Seeking partner for setting up a FX High-Frequency Trading Desk
forex_trader_28881
Member Since Feb 07, 2011
724 posts
Dec 21, 2011 at 10:49
(edited Dec 21, 2011 at 10:51)
API is $60 a month. Not a train smash.
Look - I'll use them before any of the other MT brokers, but Oanda for now is king, those 1 unit trades means I can compound per tick really.
Sadly the API fee there is a bit steep, $600 a month, but I've done serious volume on them, 100 000 trades plus a day.
Look - I'll use them before any of the other MT brokers, but Oanda for now is king, those 1 unit trades means I can compound per tick really.
Sadly the API fee there is a bit steep, $600 a month, but I've done serious volume on them, 100 000 trades plus a day.
Member Since Aug 16, 2010
453 posts
Dec 21, 2011 at 10:54
But LMAX doesn't have MT4, correct? 60 a month is ok, with some traffic it's free as well.
I was trading MDP at OANDA successfully, but once I made some 800 bucks in a matter of hour I started to get rejected orders on modifications right away. Created new account, started smaller, all fine, bigger orders coming through - got blocked again. Withdrawn in the end. Makes no sense for me.
I was trading MDP at OANDA successfully, but once I made some 800 bucks in a matter of hour I started to get rejected orders on modifications right away. Created new account, started smaller, all fine, bigger orders coming through - got blocked again. Withdrawn in the end. Makes no sense for me.
Was mich nicht umbringt, macht mich stärker.
forex_trader_28881
Member Since Feb 07, 2011
724 posts
Dec 21, 2011 at 10:56
We'll - all I can say is I had no issues at all with them despite my volumes.
If you were using MT I'm not surprised, they haven't gotten that sorted there at all. I used the JAVA API.
If you were using MT I'm not surprised, they haven't gotten that sorted there at all. I used the JAVA API.
Member Since Aug 16, 2010
453 posts
Dec 21, 2011 at 10:59
Alright. Thanks for the info. Yes I used off-the-shelf MDP bot with MT4 more to test and see, but now moving onto my own algos and with API as well.
But do you refer that LMAX has MT4 or not? I haven't see this offering on their web, just some Multitrader in addition to crappy web trader.
But do you refer that LMAX has MT4 or not? I haven't see this offering on their web, just some Multitrader in addition to crappy web trader.
Was mich nicht umbringt, macht mich stärker.
forex_trader_28881
Member Since Feb 07, 2011
724 posts
Dec 21, 2011 at 11:02
(edited Dec 21, 2011 at 11:11)
They have it yes. I saw it somewhere. But MT is so slow. Every time you add to a position it generates a new trade, so if I been working a position for a year I end up with 100 000 trades or more.
To close them in MT takes about 18 hours. It's a joke.
To close them in MT takes about 18 hours. It's a joke.
Member Since Aug 16, 2010
453 posts
Dec 21, 2011 at 11:13
That the way MT4 works - it doesn't work with positions rather with orders/tickets. It's a dealing desk oriented system. I am interested in regards to manual trading though, no intention to use MT4 for HFT stuff.
Was mich nicht umbringt, macht mich stärker.
forex_trader_28881
Member Since Feb 07, 2011
724 posts
Dec 21, 2011 at 11:17
Manual trading I wrote of years ago, no quality of life sitting there staring at a screen. That's why we have computers. Can stare at the screen for me....
Member Since Aug 16, 2010
453 posts
Dec 21, 2011 at 13:55
Agree in part but I don't mind spending 2-3 hours a day on session open scalps and managing some end-of-day trades. Although I do believe algo will rule trading soon. It's quite substantial already.
I have gotten Multitrader platform and demoing it. I have to say it's amazing, best platform I have seen to date including Ninja trader and ATC Trader for Futures, not mentioning JForex from Dukas, and crappy slow metatrader. Lovely software, mega powerfull and ideal for scalping. Similar but better equipped than Ninja, awesome looking.
I have gotten Multitrader platform and demoing it. I have to say it's amazing, best platform I have seen to date including Ninja trader and ATC Trader for Futures, not mentioning JForex from Dukas, and crappy slow metatrader. Lovely software, mega powerfull and ideal for scalping. Similar but better equipped than Ninja, awesome looking.
Was mich nicht umbringt, macht mich stärker.
forex_trader_32776
Member Since Mar 28, 2011
1008 posts
Dec 22, 2011 at 19:24
TheCyclist posted:
API is $60 a month. Not a train smash.
Look - I'll use them before any of the other MT brokers, but Oanda for now is king, those 1 unit trades means I can compound per tick really.
Sadly the API fee there is a bit steep, $600 a month, but I've done serious volume on them, 100 000 trades plus a day.
API for $60 a month and with a small 1k min is surely worth checking out the orderbook, realtime quotes for testing purposes. I've never heard of lmax so after some google searches I through the info together and stuck it here - https://www.pipinvestment.com/forex/fx-wall#comment-1686
There's a good presentation from the developers.
spreads +comm. are quite low and ave exec is <20ms. I'll be checking this out for sure. Thanks for sharing.
