On the purpose of Monte Carlo methods in trading

I would like to gauge the community's knowledge on Monte Carlo methods, and especially its utility in Forex trading. Before entering a technical discussion on the topic, I would like to invite you to share your knowledge on the application of such methods in trading.

What would be the purpose of using MC simulations, and what properties would be the model have in order to fulfill this purpose?

My guess is that a less than a handful of people at FF actually apply MC methods. If you are not a practitioner, please take a minute and theorize about what MC simulations could be applied for.

Thanks

I've used simple MC simulations to try to gain an idea as to whether I could safely risk more than 1%-2% per trade.

I simply used an Excel spreadsheet to create samples of 10,000 trades, given variables like win rate, average net win size, average net loss size, trade risk. Then had it plot an equity curve.

In the end, the most important conclusion that I reached was that Excel's random number generator is flawed. LOL.

Got a similar excel Hanover! Not sure I'd really classify it as MC though.

Trading wise, as in picking sentiment/entries/exits etc, I don't believe it provides an edge, or at least not one worth looking into. I mean, the banks spend millions in electricity just running MC simulations, I can't compete with that.

I am not sure whether you can apply this kind of random statistical outcome simulations to a "game" like forex. These methods cannot account for one of the most important factor in the game, which is the player's own psyche - overconfidence, overtrading, taking revenge on the market and so forth.
How can I for example implement into calculations my great ability to generate a nice long losing streak?
I think these methods work much better when applied to games, in which player's actions have no influence on the result of the game whatsoever. They're for example great for showing a gambler that no matter how smart a money management slash staking slash bet placing slash Fibonacci progression system he uses for the roulette wheel, the casino will get all his money sooner or later anyway.

that doesn't make any sense to me at all. how could a monte carlo sim ever provide a sentiment, entry, or exit edge? and where did you hear that the banks run mc sims? why would they want, or need to?

There are pros and cons.

As a positive, MC can help in optimization, robustness, and analyzing sensitivities to unseen (esp. noisy) data.

A large range of synthetic input data can be created to stress test the system over different conditions, and give the designer more insight.

On the output side, it allows for modeling many statistical performance measures like range of profits, draw-downs, streaks, and maximum liklelihood estimates of estimated trade parameters.

As a negative, it is not trivial to capture true market behavior using synthetic data (particularly under IID assumptions). Also, too much data, while adding conviction, can have the effect of filtering out potential opportunities that are temporally finite in nature.

Like any other method, it's useful to know specifically what you are using it for, and how it is being used. Nice topic, btw. 2c

1-I don't know, tbh I haven't though on it much, hence why I said I didn't believe it provides an edge, but who knows. I know there's a guy on myfxbook who has an EA on it.

2-You're kidding me right? How do you think they price stuff? What do you think all those people with PHDs going into banks do?!? It's sure as hell isn't technical analysis and news trading! As to "where I heard it", ummm, on the trading floor of major banks?

If only we could ban this thread from those who haven't read Hull.

1. i suspect we are all thinking of different things here when referring to monte carlo sims

2. they go to swim in big pools of money like scrooge mcduck. as for hull.... i'll just go sit in the corner until i can wear big boy pants.

Knew I should of done that PHD!

Before I go on holiday, I would like to give a final post. When I get back I hope to be able to enjoy a lively discussion on the topic


I must say that I like the responses so far!

You've come to a point that I agree with. I cannot figure out a way to use Monte Carlo to obtain a directional(!) edge in financial markets. However, as the great Benoit Mandelbrot indicated, there is room for application of MC methods in financial markets. To me, this seems to be primarily in the area of the valuation of (non-linear) financial products.

For example, given a certain volatility; strike price; spot price; interest rate and expiration period, one can use Monte Carlo methods to value all kinds of options. And I mean all kinds: American, European, Asian, digital, and whatever else you can think of! However, the most simple MC models would yield no better estimations of value than the Black-Scholes-Merton equations.

If we introduce the assumption of, for example, stochastic volatility (i.e. that volatility itself follows a random process,) provided that the parameter values are close to empirical, then MC simulations can easily yield more representative results than deterministic equations.

Besides the valuation of financial instruments, like Hanover did, Monte Carlo methods could be employed for money management issues. In theory, given ideal probability density functions and a known edge, the results would be no different than that of optimal F calculations (Kelly criterion). However, from my trading experience, I feel that my winners and losers tend to come in waves. If this is true - and it has to be investigated statistically first! - then this knowledge could be applied to a MC model on money management. The conclusion would probably be that the optimal fraction to risk is much lower than what the Kelly criterion dictates!

To sum this post up: if problems have a low degree of complexity, there probably are some deterministic equations to solve it. However, as the complexity increases, MC methods may provide a relatively simple way to work around it and quickly provide useful results.

On the purpose of Monte Carlo methods in trading (reloaded)

I checked what is available on FF regarding the use of Monte Carlo simulation. Here I found spread sheets for Monte Carlo simulation.
However I think not every system can be modeled using those spreadsheets.

In those spreadsheets it is about modeling system expectancy. So it would be OK if your system has fixed stops and take profit levels so you can calculate exactly the win/loss. By doing that you can model it.

However if you use stop and reverse signals and no stop loss, or if you use a time filter or trailing stop. I think that in those case the system expectancy is not modeled properly.

So there is another way of doing things. It is about selection without replacement. In one article I gathered some information about it. Or you can download the pdf file here.

Unfortunately I did not found a free software doing exactly this. There are some commercial software including excel ad-dons.

So what does it means in practice. In fact the most important aspect of this kind of Monte Carlo analysis is the analysis of the drawdowns. For example with a given confidence level you can say what that worse drowdown can be.

http://www.tradelabstrategies.com/cu...tegiesrev2.pdf

This is a free e-book from Larry C. Sanders from his website tradelabstrategies.com. On this discussion forum you can read an interesting discussion about it.

Larry C. Sanders addresses the topics of probability, marble game and MonteCarlo simulation. He has also designed a software program TradeSim which he sells at his website. The book can be downloaded at this site for free (this is the bookmark link). This book also talks about the Monte Carlo analysis of trades.

I created sometime ago now some FREE trading spreadsheets relating to
Monte Carlo & Mersenne Twister’ Trading Simulators.
Also I created one that incorporated a import utility that imported random data from Random.org
Post & FREE downloads to the excel sheets are here;

Free Excel Day Trading Simulators

Some traders might find some value in my above post
Regards
DowBoy