You've mentioned that before. Buying financial data from Ebay sounds insane. There's all sorts of issues with that recommendation...just for one, what kind of data is it? OHLC? 1m? 1s? Tick data?
Real historical tick data costs five to six figures per year from a well-regarded source. There's a price signaling issue here - if you have legitimately identified alpha, the cost of the data is not unreasonable (just as a colocated server is not unreasonable expensive if you can capitalize on it). If someone has high quality data, why would they sell it on Ebay for that price? It signals several things:
1) They don't care about selling where their customers are,
2) They don't care about leaving (ridiculous amounts of) money on the table,
3) They are probably not going to provide any verification or due diligence for the data, let alone any support for it afterwards.
What are these people doing, scraping Yahoo Finance, throwing it in a CSV and selling it? I find it exceptionally difficult to believe that "with a little Python" you'll be testing anything close to "any strategy imaginable" using this data. Vendors that sell real data do not share many of the characteristics of a fly by night operation.
I am talking about daily data for long term trading/investing, obviously if you plan to run high frequency or day trading operation and need one minute or tick data, neither eBay nor python will help you :-)
Real historical tick data costs five to six figures per year from a well-regarded source. There's a price signaling issue here - if you have legitimately identified alpha, the cost of the data is not unreasonable (just as a colocated server is not unreasonable expensive if you can capitalize on it). If someone has high quality data, why would they sell it on Ebay for that price? It signals several things:
1) They don't care about selling where their customers are,
2) They don't care about leaving (ridiculous amounts of) money on the table,
3) They are probably not going to provide any verification or due diligence for the data, let alone any support for it afterwards.
What are these people doing, scraping Yahoo Finance, throwing it in a CSV and selling it? I find it exceptionally difficult to believe that "with a little Python" you'll be testing anything close to "any strategy imaginable" using this data. Vendors that sell real data do not share many of the characteristics of a fly by night operation.
reply