Trust historical data

coiler
284
coiler  

Hi all,

please, is there trust historical data out there?. I would like to know if there is good historical data to set the correct enviroment to develop EAs, if so, could you give me way to go or a good approach please.

sorry about my english,

Vince.

Mladen Rakic
225960
Mladen Rakic  

Testing data ...

I think that whatever data (in forex) you use, it will be the same. But not in the bad meaning. There is no "absolutely correct" price for any of the forex symbols since it depend on brokers and a lot of other things. So, it is not uncommon that prices differ from broker to broker by even a significant number of points.

What you should try is actually to test on as many different brokers sets of data your EA as you can. That way you can see with a certain degree of confidence, is your EA robust enough to battle the market (and different market conditions) or not. If, for example, it works OK just on one broker data and does not work OK on other brokers data, that is a cause for an alarm : it means that it is probably over fitted to that broker data and that it can easily fail even on a slightest conditions change.

In testing, testing on more sets of data is not bad. Quite the opposite ... so, if you want advice, do not look for "best data" but look for a lot of "average data" sets and test the EA an all of them.

vibarco:
Hi all,

please, is there trust historical data out there?. I would like to know if there is good historical data to set the correct enviroment to develop EAs, if so, could you give me way to go or a good approach please.

sorry about my english,

Vince.
coiler
284
coiler  

Thank mladen.

Best,

Vince.