Hello Friends,
Was reading this
Algoji article and wanted to ask your opinion on
choosing the right datafeed vendor for AmiBroker.
Below is an excerpt from the above article:
------------------------------------------------------------------------------------------------------------------------------------------------------------------
Sources of Real Time Datafeed
NSE RTD is available usually through three channels- either by DotEx (an NSE subsidiary), directly from NSE via TAP server or at co-located server rack. There are two primary source for RTD- Data vendors and Brokers.
Data Vendors receive data from DotEx through leased lines, may store it on their own server, then normalize/filter and pass the data to client machine. Some of the authorized data vendors are Globaldatafeeds, Reliable, Spider, Icharts, Tickerplant etc.
Brokers receive data from NSE server via TAP server which is located in their premises. TAP facilitates IT Infrastructure consolidation and routes the orders and trades between Client and NSE Server in an optimized protocol.
Sources of Tick Data
Sorry to disappoint, but the above two sources do not provide true tick-by-tick (TBT) data. TBT data includes each and every L.T.P. (traded price) and is available only at NSE co-lo servers. The TBT data is not available at TAP Server or through DotEx for further broadcast.
The broker’s TAP servers receive filtered data, typically one tick every millisecond for Nifty Futures. The broker’s server further filters the data to reduce bandwidth requirements and then passes to client machine.
Comparatively, the data received by data vendor has still less ticks then TAP server. Similar to the broker’s server, the data vendor may also filter it again before passing down to clients.
The quality of data will depend upon 1) the Broker’s or Vendor’s server, and 2)Your own internet connection. Generally a broker may be able to provide better datafeed than a data vendor, but there is no way to ascertain this other than testing on your own computer. Both sources apply a data filtration algo to reduce the size of data for transmission over internet. Because of this filtration, different sources may broadcast slightly different ticks and the live ticks may also lag in at different time of the day. Neither your broker, nor your vendor, is likely to be generous in disclosing their server infrastructure and filtration technique.
Examples of data filtration
- Broadcasting a tick only when it is different from previous tick
- Broadcasting max n number of ticks per second
- Broadcasting max n per second number of elements inclusive all data arrays like bid, bid size, %Change, change etc.
- Retaining redundant data like vwap, total volume
- Testing the download speed to client and adjusting broadcast data accordingly
------------------------------------------------------------------------------------------------------------------------------------------------------------------
I have noticed considerable difference between NEST/NOW DataTable/VWAP backfill/rtd and Truedata/GFDL datafeed. Also in my three days trial with GFDL, I found that data appeared to be less noisy (almost similar to TrueData but slightly different). Don't know whether that is happening because of missing ticks or because of real proper datafeed?
Requesting seniors to please shed some light.
Is GFDL data quotes best and absolute?
GFDL is hell lot expensive for a chotu trader like me.
Or TrueData can still be trusted despite their recent disruptions?
Or any other hidden gem?
Thank you
