You're now on the archive server. Commenting has been disabled.
Introduction To High Frequency Finance
In the absence of anything else being even remotely relevant in today's market, here is some good bedside reading on this increasingly more pertinent topic.
An Introduction to High-Frequency Finance -
h/t Economics of Contempt.
»
- 4560 reads
- Printer-friendly version
- Send to friend
Td,
Sprott will be interviewed by Eric King at
http://www.kingworldnews.com/kingworldnews/King_World_News.html
King does a good job with his interviews.
And while there is not much going on I thought an OFF TOPIC update on Swine Flu would be worth posting as it is revealing itself in the S. Hemiphere.
Our hospitals are having a surge in pregnant women on ventilators and babies being born prematurely, this is highly unusual in our hospitlals, it has gone from 0 historically to about 6 today.
People are on cardiac respirators too ( thier blood is taken from their bodies and ventialated due to lung damage)
So, this thing is turning ugley and out peak flu season is August!
Sorry for being Off Topic, but I think this important, esp if pregnant women are planing a visit downunder!
"UP TO six new mothers infected with swine flu are on life support after giving birth prematurely because the virus was threatening the lives of their babies. "This situation has become very, very grave," a midwife said. "Usually we don't have any pregnant women in ICU with influenza so this is worrying," he said.
Swine flu can pass to a foetus through infected membranes in the placenta or can cause a baby to overheat if the mother is febrile. "
http://www.smh.com.au/lifestyle/wellbeing/swine-flu-attacking-mothers-newborns-20090715-dlly.html
July 16, 2009
swine flu...yawn
Got Clift-Notes versoin?
Much of the material in this book is about high frequency statistical arbitrage. They take some pretty simple concepts and make them much more complicated than they should. Just look at Goldman's Global Alpha fund. They pretty much followed they formulas and methods in this book. Until one day things stopped correlating or stopped being cointegrated and then they all realized that it was much easier to make markets, stat arb bid-ask spreads and electronically front run clients as well as add momentum algorithms. HF has moved way beyond this book into things you can not publish because they are either illegal or market manipulative.
Apocalypse Now-
350+ pages, impact of new technology comments:
CAPM is static/linear, and does not take into account different investor time horizons and trading methodologies. "From a dynamic perspective, the notion of reduced friction should be central to the notion of efficiency."
"A quantitative measure of efficiency might be derived from the bid-ask spreads. Spreads are not only a measure of "friction", they also contain a risk component". Interesting, wonder if spreads could be modeled like options using Black-Scholes.
"It takes differences in expectation among market participants to explain why they trade." including geographic lags and access to news updates. Clearly true unless robots goal seek and trade amongst themselves for the liquidity rebates.
Also "Numerous studies have shown that simple trading rules do not work in efficient markets."- if we extrapolate, it might explain that in order to have a market where traitors do make money, the system associated with our markets(gov, media, banks, & related corps) need/want inefficient markets with heterogeneous participants split into groups with different expectations by way of social engineering. Think right/left paradigm, deflation/inflation, optimist/pessimist, green shoot/green weeds.
Homoskedastic Fat Tails of the Ljung-Box variety with excess degrees of freedom scare the shit out of me!
where were you?
we just finished a 3 day conference on HF-finance
Modeling High Frequency Data in Finance The workshop will take place at Stevens Institute of Technology between July 10 and July 12 2009.The conference took place between July 10 and July 12 as planned. There were 123 participants at the conference. We wish to thank all the conference participants, and we hope that the presentations will help you with your future research and development of algorithms and models for studying data sampled with very high frequency.
A report about the special session and some of the talks are available here.
This is a joint workshop (Stevens Institute of Technology and New Mexico State University) in high frequency data modeling. The main purpose of the workshop is to provide synergy between Econophysics and High frequency data modeling. Specifically, the focus of the conference is on models for high frequency data and in particular applications of statistics and statistical mechanics to this modeling problem. Furthermore, we are interested in complex systems and system of systems as applying to this problem.
