Search results

Is anyone aware of any methods to calculate the volatility of price points deeper in the book than just the last-executed price? Let me clarify. Standard models (garch,etc) operate on trade prices to calculate volatility. Given a full depth view (TotalView,Ecns,Openbook,etc) you can...

RedSky Financial Genesis Lime Brokerage

What do you mean by "works well"? You need some kind of success criteria. Seems like glorified TA to me.

Potentials of Unbalanced Complex Kinetics Observed in Market Time Series Abstract: As a model of market price, we introduce a new type of random walk in a moving potential which is approximated by a quadratic function with its center given by the moving average of its own trace. The...

Perhaps you should use some sort of continuous time model.. e.g., the time between events is another parameter. quantatizing tick data is like putting a square peg into a round hole.

End-of-day data is nearly useless for this kind of analysis. Closing prices are completely arbitrary and don't have much more significance than say.. the prices at 2PM.. or 1PM each day. Tick data should really be used in these cases. The larger the timeframe the more unpredictable it is.

Denton.. hippy college town 30 miles north of downtown dallas. Lots of beer drinking and shitloads of young hot girls!!!

quickfix is actually pretty flexible.. you can define more messages in the master config file and it'll regenerate all the datatypes and marshalling code for you. --Stephen

Right, fix is usually for orders only.. it CAN handle data, but nearly no one supports that because the protocol is so bloated and slow. I've used quickfix for a bit.. never in production and I found it to be pretty annoying how the sessions and reconnect/sync logic works. I think there...

In case anyone was wondering. The Econometrics of Ultra-high-frequency Data can be found here.

Ahh yes you are right. I remember this now.. haven't used a matrox card in years. The hardware situation is still very depressing. I wonder if anyone has ever actually proven that opening specs can help the competition?

Dude, Matrox was one of the first to open up all of their specs. I led a project way back in the day to create OpenGL drivers based on their specs. See Utah-GLX. This work was later ported to be part of XFree86 proper.

Are you serious? Any ATS worth it is going to run in a data center on some serious hardware and not rely on some slow/flaky residential connection(s).

lol. You definately won't have any luck with old TA methods like that.

Go and read the QQQQ prospectus. Yes, the NAV is updated every 15 seconds but you can only create/redeem at the end of each business day and in increments of 50,000 shares plus $1000 commission costs, not to mention the commission for buying/selling all the underlying securities and accounting...

What about data? When are you going to publish the protocol for the data feed? Entering orders from linux is useless if you cannot receive tick data.

Unless every single quote on the buy side got lifted within 5ms of each other then it's not hard for an algorithm to think "oh shit, 25% of the quotes were just lifted, sell!"

yes that is overly simplistic. A bot would see the start of the stampede and jump in with it long before it was over.

Where do you get the idea of high risks? High frequency trading is actually less risky because there is less chance of you getting caught in one of the fat-tails.

Dude, this is standard timeseries analysis. There are numerious books on this subject and they are using very well known methods.