Recommend 'Time Series Analysis for Dummies'?

I am a nuclear engineer by training with several patents for radiation shielding design in the early '90's. I've read Zivot & Wang's "Modeling Financial Time Series with S-Plus", I own MatLab, the MatLab Neural Net Toolbox, and Statistica.

I don't make money trading with them per se. No holy grail or black box that I've been able to find. For me, at least, I have never been able to forecast flat price directional markets but I have been able to model strong correlational tendencies between market instruments that behave much better statistically than flat price instruments. Enough of a positive risk/reward skew to make money at it for 18 consecutive years and counting, with every market sector and with just about every market condition.

Most practical time series analysis is used by professional risk managers.

Using statistical tools to mine and exploit arbitrage opportunities IMO is a much more worthwhile goal than trying to predict forward price.

In the end, it's all about the money. You don't have to predict future outcomes to make money in the markets. If your reasoning is that you might one day be able to identify statistically relevant tendencies, then fine. If you want to build forward price curves, then fine. But please don't think that success in the marketplace requires looking at one instrument in isolation and then applying GARCH or ARIMA or whatever to that one instrument.

Look for relationships.
 
I urge you to seek out the work of John Ehlers. Lots of his code is freely available on the web if you can program a little bit. Also, his book Rocket Science for Traders is very helpful.
 
Ehlers is so shallow mathematically that he doesn't even understand when the existence conditions of the tools he uses are invalidated by the inherent nature of the market.
 
Quote from bone:

I've read Zivot & Wang's "Modeling Financial Time Series with S-Plus",
More...

That's a good book, and Zivot is an excellent teacher / explainer, both in person and in writing.
 
this book is on my wish list. According to the review on amazon you don't need to know calculus to understand this book.

http://www.amazon.com/Forecasting-T...UTF8&coliid=I4N9FM64758PE&colid=2H7BWO8P6JQKJ

Introduction
An introduction to forecasting p. 1
Basic statistical concepts p. 27
Regression analysis
Simple linear regression p. 79
Multiple linear regression p. 139
Model building and residual analysis p. 221
Time series regression, decomposition methods, and exponential smoothing
Time series regression p. 279
Decomposition methods p. 325
Exponential smoothing p. 345
The Box-Jenkins methodology
Nonseasonal Box-Jenkins models and their tentative identification p. 401
Estimation, diagnostic checking, and forecasting for nonseasonal Box-Jenkins models p. 449
Box-Jenkins seasonal modeling p. 489
Advanced Box-Jenkins modeling p. 539
Table of Contents provided by Blackwell. All Rights Reserved.


If you buy this book let us know how you like it.
 
I traded commercial OTC electricity for a while in the '90's, and recently financially-cleared ICE and Clearport Power Swaps - "Modeling and Forecasting Electricity Loads and Prices, A Statistical Approach" by Rafal Weron is the most marketplace-applicable and one of the most relevant books I have ever read for time series analysis. Rocked my world. Pumped some wattage into my cottage.
 
You must log in or register to reply here.
Back
Top