<< 1 >>
Rating: 
Summary: Poorly explained
Review: I have a university maths degree and found the book very obvious and drawn out for the first few chapters. In spite of this I looked forward to what was going to be explained later. Suddenly from a very simple and easy to understand explanation on the EMH he starts to use mathematics in his equations that I had a lot of difficulty following. There was very little or no explanation of how these equations were arrived at and a lot of mathematics and statisics is assumed. This book does not apply the theory in ny meaningful way to the markets let alone the capital markets in my opinion. I found that I took very little away from this book and would not recommend it to anyone who has basic mathematics like myself or is looking for some deeper insight into the markets. I would hate to have Mr Peters as a teacher based on his book.
Rating: 
Summary: A very good introduction
Review: I read this book, the 1991 version, years ago. Around 1980 my own attempts to crack share prices statistically convinced me that all share prices behaved like a Gaussian random walk meaning that all speculation was comparable with playing roulette and I am not one of those guys who usually wins when gambling. This view was strengthened when the option pricing model came up, meaning that even the real pro's in the field assume that share prices are nothing but a random walk. This book has opened my eyes to the fact that there is much more to randomness than just the Gaussian curve. Share prices are not fully random. Impressive is the demonstration that an RS analysis on the real data is different when applying the same RS analysis on scrambled data. So there is information hidden in these time series, somewhere. Since then I have picked up the subject of cracking time series again with great pleasure. I think this book is exceptionally well written and without it I doubt if I would have been able to follow Mandelbrot's book "scaling and fractals in finance" that I bought later. The book is about understanding a subject, not about learning a simple formula to apply on a time series.
Rating: 
Summary: A dated overview, with little real meat
Review: The second edition of this book was published in 1996. The book
seems to be largely based on Feder's 1988 book "Fractals". The
dated nature of this book means that it is missing later work
on long memory processes, which Peters estimates using the Hurst
exponent.As one reviewer already noted, don't assume that this book will
provide much in the way of useful equations. For anyone who wants
more than an overview, this book is a disappointment. Peters does
a poor job of explaining the equations and I did not find enough
detail to implement the algorithms discussed (I turned to Feder's
book and various journal articles). The book does come with a
"floppy" disk containing the Visual Basic algorithms. This is
a poor choice, since C is pretty much the lingua franca for
algorithms. The various chaos and fractal techniques are applied to a handful
of financial data sets, but this is far from even a solid
suggestion that these techniques might be useful to anyone
developing real market models. Some of the conclusions that Peters draws (cycles in financial
data) do not seem to be supported the evidence he presents. In summary, if you are looking for something beyond an overview,
save your money. Feder ("Fractals") has a better description of
RS calculation. "A Non-Random Walk Down Wall Street" by Lo
and MacKinlay has a chapeter on the application of the RS
statistic and long-memory processes which is much better than
Peters. For those who need to simulate fractal brownian motion
(data sets with a particular Hurst exponent) "The Science of
Fractal Images" by Barnsley et all is a good reference.
<< 1 >>
| 
