I am happy to share my views on this data. Firstly it is quite inexpensive. You do not have to pay for all the markets, you can buy the data for certain groups, but I pay for all of them and still think the price is reasonable.
I download the data rather than get it on CD. As you have pointed out, the data goes back years, which is very useful for back-testing. I cannot comment on the quality or accuracy, but if you are using it for research then I would suggest it is good enough.
There are other things I like about this company:
A) They appear to be a one-man setup, and whenever I have emailed them with a query, I always get a prompt and courteous reply.
B) The data is available in several formats, including Ninjatrader. So if this is what you are using, you have no conversion to do. Also, you can download different timeframes. I forget the full range, but it may be something like 1, 15, and 60 minutes. This is in addition to the tick data, but beware that the timestamp is only the hour and minute of each tick.
C) As well as the intraday data, you also get the daily data.
D) They offer data for a wide variety of markets - forex, indicies, some commodities. As I say, I just download everything.
E) Finally, as part of your purchase you are entitled to 4 quarterly updates to the data. You do not get an email telling you that they are available, so you need to make a diary note to go and look at the appropriate time. Updates are usually available a couple of weeks after the quarter end. In addition, for a few dollars, you can pay for a further set of 4 updates, so you can keep your data up-to-date on an ongoing basis (which I do).
Hope this helps.
The following 2 users say Thank You to swandro for this post:
I also think the data is really affordable and getting the data for all the markets is reasonable. You don't know how accurate the data is? What is the point of backtesting if the data is not accurate? I don't understand why the timestamp of the TICK data is to the minute and not the second or millisecond. When importing the data into NT, does NT automagically sequence the ticks and "know" which came in which order?
I personally do not worry about the accuracy of data when I do backtesting. The reason is that the whole process of backtesting is vague anyway. I use backtesting to give me some statistical evidence that a method or strategy may work successfully in the future. I also use backtesting to help me identify market events which I may not have planned for when designing a strategy. If a tick is off by a pip or two, or if a few ticks are missing, it is not going to affect the outcome of my testing.
As for how NT handles tick data, I am only an occasional user of NT so I cannot say. As I said, you can actually download the data in NT format, so NT must be happy to work with it.
In attacment samples of their 1Min and tick EUR/USD 2008.01.06 - 2008.02.01 in NinjaTrader ASCII format.
Before to buy I compared EUR/USD 1Min 2008.01.06 - 2008.02.01 with Oanda's data.
Differences between disktarding quotes and ohers I checked (including 60 min candles from MetaQuotes server) was minimal.
1) minute-on-minute differences with Oanda OHLC (separately sorted in EXEL) in tenth pips: 400 means 40 pips, -100 means -10 pips.
2) Oanda's 1Min chart (top) and disktrading 1Min chart (bottom - I loaded their quotes as GBPNZD).
I sent a request to buy othes forex pairs they have.
The following user says Thank You to fx.practic for this post: