Just a quick question, which Finance database is better for finding historical open, high, low, close, etc. data on stocks for backtesting? I've used Yahoo Finance and occasionally, days are missing which trips up my backtesting program..
i have been doing some back testing myself. I would say yahoo is easily better for this. If you have a program, Matlab, which is used a lot in engineering practices, there is a built in historical data function block that imports the data from yahoo. Makes it a whole lot simpler. otherwise i would go with some sort of php type function using fopen, etc. with yahoo's finance to aggregate all the data for it. i believe it comes in the form of chart.yahoo.com?s=IBM or something along that. Ironically, if you search for it on google there will be a list of like 60 different data pieces for each stock each day to look at if you wanted
I use Yahoo. They've been around for longer and their servers are used to these kinds of requests. I don't think Google has the extensive database and diversity of data that Yahoo has. But if you're looking for just simple price data, might give Google a shot.
Hi, I am a programmer. I programmed a stock download using esignal. I also programmed one using Yahoo. I am "watching" well over 20,000 stocks programatically. I have not tried google, but I can tell you the things eSignal (which is pricey) did for me that Yahoo could not: 1. The historical prices are rounded the nearest cent in Yahoo to. As I am looking for various patterns, this is rather disappointing. Of the 13,000.00 OTC stocks I am watching, many of them are very cheap, making fractions matter. It is not much of an issue for the normal exchange stocks though. 2. eSignal does not require any symbol extentions. If I want to find AAUKY.PK, Yahoo has to know it is Pink Sheets, whereas esignal will figure it out if I merely enter AAUKY. This makes eSignal find the greatest number of stocks with the least amount of programming. Since I import these lists from the exchanges and other sources, I do not want to have to manually review everything I imported and correct the extentions. Yahoo has the following advantages: 1. It works seamlessly with multi-threaded requests and so is fast as lightening when programmed that way. This makes Yahoo seem faster than eSignal, as it gets the results faster, even though eSignal is very very fast, much faster than Yahoo if you program it sycnronously. 2. It's free. You can get around the symbol extention problem by doing extra programming. Have your program look up the symbol and if it finds no match, then look up Symbol.OB, then Symbol.PK, etc. and save the first positive match, which the user may then edit. If anyone has tried goodle or knows of other alternatives, please let me know how they matched up. I am very interested. Thanks.
I want to make sure that I don't bake in survivorship bias on my backtests. Can we query eSignal with something like "give me quotes for the 100 largest companies by market cap on 1/1/1980" have it make sure to include companies that have since gone out of business?
I'm using Yahoo for daily and DTN IQFeed for intraday streaming and historical. I could just as well use IQFeed for daily but haven't got round to it yet. IQFeed is good and reliable. It also works on Linix under wine (100% reliable and 64 bit no problem)The API is easy, though perhaps not the most elegant. Cheaper than eSignal. Recommended. As to another poster's question about survivorship bias, I think all the retail stuff suffers from it. It's a disgrace really. If anybody knows any different, then please tell.
How about this: Where can I get a historical list of Russel 3000 member stocks? That's a big enough list to limit myself to. Example, I want to backtest to 10 years ago. What were the companies on the R3k in 1999? Also, can eSignal / IQFeed give me data for delisted companies?
Only drawback from yahoo for me right now is how slow the data gets downloaded from their servers. So if you're looking at a large list of stocks/instruments, the download to update your database is gonna take a long time. The speeds rarely go above 50-100KB/s or in Mbit terms that is rarely above 0.50 Mbits/s. Way too long. If Google is faster I might give it a shot.
If your software can open multiple concurrent TCP connections (eg by multithreading), it can be quite quick. My code manages to get > 5K stocks in a few minutes. So far Yahoo hasn't objected to this for historical data download. Don't try too many connections on their fundamentals pages, though, or they will lock you out for a period.