Hi, I know some here are day trading using the second time frame. If you one of those, would you share us with: 1- what platform you are using? 2-manually or using automated system? 3-how to decrease transaction costs? Thanks
I trade 5 second time frame, automated with Interactive Brokers. The biggest difference between 5 second and 1 second time frame in my experience is the effectiveness of market orders vs limit orders and the difficulty of simulating what happens at 1 second intervals due to latency and execution time. 1 second time frame doesn't have a huge advantage over 5 second time frame unless you've got accurate level 2 depth historical data and the capacity to reliably simulate limit orders. Traders operating at sub 5 second time frame "make odds" on the order book, almost exclusively by posting and cancelling limit orders, thus taking on the role of liquidity provider. They reduce their transaction costs by arranging a "rebate" commission structure with their broker: the higher their volume, the smaller their commission bill is (it can even go negative, where the broker pays them). This is the modern equivalent to owning a seat on the exchange, which entitled the trader to operate on the floor without paying commission. My conclusion is that since I don't want to play that low-margin, high overhead game, I may as well take my directional trades at 5 seconds using market orders instead of taking on the trading desks at <= 1 second but blind without the order book and at the whims of my VMWare host's ISP.
Software is all custom made in Java. Hardware is a VMWare virtual server rented from a high availability data center. It is damn fast hardware and internet connection.
Since 1 second time frame is not suitable for your system, may i ask why you did not go to 2, 3, or 4 second time frames instead of 5? Also, may i ask if one wants a customizable second time frame so he or she can tailor the length as he wishes from 1 up , what soft ware you recommend? Thanks and sorry for many questions
It is a function of my average holding time. There is no significant price movement at 5 second time frame relative to my holding duration, but there was at 10 second time frame, so 10 wasn't granular enough. Anything less than 5 seconds and I'm wasting resources trying to store the data and simulate trading with the stored data. This gets exponentially harder the closer you get to pure real-time. My software supports trading based on pure real time bid/ask price changes, but I don't want the overhead. Ninja Trader or Meta Trader can probably do everything you need, although I do not use them except for testing out quick ideas. If you haven't looked at them you should.
Nice, you time your holding duration relative to timeframe movement. interesting approach, thank you for sharing. So, the question is , what if your holding reached the limit you have exit at, do you exit in loss regardless? or what is your MM to deal with such situation? NT is OK but Metatrader is sucks in trading. MT cannot go under 1 minute time frame mq4 is a limiation too. i am using MT for development and quick testing because it is very convenient. But for real trading, i will go with another platform.. My choice till now, NT, MC, Amibroker but didnot decide yet.. Also, i have a question, i am thinking to link my third party software with ECN, Do you think it would be better to have an external RT data or using the same data of my ECN? Thanks
If you one of those, would you share us with: 1- what platform you are using? Cunni Lin Gus 2-manually or using automated system? Manually, it would be totally assinine, hovine, bovine to automate it. 3-how to decrease transaction costs? slide in the backdoor
I have a buddy who trades natural gas using Guppy MAs and Zero lag Stochs on a 5 second chart and is minting money.