The scientific motivation is to bring together the best mathematicians, practitioners and regulators to help develop and better the modeling aspect of the marketplace.
The main objective of this meeting is to expose today's economic and modeling problems to mathematicians and current graduate students in the hope that this will improve the quality of the research problems studied at the moment of the conference.
The location of the conference is split between Bissinger Room in Howe Center and the main auditorium in Babbio (B122). A campus map can be accessed here. The registration is in Howe Bulding 4th floor on Friday July 10.
Last Updated on Wednesday, 15 July 2009 15:48
Goals and Topics of the conference The main objective of this meeting is to expose current economic and modeling problems to mathematicians and current graduate students in the hope that this will improve the quality of the research problems currently under development. A secondary objective is to strengthen the collaboration between mathematicians, physicists, economists and industry. The conference will focus on three interrelated aspects:
- Stochastic modeling and statistical analysis of high-frequency data
- Models in econophysics and application to the analysis of high-frequency data
- Systems and complex adaptive systems in finance
The first topic is the main subject of the conference. Exposing the state of the art models for the high frequency data to as many mathematicians, statisticians and practitioners as possible will help exchange ideas and create a venue for further improvement of the models. The second topic is equally important. Physics has a much longer history than economy of analyzing high frequency data. It has developed models that can be translated to other areas and improve the current state of knowledge for the high frequency data modeling. For example, Quantum stochastic processes were introduced first in the 1980's by L. Accardi, A. Frigerio, and J. T. Lewis. The general Levy processes that have been used in Finance during the current decade, were developed for practical use in physics since the end of the 80's, the theory of what was then called quantum Levy processes being developed extensively by M. Schürmann and U. Franz. Finally, the third topic aims at studying the financial systems as a hierarchy of complex systems and subsystems whose synergies at every level produce emergence. Methods of network analysis based on probabilistic graph theory, neural networks, artificial intelligence, nonlinear dynamics and chaos are the core of evolving this theory for financial systems.Last Updated on Thursday, 19 February 2009 11:45
So....Here in lies the problem. We take some of our best minds / talent and set them off on this bull$hit, unquantifiable phenomenon. But...but...but your model just needs to be a little more complex. Chaos theory, yes you need to add chaos theory. Wait! bosons, you need to add vector bosons. Ignorant. So smart, yet sooooo stupid.
The funny/sad thing is that all these bright minds try to figure out financial markets for the sole reason of becoming rich?! Where is the social/economic benefit of all the HF? Or even something that people can consume?
It's just trying to move the money from one place to the other. i.e. pension funds, retail investors,etc into the pockets of some fat cats.
Sad, the amount of human knowledge and expertise gone to waste
Couldn't agree more. Nothing wrong with wanting to be rich it's just a damn shame that this biz is such a large conduit.
a good alternative to using the GARCH model:
Probability distribution of returns in the Heston model with stochastic volatility
Quantitative Finace Volume 2 (2002) pp. 443-453
"Abstract
We study the Heston model, where the stock price dynamics is governed by a geometrical (multiplicative) Brownian motion with stochastic variance. We solve the corresponding Fokker–Planck equation exactly and, after integrating out the variance, find an analytic formula for the time-dependent probability distribution of stock price changes (returns). The formula is in excellent agreement with the Dow–Jones index for time lags from 1 to 250 trading days. For large returns, the distribution is exponential in log-returns with a time-dependent exponent, whereas for small returns it is Gaussian. For time lags longer than the relaxation time of variance, the probability distribution can be expressed in a scaling form using a Bessel function. The Dow–Jones data for 1982–2001 follow the scaling function for seven orders of magnitude."
What is the best way for me to share a document? I wanted to embed a paper concerning this topic, but was unable. I have the doc loaded to docstoc.
Yeah..that was a good read ;)
Any other suggestions for reading/resources/classes about this? I would like to learn more about high-frequency market making (front-running) than traditional stat arb. Anything on exchange order handling, rules, latency etc. would be helpful.