There are 1-2 market taking shops that dominate. There are many market making shops. Market making is not a winner take all environment because where you are in the queue determines when you get fill information which will sometimes put you ahead or behind other people. Also, the strategy is more vague. It can be defined as "sit in the market but try not to take toxic or informed flow". Many different trading profiles can accomplish this. When you are market taking the strategy is more "cross when the market is about to move". Lots of moves are initiated with rather simple and easy to understand triggers. The problem is that with everyone doing the same thing at exactly the same time there's more of a winner take all phenomenon.
A low latency E-Mini strategy
- Thread starter Rum Runner
- Start date
Let's say you successfully develop this market making system with SAR (stop and reverse), if I can summarize it as such. But how are you going set your target and stop levels on those SAR positions? The SAR component makes your system trend following. I suspect that will give you a lot of false signals with a few good signals that will generate enough profit.
Also, how do you define high volume "spikes"? One trade transaction with enormous size or the sum of all volume in an interval like 1 minute?
Also, how do you define high volume "spikes"? One trade transaction with enormous size or the sum of all volume in an interval like 1 minute?
If it is SAR, never any stops, you reversing into other direction? Extensive back testing only way to find what 90% of winning trades tells of targets, but unless you are other than retail, going to be tough doing volume of trades. Volume spikes can be caused cause of time stops, I use time stops, something has to occur in so much time or 'this" will then take place. So if a trader has a rule they must get out then press markets' supply to get out and market jumps, doesn't mean there is a trend, just someone having to get out, so often times these "spikes" have no meaning cause of price patterns or action, so your order going back to the mean becomes like a Bollinger band around price of just getting out in the middle. Think about it, your orders are the Bollinger Bands or percentage there of, you manually or automated seeking spike volumes and that is dictated as speed of market orders hitting Globex. You study Price and Bollinger bands long enough, you know when price is expanding for strong trend and when it is just flopping around. When price flopping around is when volume spikes works best.
One minute bars are like forever, if you watching the Dome, how long do you think 1000 lot market order would take during lunch time, 2 seconds? So what your program is set to cancel orders if price gets too close without quick approach.
The A/B multicast feeds you get when you plug into the exchange.
I ran something similar to this for a few years about 7 or 8 years ago with equities. Its simple market making
He's layering the book to detect toxic flow and then reversing directions (hoping for momentum ignition).
unless I am reading wrong he's putting bids and offers around the bbo (ie. making a market) and hoping for a sweep and reversal (not momentum)
why should any of the information you provided provide any edge whatsoever? (Serious question). You have no way of knowing whether that "swipe" was just a larger order and that it completed or whether this was the first reaction to some news that hit the tapes and may put you in the red by 20-30 full points within seconds and spit you out upside down....And what do you mean with "How many levels must be wiped out to make it worth it".
second that. OP, if anything I would research the opposite, the infancy of momentum driven by large orders that cross the spread...oh well, I will not give much more away...