@pipgnostic - thanks for comparing the execution speeds. For every 1 millisecond faster it seams to cost another 10k =)
forex_trader_32776
Member Since Mar 28, 2011
1008 posts
Dec 22, 2011 at 19:25
mistificator posted:
I am building up algorithmic trading business myself (www.fxalgo.net) but actually surrendered the idea of HFT along the lines of market making in favor to less demanding strategies.
One interesting is about arbitraging between forex and fx futures. Execution for both can be done via FIX.
what is your direction with your fxalgo site?
forex_trader_32776
Member Since Mar 28, 2011
1008 posts
Dec 22, 2011 at 19:30
FXBird posted:
Hi forexproptrader,
It'll be mainly spend on infrastructures (high performance trading engine, data analyzed tools, ticker plant, colo server, .. etc.), and first few months' regularly expenses. The system needed testing at least few weeks even months before trading real money.
Add: The personnel costs should be included, yet you always paid for what you got. Experienced personnel may cost more but will be worth.
can you share any links about the topics you mentioned above?
for example - https://www.exegy.com/tickerplant.htm - which costs 30k a month. how can this be done on a 10k budget?
forex_trader_32776
Member Since Mar 28, 2011
1008 posts
Dec 22, 2011 at 19:38
(edited Dec 22, 2011 at 19:39)
..
Member Since Aug 16, 2010
453 posts
Dec 22, 2011 at 22:14
Ideas sharing, development of algos, closed algo trading community, retail section for few algos for MT4 to be sold, and section for co-located trade signal providing from a portfolio of weighted-risk algos, or an option for a PAMM money management. I have a company behind it and few bright minds working already on the first product. Just need to prioritize, so much of stuff is happening same time. And I am full time trading manually too.
ForexScam posted:
what is your direction with your fxalgo site?
Was mich nicht umbringt, macht mich stärker.
Dec 23, 2011 at 07:33
(edited Dec 23, 2011 at 07:33)
Hi ForexScam,
The 10k budget was a rough estimated. The simple idea was using my per-HFT strategy to self-support after verified it in the real account. Converted the strategy to automation trading system could increased it performance at least 30% (ex. $500/h to $650/h). But in this preliminary stage it'll be highly risky! You need to monitoring the operation 24h if you want to avoiding any 'spot news'. The strategy can be implementing by Excel with API or FIX bridge/add in. Ticker plant wouldn't required by this stage but for the latter stages.
The 10k budget was a rough estimated. The simple idea was using my per-HFT strategy to self-support after verified it in the real account. Converted the strategy to automation trading system could increased it performance at least 30% (ex. $500/h to $650/h). But in this preliminary stage it'll be highly risky! You need to monitoring the operation 24h if you want to avoiding any 'spot news'. The strategy can be implementing by Excel with API or FIX bridge/add in. Ticker plant wouldn't required by this stage but for the latter stages.
forex_trader_28881
Member Since Feb 07, 2011
724 posts
Dec 23, 2011 at 08:48
We just wrote in a spread check on O. Spread goes wide we simply stop trading. Easiest way to avoid any news.
Member Since Aug 16, 2010
453 posts
Dec 23, 2011 at 09:43
But slippage would be. In case of episodic volatility ('spot news') all jump on the same bandwagon, e.g. all are buying, so if you want to buy too, there will be no one to sell to you, but quite far away. And price discovery mechanism will give you terrible slippage in many cases, unless the move is so violent that CME will apply circuit breaker and reject trades. Especially in ES, the most HFT-traded security ever, where perhaps 90% of trading is done by HFT.
Was mich nicht umbringt, macht mich stärker.
Dec 23, 2011 at 10:06
Hi MistiFx,
Although I need doing more tests for verified the strategy's performance in these violent time which you pointed out. The simplest solution could be just avoided trading at these occasion like PipGnostic suggested. I have nothing to add at present.
Although I need doing more tests for verified the strategy's performance in these violent time which you pointed out. The simplest solution could be just avoided trading at these occasion like PipGnostic suggested. I have nothing to add at present.
Member Since Aug 16, 2010
453 posts
Dec 23, 2011 at 10:17
I think the easiest way for your strategy to justify when to stop trading is to look at the order book and validate nearby liquidity let's say within 1-2 points (i.e.4-8 ticks) and compare with an average liquidity level 10,30,60 seconds ago for example. If it's dried down 2-3 times or more - smash the brakes. Way better than trying to analyse spread which is not indicative in futures whatsoever.
Was mich nicht umbringt, macht mich stärker.
Dec 23, 2011 at 16:25
(edited Dec 23, 2011 at 16:27)
MistiFx,
Very good idea! My strategy was working at low liquidity time well too but with plenty liquidity the performance will better. I just testing it at high liquidity trading time (between 9:49 to 10:00 EST), the result was quite good in my 'high latency manual trading environment': P/L +$300 within 10 min and in total 57 orders without any draw-down. Yet some kind of additional risk management modules/functions should be build into the automation trading system certainly.
Very good idea! My strategy was working at low liquidity time well too but with plenty liquidity the performance will better. I just testing it at high liquidity trading time (between 9:49 to 10:00 EST), the result was quite good in my 'high latency manual trading environment': P/L +$300 within 10 min and in total 57 orders without any draw-down. Yet some kind of additional risk management modules/functions should be build into the automation trading system certainly.
