NEWS FLASH

Meetup Dec 12 in North York! Quant ‘Secret Sauce’ tricks of Matlab, bridge to C++/C#, FREE .NET open source HFT trading platform, and MYSQL historical database for back testing

Register and details at:

http://www.meetup.com/quant-finance/events/42092262/

Quant Book Opinion

Goldman Sachs quant PDF research papers added

Written by Administrator Tuesday, 01 February 2011 16:22

Goldman Sachs quant PDF research papers added

http://quantlabs.net/labs/quant-books/cat_view/22-goldman-sachs

 

9 Morgan Stanley quant PDF articles added

Written by Administrator Tuesday, 01 February 2011 16:21

9 Morgan Stanley quant PDF articles added

http://quantlabs.net/labs/quant-books/cat_view/23-morgan-stanley

inside the black box: the simple truth about quantitative trading..

Written by Administrator Tuesday, 01 February 2011 14:45

inside the black box: the simple truth about quantitative trading...

my new book is now available at http://www.amazon.com/Inside-Black-Box-Quantitative-Trading/dp/0470432063/ref=ntt_at_ep_dpi_1 . more info on the book can be found athttp://www.thequantbook.com. a couple of the endorsements it's received so far: 

Blair Hull, Founder, Matlock Capital: “In Inside the Black Box: The Simple Truth About Quantitative Trading, Rishi Narang demystifies quantitative trading. His explanation and classification of alpha will enlighten even a seasoned veteran.” 


Peter Muller, Head of Process Driven Trading, Morgan Stanley: “Rishi provides a comprehensive overview of quantitative investing that should prove useful both to those allocating money to quant strategies and those interested in becoming quants themselves. Rishi’s experience as a well-respected quant Fund of Funds manager and his solid relationships with many practitioners provide ample useful material for his work.”

----

I read Rishi K Narang's Inside the Black Box, The Simple Truth About Quantitative Trading, as a person who has many, many years working with gigantic amounts of economic and financial data with the specific purpose of forecasting future outcomes. 

This book is a must read for those who want an insight into the world of "high finance". Well written, simple, clear and does not require an understanding of the mathematics of finance. He lays out the how and why of quant trading, in non-technical term, that allows all of us to appreciate that quants are here to stay. Financial services cannot do without quants, just as much as we cannot do without markets. 

Valuable to both newbies and seasoned statistical modelers such as myself who have not been directly involved in trading. He points to sources of errors in financial strategies in both the quant and non-quant world. I particularly liked that he had pointed to the problems with correlations. 

Narang provides an insight into the world of quants. He explains how and why quants implement strategies and the choices available to them. That at a high level there are a limited set of strategies - one can count on one hand - but at the detail level the number of choices explode. 

Reading this book told me that Narang was not just another quant. He is a quant who is always asking the question, how can we build a good model with this data? As it is one thing to build a model but quite another to dig deep into the data to find out why it ticks.

 

Simple models that generate heavy tails?..

Written by Administrator Sunday, 30 January 2011 14:48

Simple models that generate heavy tails?..

One thing that always interested me in university is: "why do people use brownian motion and normal distribution to model financial time series, while these series obviously DON'T behave that way??!".
Now that several years have passed since I graduated, I see that it is hard to use any other model.. But it is still possible, isn't it?
What I'm working on at the moment, is to build some sort of SIMPLE discrete process or a stochastic process that would show heavy-tailed behavior. Any suggestions??
--
The problem with not assuming normality is that the resulting model is not precise enough to make money on a daily basis. I believe that this is the fundamental weakness of quantitative finance and is one reason why human experience and expertise will never be fully replaced by math.
--
'm not arguing that prices actually exhibit Brownian motion. I'm saying that there is not a superior quantitative methodology for the short run. Managers will always have an incentive to misinterpret any risk methodology because of the ability to profit in the short run. You would be amazed at how complicated a problem becomes when people have an incentive to not understand it. It all boils down to whether you would prefer to be right most of the time and collect your annual bonus or be right at the critical time and be unemployed.
---
Because of the tails, Black-Scholes prices in-the-money options too high and out-of-the-money options too low. Why not find someone who has deep pockets and uses Black-Scholes because it's "precise", rather than accurate, and then systematically write and sell them in-the-money options, and buy from them out-of-the-money options? Are the trading expenses higher than the expected profit? Has anybody modeled this? One would just need historical data, including option prices, and no new model of the tails -- the tails are in the data.

It's amazing to read that "a novice need not go far beyond Black-Scholes to make money in the options markets"... (source: product description for "Basic Black-Scholes..." by Crack). Is this market *that* inefficient? Disclaimer: I've never traded options.
--
Because of the tails, Black-Scholes prices in-the-money options too high and out-of-the-money options too low. Why not find someone who has deep pockets and uses Black-Scholes because it's "precise", rather than accurate, and then systematically write and sell them in-the-money options, and buy from them out-of-the-money options? Are the trading expenses higher than the expected profit? Has anybody modeled this? One would just need historical data, including option prices, and no new model of the tails -- the tails are in the data.

