tradier API

thanks, that makes it a lot clearer. It would be very helpful to include all this information on your website because I at least find the information on the website greatly lacking in this regards. How to code against the API is explained but the big picture is entirely missing.

JCBarry said:
Hi volpunter,
Thanks for your questions.

a. We handle quote and trade data from NYSE, Nasdaq and OPRA based securities. The data points are documented for quotes here: https://developer.tradier.com/documentation/markets/get-quotes and for chains here: https://developer.tradier.com/documentation/markets/get-options-chains. We do deliver the entire OPRA feed through our REST (and streaming) APIs.

b. On average, multi-symbol quote requests yield 60ms responses. Option chain requests returning entire chains for an expiration average 90ms. We're working on a better way of continuously publishing our performance metrics for the APIs. Timestamps are provided by our data provider and do go down to the millisecond.

c. Let me know which examples you'd like to see. My guess is you're looking for .NET. We'll be open-sourcing a .NET SDK in the next couple weeks. If you'd like early access fire us an email and we'll send it your way.

d. We don't currently publish any symbol limits to our streaming APIs. However we do actively monitor for reasonable use cases to protect our systems and maintain compliance with the exchanges.

e. We don't currently offer data services separate from a brokerage account (but we're working on instant access to delayed and historical data). We don't mention pricing for our APIs because they are free! :) API access is offered with a Tradier Brokerage account and we currently don't charge for use of the APIs.

I hope this helps and apologize if I wasn't able to deliver the amount of detail you're after. However, if you're still curious about our APIs, don't hesitate to send an email to techsupport@tradier.com. And, if you've got specific brokerage related questions, you can email service@tradierbrokerage.com or call 980-272-3880.
More...
 
This is good feedback. We're actively re-working some parts of our web sites and we'll make sure to clarify some of the details above.
Thanks!

volpunter said:
thanks, that makes it a lot clearer. It would be very helpful to include all this information on your website because I at least find the information on the website greatly lacking in this regards. How to code against the API is explained but the big picture is entirely missing.
More...
 
JCBarry said:
Hi volpunter,
b. On average, multi-symbol quote requests yield 60ms responses. Option chain requests returning entire chains for an expiration average 90ms. We're working on a better way of continuously publishing our performance metrics for the APIs. Timestamps are provided by our data provider and do go down to the millisecond.
More...

Thanks for the answers. What kind of latency does the streaming data feed deliver?
 
That's a lot trickier to measure, because so much depends on where you're connecting from, quality of your ISP and lines, etc. That being said, I'll see if there is something I can dig up.
 
JCBarry said:
That's a lot trickier to measure, because so much depends on where you're connecting from, quality of your ISP and lines, etc. That being said, I'll see if there is something I can dig up.
More...

Thanks. Are your data servers in NY4?
 
jtrader33 said:
Thanks. Are your data servers in NY4?
More...

We worked closely with Amazon to host as much as possible using high availability/performance instances in the cloud. That being said, as executions go, we still needed to establish direct connectivity to market makers. We accomplished that with the help of folks at AWS and the major data centers in the NY/NJ area (including NY4).

For market data, we consume a consolidated feed from our data provider who houses their plants in Amazon as well. We tried to measure a bit for you. Comparing NTP and our streaming data, we were able to compare local time and the timestamps of bid/ask prices as marked by the market centers.

Here is the script and results: http://pastie.org/9319852
We ran this from a laptop connected through wifi in Charlotte, NC.

Note: We're aware this code can be significantly optimized (Ruby?), but we wanted to put something quick together to get some sample data.
 
volpunter said:
why does THIS friggin matter? You are submitting orders through a REST Api, it really does not matter whether their servers are located in NYC, Omaha Nebraska or Tokyo, Japan.
More...

Ha, dude I love how angry you are 100% of the time. For some reason it makes me happy. Thanks.
 
JCBarry said:
We worked closely with Amazon to host as much as possible using high availability/performance instances in the cloud. That being said, as executions go, we still needed to establish direct connectivity to market makers. We accomplished that with the help of folks at AWS and the major data centers in the NY/NJ area (including NY4).

For market data, we consume a consolidated feed from our data provider who houses their plants in Amazon as well. We tried to measure a bit for you. Comparing NTP and our streaming data, we were able to compare local time and the timestamps of bid/ask prices as marked by the market centers.

Here is the script and results: http://pastie.org/9319852
We ran this from a laptop connected through wifi in Charlotte, NC.

Note: We're aware this code can be significantly optimized (Ruby?), but we wanted to put something quick together to get some sample data.
More...


Thanks for doing that, appreciate it. I'm assuming your provider is NxCore. Are they providing timestamps to the ms yet? If so, any chance you could rerun the script with ms granularity?
 
Hey jtrader33,
I was mistaken earlier regarding the timestamps -- the precision of the stamp goes to milliseconds but the actual stamp only goes to decisecond (1/10th of a second). You can check it out here: http://pastie.org/9320294 (Same deal as before: wifi, Charlotte, etc.)

Hope this helps.
Thanks!
 
You must log in or register to reply here.
Back
Top