The date range is 8/1998 through 7/2001 on about 300 CDROMs This data contains every trade and quote for every issue traded on the NYSE, NASDAQ and AMEX between 8/98 and 7/01. There is no survivorship bias. basic info: http://www.nyse.com/marketinfo/1022221393337.html retail pricing: http://www.nysedata.com/info/productDetail.asp?dpbid=36&dptid=41&page=3 According to the NYSE, 36 months will cost between $7200 (marginal cost if you already were buying 60 months) and $24,000 (single 36 month purchase). Any interest?
I'm actually looking for the most recent 3-6 months of data on dvd. Do you have this or do you know who might want to sell? I'd prefer to buy the data 'used' rather than pay NYSE's prices. -- Joe
Sorry, no I don't have any. I don't know where else to get it except from the NYSE. If you only need a few symbols, there are much cheaper sources. If you need a lot of symbols and only basic T&S (time, price, size), you might get it from Realtick (via real time or EOD-batch downloads). Just be careful not to hog RT's servers, which RT might not tolerate for too long. The advantage of NYSE TAQ is you get at least 20K symbols with full T&S and bid/ask quote information, CUSIP and symbol lists, plus dividend and split information.
â¦bringing this thread up to the top. Ok, I see no takers yet. Here is what I have available right now: 1) The 36 month NYSE TAQ data described in this thread's first post, on the original CDROMs. 2) 23 months of tick data for 8000+ stocks (8/01 to 6/03) in a different format (not TAQ). This is also survivorship-bias free. Combined with the TAQ data this is 59 continuous months of tick data, and represents about 200 GB of data. 3) C and Matlab code to process (1) and (2) into a database that can be queried from Matlab, indexed by date, CUSIP, symbol and channel#, and free of survivorship biases. You define the universe of stocks you want to analyze. The database provides about 100 different statistics (channels) covering price, price changes, volume, money flow and other fundamentals. Each statistic is accumulated versus trade size and tick direction/magnitude. There are very different dynamics for different trade sizes and price move sizes. These are great inputs to Neural Nets and Bayesian networks. You can modify the code to generate your own statistics or indicators too. Overall this is an excellent foundation to explore and design systems. Build either a forecasting system, trading system, stock-picking system, or an all-in-one as I did. The Matlab codes can easily be ported to C if desired. Why am I selling? Because this is MUCH more valuable to those with time and computational resources willing to devote to it. I will eventually return to it myself. Right now Iâm busy with other systems and could use some extra money. Iâm not interested in trading for data, code or other materials. Thanks.
I'm interested in the data - cdroms and/or the later tick data. But not at prices in the ballpark of those required by the NYSE. I also am not interested in the code - just the tick data. You can email me by clicking on my account name to continue this discussion...
After a long hiatus from trading system work, I am getting back into the game. PM me if you are willing to purchase or trade other data for this NYSE TAQ set.