
The data I use to manage a trade is through a recognised broker so all data is true and correct every day.

Actually many of our support queries are from people that ask why our data doesn't match:
ASX web site, CommSec, CMC Global, Interactive Brokers, E*TRADE, Yahoo, Google, and other data vendors etc.

Reasons include
Such sites don't (or rarely) plot adjustments for corporate actions, use uncorrected price data (the ASX do cancel quite a few erroneous trades), use different opening/closing times (eg. 4:00pm instead of 4:11pm), inaccurate data (reporting only to cents not to tenths of cents) etc.

I'll put together a few examples for you all to check the quality of your own data to see how yours rates.

PS - Yes I did mention the ASX web site! We've seen many examples of inaccuracies on there.
 
Actually many of our support queries are from people that ask why our data doesn't match:
ASX web site, CommSec, CMC Global, Interactive Brokers, E*TRADE, Yahoo, Google, and other data vendors etc.

I do realise that Premium Data is compiled at the end of a daily market session but would you go so far as to state that broker quoted share prices during trading sessions are different?
 
Below is part of the output of tonight's data update of yesterday and today's data using Premium Data.

The output is to my backup laptop which is slower than the desktop.

Pretty impressive detail I think.
.
 

Attachments

  • Premium Data Example.png
    Premium Data Example.png
    58.8 KB · Views: 92
As promised, I've put together a few different types of things that quality data vendors do to the data to allow you to compare against what you are currently getting.

Here's a few examples of data maintenance that most (all?) free data vendors/brokers and quite a few data vendors would not have performed.

1. Code changes

Pallane Medical (changed codes from DIA to PNM on 20 November 2009)
Accurate data will show: complete history of PNM from 24 Jan 2005 through to today
Bad data will show: Short history of PNM starting 20 Nov 2009 and DIA last date showing 19 Nov 2009

2. Merging of deferred settlement periods (eg. around reconstruction time)

When a stock undergoes a reconstruction or restructure it trades as a deferred settlement security.

Recent examples include:

Astro Japan Property Group: Recently became a stapled security.
Traded as AJA until 4 Nov 2009
Traded as AJADA from 5 Nov 2009 through 18 November 2009.
Trading now as AJA from 19 November 2009 onwards.
Good data will show: Complete data history from 4 April 2005 through to today with no periods of missing trades.
Bad data will show: Missing data between 5 Nov and 18 Nov 2009 - you are missing trading history.

Nuplex (NPX) had a 4:1 consolidation exdate 11 June 2009
Good data will show: Complete data history from 24 Nov 1999 through to today
Bad data will show: Missing data between 11 June 2009 through 24 June 2009.

3. Dilutions

Many corporate actions are dilutionary to existing shareholders. However, if you participate in the corporate action (eg. a rights issue, non-renounceable issue etc.) your shareholding will not be diluted. The chart should reflect your participation in any such actions.

Recent examples include:

Lynas Corporation (LYC) has a 1:1 Non-Renounceable issue of shares @ 45C with exdate 1 Oct 2009. This resulted in a dilution factor of 0.75
Good data will show: Close price of 0.675 on 23 Sept 2009
Bad data will show: Close price of 0.9 on 23 Sept 2009

Euroz Limited (EZL) declaed a 1 for 2 share split exdate 9 Oct 2008
Good data will show: Close price of $1.21, volume of 16706 on 7 Oct 2009
Bad data will show: Close price of $2.42, volume of 33412 on 7 Oct 2009

Toll Holdings (EQF) declared a 1:1 in specie distribution (spinoff) of Virgin Blue exdate 17 July 2008 with a resultant dilution factor of 0.8967
Good data will show: Close price of $5.775 on 16 July 2008.
Bad data will show: Close price of $6.44 on 16 July 2008.


4. After market trade cancellations/corrections

The ASX regularly cancels erroneous trades. This can affect the high, low, close and volume figures. It may also affect any indexes that the stock is in too.

Recent examples of BIG erroneous trade corrections include:
Brickwords

2 June 2009 Austereo Limited (AEO)
Was: $1.74 as Close, erroneous trades cancelled - close should be $1.475

2 June 2009 Brickworks Limited (BKW)
Was: $15.31 as Close, erroneous trades cancelled - close should be $12.25

In the last year, the ASX have made:
Changes in O,H,L or C due to trade corrections/cancellations: 49 = 4/month
OHLC same, but reduced volume due to trade corrections/cancellations: 5562 = 463/month = 22/day

5. Adjustments for special dividends

The standard methodology adhered to by Standard and Poors, Dow Jones and ASX adjusts for special dividends.

Some examples:
Flexigroup (FXL) declared a 3c special dividend exdate 26 Nov 2008.
Good data will show: Close price of $0.255 on 25 Nov 2008.
Bad data will show: Close price of $0.285 on 25 Nov 2008.

Keycorp (KYC) declared a 5c special dividend exdate 25 May 2009
Good data will show: Close price of $0.44 on 22 May 2009
Bad data will show: Close price of $0.49 on 22 May 2009

6. Merging of FPOs with CDIs

Some companies in Austrlaia decide to be re-domiciled in another country. The Australian shareholders' shares are converted into trading instruments known as CDIs (Chess Depositary Interests).

Some examples include:
Heartware Intl (HIN) converted to CDIs on 4 Nov 2008. It was previously trading as HTW.
Good data will show: Trading History of HIN back to 13 January 2005.
Bad data will show: Trading history of HIN back to 4 Nov 2008 only.

News Corp (Class B Voting CDI) (NWS) redomiciled their operates in 2004 to the US.
Good data will show: Trading history before Nov 2004 (back to at least 1992).
Bad data will show: trading on NWS only starting on 4 Nov 2004.

Henderson Group Plc (HGG) converted to CDIs in October 2008.
Good data will show: trading history back to 23 Dec 2003.
Bad data will show: trading history only back to 27 Oct 2008.

7. Actual index pricing

Vendors who subscribe to the official S&P index feed will be able to provide the actual open of the index when stocks start trading at 10:00:00am. Vendors who do not subscribe to this will have the close equal to yesterday's open. This has implications for analysis of index data.

8. Index corrections

From time-to-time the ASX makes corrections are required to various indices due to trade cancellations.

For example, the low of the S&P/ASX 100 was corrected from 2535.4 to 2590. That's a 2% error.

9. Incorrect trading increments on ASX data

Particularly an issue with free data, is that they round to two decimal places. Any security traded on the ASX that is under $2.00 trades to three decimal places (either 0.005, 0.001 ticks according to its price). If data is only quoted to two decimal places the data can be dozens of percent inaccurate.

I hope that highlights a few examples of where paid data is different to free data.
 
I do realise that Premium Data is compiled at the end of a daily market session but would you go so far as to state that broker quoted share prices during trading sessions are different?

By "quoted share prices" I assume you mean either best bid or best ask, last traded price and/or market depth. These should remain the same unless the broker has any sort of technical difficulties.

Note that derivatives of the share prices may differ (eg. CFD providers may incorporate a haircut into the quoted data - their trading hours may also be different too such as not giving you the ability to participate in the opening and closing auctions).
 
Top