You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Infinity AT
- Thread starter Ares
- Start date
this is the thread i meant to post to.
my point is that comparing AT to TT with ninja isn't apples to apples as its not unfiltered tick by tick data. TT on ninja runs on the FIX protocol and is coalesced. xtrader, like AT does not coalesce.
i too like the AT DOM and the broker but experience lagging data. my assumption is that the nature of the unfiltered data stream will perform much better on a separate chart-less execution-only computer.
my point is that comparing AT to TT with ninja isn't apples to apples as its not unfiltered tick by tick data. TT on ninja runs on the FIX protocol and is coalesced. xtrader, like AT does not coalesce.
i too like the AT DOM and the broker but experience lagging data. my assumption is that the nature of the unfiltered data stream will perform much better on a separate chart-less execution-only computer.
dameh. who said anything about charting or PCs?
AT chockes itself. They have to rewrite it. They have good potential and good data, however software cannot cope.
Is nothing to do with us users or our computers.
If you put it on a supercomputer it will behave the same way. I can bet money on it.
AT chockes itself. They have to rewrite it. They have good potential and good data, however software cannot cope.
Is nothing to do with us users or our computers.
If you put it on a supercomputer it will behave the same way. I can bet money on it.
Quote from dam5h:
this is the thread i meant to post to.
my point is that comparing AT to TT with ninja isn't apples to apples as its not unfiltered tick by tick data. TT on ninja runs on the FIX protocol and is coalesced. xtrader, like AT does not coalesce.
i too like the AT DOM and the broker but experience lagging data. my assumption is that the nature of the unfiltered data stream will perform much better on a separate chart-less execution-only computer.
I think you may have a good point...but I'm wondering:
In the end, what practical purpose do these ultra fast unfiltered tick feeds serve (especially to discretionary point and click in the DOM traders), if they freeze/hang/bottleneck the software/PC?
Will someone who is even scalping for 5-10 ticks, but manually entering orders this way even notice a useable difference between TT, AT, and ZF in price updates, fills, and execution?
Legend, make an effort and read the subject of the thread before coining your 2p.
This is not about comparsion of platforms. AT has serious issues when DOM freezes for 10-15 sec. And the OP was unfairly bullied by some.
There are plenty of threads about Ninja etc. Please keep it clean.
P.S. having seen numbers of threads about Ninja and Zen - it seems as the users are trying to convince themselves that they are buying a good value for cheap... Yeah.. Good luck.
This is not about comparsion of platforms. AT has serious issues when DOM freezes for 10-15 sec. And the OP was unfairly bullied by some.
There are plenty of threads about Ninja etc. Please keep it clean.
P.S. having seen numbers of threads about Ninja and Zen - it seems as the users are trying to convince themselves that they are buying a good value for cheap... Yeah.. Good luck.
Quote from legend4life:
I think you may have a good point...but I'm wondering:
In the end, what practical purpose do these ultra fast unfiltered tick feeds serve (especially to discretionary point and click in the DOM traders), if they freeze/hang/bottleneck the software/PC?
Will someone who is even scalping for 5-10 ticks, but manually entering orders this way even notice a useable difference between TT, AT, and ZF in price updates, fills, and execution?
Did you see Ninja freezing in the video comparision I uploaded?
No.
The problem is Infinity AT (Transact) software
Period.
personally i have worked with transact on the problem and they sent me a new version of the software that dealt with long freezes, they had found a bug (this was in november). the newer version helped and i never have (and never did for that matter) seen anything like what ares has experienced
while improved the price can still lag a bit (i know from experiences with the flatten button) during high volume price surges, its less than a second but it is still a lag. i have a lot of charts (that read the same tick feed) on this machine and a few other processes, their position on the matter is i will see much less data lag on a fast execution only machine. i compared price changes beside a hard wired laptop (that I borrowed) running only the DOM and noticed a difference in this regard. i am building out a new desktop for this purpose, hopefully the parts will be in this week and i can update my experiences.
legend here is my best shot at addressing your questions:
most of the difference on execution will depend on the backend not the frontend. this will depend on how loaded the broker's servers are to rated performance/capacity and where they are located relative to the exchange. the best case is prob 50% max load and worst is well over 100%
the front end is important in disseminating price data to you, you have zero chance of initiating execution properly if what you are seeing is not reality.
as far as coalescing goes, i am not completely clear on whether or not price updates with each change or not, if the platform only updates price 10 times per second (100ms intervals) and you get a touch and immediate rejection of price (where only one trade took place there for a few miliseconds). if the answer is that only the volume would show price there and not the DOM ladder on a coalesced feed then i would say therein lies one advantage of an unfiltered feed. as seeing every price change is more important than every single trade at the same price, imho. i intend to do more research on this matter. please chime in if you know.
while improved the price can still lag a bit (i know from experiences with the flatten button) during high volume price surges, its less than a second but it is still a lag. i have a lot of charts (that read the same tick feed) on this machine and a few other processes, their position on the matter is i will see much less data lag on a fast execution only machine. i compared price changes beside a hard wired laptop (that I borrowed) running only the DOM and noticed a difference in this regard. i am building out a new desktop for this purpose, hopefully the parts will be in this week and i can update my experiences.
legend here is my best shot at addressing your questions:
most of the difference on execution will depend on the backend not the frontend. this will depend on how loaded the broker's servers are to rated performance/capacity and where they are located relative to the exchange. the best case is prob 50% max load and worst is well over 100%
the front end is important in disseminating price data to you, you have zero chance of initiating execution properly if what you are seeing is not reality.
as far as coalescing goes, i am not completely clear on whether or not price updates with each change or not, if the platform only updates price 10 times per second (100ms intervals) and you get a touch and immediate rejection of price (where only one trade took place there for a few miliseconds). if the answer is that only the volume would show price there and not the DOM ladder on a coalesced feed then i would say therein lies one advantage of an unfiltered feed. as seeing every price change is more important than every single trade at the same price, imho. i intend to do more research on this matter. please chime in if you know.