Modeling Trading System Performance: Monte Carlo Simulation, Position Sizing, Risk Management, and Statistics

Thanks

Perfect.

I became an active investor and a serious student of trading in 2008. In subsequent years, buoyed by my initial success I began intraday trading. After a lucky start in a bullish period, inadequate money management techniques caught up with me and led to disaster. During that period of soul searching in late 2011, I ordered Modeling Trading System Performance (MTSP).Excitedly, I skimmed through it and at first my heart sank. Why? I have Dr Bandy's other books and I consider them excellent. Why didn't I like this one at first glance? Because it laid out all the concepts I needed to *really* understand in order to build robust systems. My dismay was not at the quality of the book but at the recognition of all the things I needed to learn. Is learning to model trading systems useful? Turns out it is an extremely valuable and necessary first step before any system development should begin!Like all of Dr Bandy's books, this one is extremely well written with his characteristic attention to detail to every aspect including the layout, font-size, illustrations and step-by-step instructions. Complex topics are explained astonishingly well. At first read I could understood the individual concepts. It took me repeated readings to get the big picture but since that 'Aha!' moment, MTSP has become my *favorite* book on the topic of trading systems.This book provides the best framework that I've come across to begin thinking about trading systems and now I understand why this book omits specific buy/sell rules as they are just a small part of system development (and besides Dr Bandy has other books that address these topics).This book covers:- How to think about the basic statistics of a trading system -- expectancy, CAR, MDD.- An excellent discussion on how longer holding period increases risk.- Position sizing and the importance of understanding the distribution of trades. This is a pernicious problem that plagues every rosy back test result. The book links to an excellent excel plug-in to compute a realistic estimates of draw downs using Monte Carlo simulation and complete step by step instructions. I found the hands-on exercise in the appendix to be immensely beneficial in helping me learn and internalize the concepts.- How to determine the health of the trading system. This is a goldmine of information with practical techniques for this very important phase of monitoring trading system you placed into production. The treatment of this topic alone is well worth the price of the book!Will this book alone change anything? Yes! It changed the way I approach system development and validation. It has provided me a solid logical foundation and as I build new systems I find myself constantly referring to my dog-eared copy resplendent with neon post-it page markers. I think this book is one of those timeless classics and I can't thank Dr Bandy enough for writing it.

I had to read the book several times over to make sense of the ideas and how they fit around my existing knowledge of mechanical trading systems. To me, my main criticism is that the most important part of the book, i.e. step by step guide to conducting the Monte Carlo process to generate the drawdown and terminal wealth relative distribution curves were in the Appendix! I felt all the results would have been more convincing the first time around if this part was introduced first.The comparison of trading to games like roulette, blackjack and poker were both interesting and entertainment - probably the easiest part of the book to read.I'm still in the process of understanding and trying out the ideas in "Is it broken?". I agree with the ideas on a fundamental level - my statistics knowledge is a bit rusty (I'm an engineer) so I struggled to follow the logic line by line. Some more clear examples could have helped.Chapter 8 - Bar Length and Holding Period - I felt could've been written a little better. Fell asleep after page 3 of the chapter. Skipped to end of chapter summary when I next picked up the book.Chapter 10 - Position Sizing - this was not as I expected. I expected to see the effect of various mainstream position sizing techniques such as scaling in, scaling out, volatility adjusted (=f(ATR, account risk) ) on the drawdown and final equity curves. The appendix calculation shows an example of a fixed fraction (vs fixed ratio) position sizing in the Monte Carlo simulation. It would have been good to expand those appendix calculations to show how fixed ratio, scaling in and out would have been incorporated in those calcs.Chapter 9 - Tradables - this chapter felt like filler material. I failed to see the relevance of this to modeling trading system performance.Overall, it was a good book. I just wished the sequence of information presented in the book was a bit more orderly.I greatly benefited from reading the book, and it's given me some sound ideas on using MC simulation to measure risk, and some fundamental ideas on testing whether the system is broken - tools I can add to my belt when constructing these mechanical trading systems.Thanks for the read Howard.