It's amazing to read that "a novice need not go far beyond Black-Scholes to make money in the options markets"... (source: product description for "Basic Black-Scholes..." by Crack). Is this market *that* inefficient? Disclaimer: I've never traded options. The problem that you facew is that you need a distribution that when combined - it produces the same (or at least a known) distribution. Normal and LogNormal does this,but it does not have fat tails. There is a family of distributions - The Levy Distribution - that do have fat tails and do produce another family member distribution when combined. These are often used in finanance and a quick Google search will find documentation.

On the rest of the discussion - I would say that the reality is the other way round. Short term stock movements are highly non-Normal. Longer term is more Normal. The Central Limit Theorem would tend to support this.
--
• Longer-term is closer to normal than short-term, yes, but longer-term is still decidedly abnormal. Whether it's abnormal enough to violate your assumptions in your projects is your decision.

Getting the size in the tails correct is only half the battle, and even then, you have to optimize where in the tail you want to get it right (10%? 5%? 1%? 0.5%?), knowing that THAT choice means it'll be wrong in the other places. Student's t distribution with a low degrees of freedom (3-5) will get the heavy tail you're looking for and be very simple and spreadsheet-able.

Getting the size of the hump in the middle correct is the other half of the battle. Not getting BOTH the tails AND the hump right will impact your simulations, more and more the higher your count of trials goes.
--
I havn't, but I know that a number of banks do use Levy processes.
---
Just like was mentioned above, a model, any model of such magnitude is useless to make money. However, to gain a better understanding of how the markets work, we need to consider the process, not the distribution. Besides, distribution can and does change all the time. MMAR (Multifractal Model of Asset Returns) of Mandelbrot is just such a model. This is a model which generates a random process with fractal parameters, where the distribution can change. Mandelbrot himself insisted that we can not discount the infinite variance possibility, so for any risk applications, I'd think that Levy distribution is a good limiting case.

As far as the data, it doesn't tell us much - daily data is a lot more volatile than aggregated monthly data, for example, and if people use yearly data, all of a sudden they try to claim normality, which is absolute nonsense. Also, there are preassymptotic regimes which make a huge difference (between finite and infinite variance) which can not be detected from data. Another aspect that can not be detected (especially in the presence of infinite variance) is the regime switching, which can take your random process generator from a 'thinner' to a 'fat' tailed distribution (and as we know volatility comes in clusters). Even wavelet-based methods are not developed yet to handle regime switching (they have to assume finite moments to work).

It is true, you can not use MMAR to make any calculations - it is a qualitative model. Better a good empirical model than a bogus numerical one.

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=78588

Abstract:
This paper presents the "multifractal model of asset returns" ("MMAR"), based upon the pioneering research into multifractal measures by Mandelbrot (1972, 1974). The multifractal model incorporates two elements of Mandelbrot's past research that are now well known in finance. First, the MMAR contains long-tails, as in Mandelbrot (1963), which focused on Levy-stable distributions. In contrast to Mandelbrot (1963), this model does not necessarily imply infinite variance. Second, the model contains long-dependence, the characteristic feature of fractional Brownian Motion (FBM), introduced by Mandelbrot and van Ness (1968). In contrast to FBM, the multifractal model displays long dependence in the absolute value of price increments, while price increments themselves can be uncorrelated. As such, the MMAR is an alternative to ARCH-type representations that have been the focus of empirical research on the distribution of prices for the past fifteen years. The distinguishing feature of the multifractal model is multiscaling of the return distribution's moments under time-rescalings. We define multiscaling, show how to generate processes with this property, and discuss how these processes differ from the standard processes of continuous-time finance. The multifractal model implies certain empirical regularities, which are investigated in a companion paper.

The above is from Linked In

Quant researcher: Quantitative Methods in derivatives pricing quant book review

Written by Administrator Wednesday, 26 January 2011 12:43

Quant researcher: Quantitative Methods in derivatives pricing quant book review

 

So I finally finished this book. Here is a small review of it.

I will say I wish I started with Carol Alexander’s  Market Risk Analysis Quant Methods in Finance. This book is  excellent in introducing you basics of the linear algebra relationships that is so important to quant. Excel spreadsheet examples help there .This is definitely a recommendation I would put out there.

As for this other book, I really enjoyed the first chapter which really focuses on the Weiner process. Who knew it was important in the world of finance? I thought it was just Black Scholes. Some may not agree. Another highlight was how the author stressed the importance of PDE over Monte Carlo. I am definitely more of a PDE guy because of this book. One of the basic like martingales and jump processes were nicely defined. Good examples too.

As you can tell, this book is more for basic knowledge reader but it is important to stress that professional practicioners would hate this book. I don’t think it is targeted towards them

Some of the more advanced topics like HJM and LIBOR were not as stressed in books like Wilmott. It might have been due to the lack of space. I like the description and examples laid out in the control variates which appear to be crucial to QuantLib.

I found Chapter Six to be the best as it explained a variety of Least Squares in Monte Carlo, lattices, etc. The two supplied case studies help to understand things like early exercise on Bermudian calls as they offer best of both worlds between American and European. The swaption helps too.

All the explanations on barriers were helpful as other books don’t provide solid examples. I did find some parts of the book rushed which I was not a fan of. But there lies the advantage of using others fill in those gaps. Remember, I am not sure many 300 books can complete the kind of topics you are seeking in the world of finance math.

 

Page 4 of 5

«StartPrev12345NextEnd»