March 2, 2010
Multi-agent based order book model -> paper
Simulation of limit order driven market -> paper (JOT), (check reference as well)
Matlab code -> here
March 2, 2010
Harvard senor network Lab(robotBee, Mercury). There is really some interesting thoughts, hardware implementation of financial data processing application (analogue to Soundblast and GPU). and Tito Ingagiola point out this, apparently ICAP is using this. Tervala (switch technology, option MM case) is backed by GS, Sigma(database VC), Acartha(financial tech PE), Nothhill( financial tech VC). Tervala using ( Arista switch technology(cisco, stanford cs guys), Celoxica( system design, UK based, former INSTINET, IB IT, HF Operation), Fix Flyer(process automation, courant math, nasa software engineer, IB op engineer), Portware (trading app design, former trader and IT), softModule(grid computing), Volante (sys integration).
on hardware side, it’s ASICs, FPGAs, network processors, GPU etc. etc.
even valuation and risk guys has interest in GPU, for monte carlos?
this two webinar about algo trading and infrastructure.
Yahoo Pipes, graphic programming. this is an awesome application!
Exchange simulator JessX
February 28, 2010
Unscented Particle Filter Talk
Option MM Baird
General direction: Liquidity – Arbitrage – Understand details.
Elwyn Berlekamp, i’m late to the story but Misha Malyshev(ex-citadal HFT, Teza technology LLC & the guy behind Aleynikov GS saga(similar story at UBS, and this interesting post)had similar track/background with a few authors in previous post (MIPT, Princton, hedgefund etc. etc.) Tactical (Turtle trend following fund The System incorporates mathematical market models that integrate key elements of modern portfolio theory, chaos theory, and proprietary money management concepts. sounds like snake oil, no?)
Detection of security fraud would be an interesting topic. e.g. how can madoff return being detect? not just suspicsion but quantifiable measurement of likelihood.
Ad Hoc inference, a lot of talk about boosting, bagging etc.
Time Series shops ( inefficiency/prediction from time series):
AHL (ManGrp, oxford based, stats, time series),
Winston (ex AHL stats, trending follow, old people)
Transtrend( Robeco grp. Netherland based systematic 7b AUM), QIM(2b AUM)
Cantab Capital (ex GS quant strategy euro head, Cambridge based)
Crabel(Milwaukee based Toby Crabel had a short stint with Niederhoffer)
John Locke(french fix-income arbitrage)
February 28, 2010
Two Sigma: D.E.Shaw math + Tudor IT
EWT: ex-nymex CEO
Getco (Chi market making), Forbes on Getco, WSJ, Bloomberg story, the prophet is being fulfilled, and market kept on evolving. with different player join the arena, the natural question is: What’s NEXT?
Ikos London based HF shop. they are not even hedge fund<what’s the point to be a fund?? dilute return??>
Chopper Trading, Chicago based, said nothing
World Quant, igor Tulchinsky, Upenn/Millennium stat arb.
Capstone, (organizer+ ex-pit trader+ finanical engineer+???)
February 27, 2010
Klipp’s saying was “You have to love to lose money, to take small losses, and hate to make money to be successful. There’s no other way.”
The hard and stiff will be broken, the soft and supple will prevail.
In July 2009, Spitznagel opened a fund betting on hyperinflation
Clone of low turn over value focused fund actually make sense(?)
Kaching: another Clone.
SharePost: private company research and exchange.
Second Market: illiquid asset market place(?)
Portfolio Solution, diversified(?) index fund ETF portfolio charging 0.25% management fee.
IASG: many listed CTA, the strategy description could be interesting.
Investment outlook for 2010: 立此存照
Mebane Faber’s idea for 2009, all seems reasonably good idea with hindersight.
February 27, 2010
February 23, 2010
I just finished re-reading Tasy’s Finanical Time series analysis, here is a list of points worth taking notes of:
- What time series analysis is trying to model here is f(x_t|F_t-1), so what we are trying to understand here is how mean and variance related to the past, through ACF.
- portmanteau Test Q(m) (Ljung-Box stats for testing sufficiency of model) i.e. residue of dynamics, i.e. no series correlation no conditional heteroscedasticity). Q(m) ~ X^2(m) link to p stats to reject null hypothesis( no series corr in residue, no CH in residue square etc.)
- model identification( PACF or Akaike Information Criteria) -> parameter estimation (OLS with significant level of parameter estimated)-> model checking ( check if residue series is close to white noise, using Ljung-box Q(m) ~X^2(m-order(AR model used))->conditional forcast with model(i.e. error in model is not taken into account).
- unit-root nonstationary, long memory, ARIFIMA(d), ARIMA(1), dicky fuller test for unit root. Ljung-box test, look at ACF, same effect.
- seasonality model, ACF contains information for AR and seasonality. (i.e. a tool to detect seasonality & predict shape of futures curve).
- nonlinearity test( Q, BDS, F, Threshold Test), parametric( TAR, Markov switch model), nonparametric( NN, kernel, MCMC etc.
- nonsynchronize trading, var(r_o) v.s. var(r). bid ask bounce, and high frequency dynamics(negative lag-1 AR)
- order prohibit model(lo et. al.),A Decomposition Model(ADS)* ( using partition I(i) indicative function and MLE to estimate parameters) ; duration model( using combination of quadratic function to remove duriual effect), using ACR model with exponential or generalized gamma innovation; non-linear duration model with two regime. bivariate model for both price change and duration( with method similar to ADS, this is close to what altreva is doing, i.e. simulate the market)
- cross-correlation matrix contains linear dependency information( couple, direct, independent) .
- VAR model able to absorb all dynamic dependency and concurrent dependency information into transition matrix, with cholesky decomposition. stationary condition, how to test for cointegration in real applicaton, Tasy mentioned about difficulty. And erro correction form( still fuzzy about this concept).
- PCA, FA, MCMC