The New Trading Systems and Methods

by Perry J. Kaufman

Get the bestselling guide to trading systems, now updated for the 21st century. For more than two decades, futures traders have turned to the classic Trading Systems and Methods for complete information about the latest, most successful indicators, programs, algorithms, and systems. Perry Kaufman, a leading futures expert highly respected for his years of experience in research and trading, has thoroughly updated this bestselling guide, adding more systems, more methods, and extensive risk analysis to keep this the most comprehensive and instructional book on trading systems today. His detailed, hands-on manual offers a complete analysis, using a systematic approach with in-depth explanations of each technique. This edition also includes a CD-ROM that contains the TradeStation EasyLanguage program, Excel spreadsheets, and Fortran programs that appear in the book.
CD-ROM/DVD and other supplementary materials are not included as part of eBook file.

Genres: Finance

1200 pages, Hardcover

First published February 28, 2005

Reviews

[Nick · Rating 5 out of 5]

Great book on system testing and system design.

Chapter 21: System Testing
-Expectations are an essential part of system development. They force you to define a plan in advance that can achieve certain goals.
-The merits of the system would be questionable if the trend could not show profits on its own, but the combo of the RSI and the stop turned a losing trend into a net profit.
-A time series has point to point correlations that we are trying to exploit, thus synthetic data is of no use unless it exhibits those attributes.
-Adjacent test values should be used to interpolate and replace actual observed value.
-If the rule change improves the specific situation that it was intended to improve, but decreases performance in other situations it is likely not a robust improvement (think G3 end of 2011 for dmr filter).
-Searching for robustness: Must be based on a sound premise (fundamental reason for existence of the pattern), it must adapt to changing market conditions, it must be tested properly with proper expectations, it should be modestly sensitive to changing parameter values, parameters should be derived from market environment whenever possible.
-Retesting is necessary after 5% to 10% of new data is available relative to the original test interval. The size of the shift in optimal parameter values should be small when new data is added.
-Markets have individual characteristics that might justify differences in performance; however, a strategy that can generate profits in a broad range of markets with the same rules and calculations in robust.
-Viewing the results of a set of similar indicators gives a much better idea of the success of the approach than testing any one of them. Not all have to work equally well, but their results should share statistical similarity or the approach is not robust. A market that performs well using a daily breakout, but fails with a dual trend crossover is not a good candidate for trending profits in the future.
-When a price shock is an important part of the data being tested, the best optimized result is the one that traded around the shock most advantageously.
-Be critical of good results

[Tadas Talaikis · Rating 5 out of 5]

If you're not acquainted with math, this book is very good starting point, a lot of classic concepts presented in very simple manner. You'll not find actually working systems (a lot of them are dead long ago), but for me it was good source of further ideas. I've programmed more than 10 indicators based on those ideas, so it was a long, but worthwhile read.

[Kem Bong · Rating 5 out of 5]

This is the bibble of trading systems, the author not only presents simple techniques like MA, Stochastic, Bollinger band... but also modern techniques like Genetic Mutation, Neural Network (AI), Anti-Martingale, News effect (in math form)... every chapter is full of info, if you are a beginner you will be overwhelmed by these techniques, don't be, take it slowly, read each chapter, write some code to test each idea, evaluate the performance of the idea, then you can move to the next. You can read the later chapter first to see how you can evaluate a system, what is a good test, what is a good performance, how to Not overfit ! For veteran, this is a very good reference book, even if you are FA analyst you will find good system here, choosing sector, compare stocks, when to get out (temporary), "hold forever" is a myth. FA is hard not because it's about domain specific knowledge, it's hard because there are not enough data point, you have 4 quarterly report, that's only 40 data points for 10 years of price data. In AI terms you need at least 200 data points to properly optimize 1 parameter, so 40 data points for 10y is too few. When you put all the FA terms into number you can apply many method in this book to analyze and improve your evaluation. Numbers dont lie!!!

[Lukas · Rating 5 out of 5]

This is the most comprehensive book on trading I've come across. The information inside is not only incredibly detailed but also of a particularly high quality and very concise.

For full disclosure: I haven't read this one from cover to cover. It's quite large (~ 1200 p.) and functions more as an encyclopedia than a "narrative-driven", chronological textbook. It's better used as a deep reference text you can jump into for specific topics.

[Keith Lansford · Rating 3 out of 5]

Dry, dry, dry....tons of information, but Kirkpatrick, Pring, Edwards and Magee cover technical analysis far better with their respective books. But, this gets three stars because it is a good reference, as it has a slew of equations, numbers, and TradeStation/Excel programming for oscillators/indicators/etc. that could prove useful. Use it for reference...unless you can't sleep...then whip this baby out and you'll be down in 10 min.

[Bernd · Rating 4 out of 5]

Very thick volume, a treasure trove of trading systems, methods, supporting math and concepts, all the while very readable to the motivated trader. It can be read non-sequentially and in piece meal, without losing context. I use it frequently as a reference, or for occasional inspirations of different methods.

[Ed Ball · Rating 4 out of 5]

Felt like I needed a college math course to brush up on before reading this. But lots of good systems, the trick is finding the system that works in today's market.

[Rohit · Rating 4 out of 5]

One of best book ever on trading systems, surely must read before developing your own system..