Seems like you'd need a lot less trades than 120 with MC, but that depends on the distribution of trades itself. Hey, I guess that's one advantage of a MC. It takes this effect into account, so perhaps we could say that MC may save money for certain trade distributions. Perhaps you could analyze that dist. I posted with the gaussian methods for comparison?
Activate/Deactivate System?
- Thread starter EricP
- Start date
Seems like you'd need a lot less trades than 120 with MC, but that depends on the distribution of trades itself. Hey, I guess that's one advantage of a MC. It takes this effect into account, so perhaps we could say that MC may save money for certain trade distributions. Perhaps you could analyze that dist. I posted with the gaussian methods for comparison?
a-ha moment:
Any probability distribution of trades (gaussian or not) will yield a gaussian distribution of results after N trades, as long as N is high enough and the trades are independent (central limit theorem). So if you want to compare the mean and variance of of your live trading results with that of your backtested results, you have two problems:
(1) You need a large number of trades N to generate your gaussian distribution of results.
(2) If the trades are not independent in the first place, you have to be careful with how you interpret "mean" and "variance". Your distribution of results is not gaussian.
Using a monte carlo, you get rid of both of these problems at once.
(1) Many less trades are required, since you can now compare distributions directly. By using a set of 100,000 traders, you can see how likely it would be to achieve a certain result with only 10-20 trades.
(2) Since you are not using "mean" and "variance" to compare your results anymore, it is a moot point whether your results are independent or not. You are comparing distributions directly.
This all makes sense to me now... but i'm open to re-education if neccesary.
Any probability distribution of trades (gaussian or not) will yield a gaussian distribution of results after N trades, as long as N is high enough and the trades are independent (central limit theorem). So if you want to compare the mean and variance of of your live trading results with that of your backtested results, you have two problems:
(1) You need a large number of trades N to generate your gaussian distribution of results.
(2) If the trades are not independent in the first place, you have to be careful with how you interpret "mean" and "variance". Your distribution of results is not gaussian.
Using a monte carlo, you get rid of both of these problems at once.
(1) Many less trades are required, since you can now compare distributions directly. By using a set of 100,000 traders, you can see how likely it would be to achieve a certain result with only 10-20 trades.
(2) Since you are not using "mean" and "variance" to compare your results anymore, it is a moot point whether your results are independent or not. You are comparing distributions directly.
This all makes sense to me now... but i'm open to re-education if neccesary.
1. What software are you using to generate MC analysis?
2. Peter, I don't know if you answered EricP's question but how do you activate a system once it has been deactivated using your MC analysis?
Thanks
another Eric
2. Peter, I don't know if you answered EricP's question but how do you activate a system once it has been deactivated using your MC analysis?
Thanks
another Eric
The program is freely available in the files section of the AustinTraders Yahoo group. You have to download all nine data files (1MB each) plus the setup program.
Interesting point: the max drawdown distribution in that example I posted is NOT gaussian. Looks like poisson. If any stat gods out there could explain why I would appreciate it.
The market has not yet pissed on my cheerios, so to speak, so I have no experience with this. I suppose I'll do what Eric does, keep an eye on the last 20 trades or so to see if it starts working again (the results should fall in line with the MC of the backtest).
I do.
I would add, however, that liquidity can affect how many shares of a stock (or futures contracts) can be bought at a given price, without causing additonal transaction costs due to slippage. If you are trading sufficient size to affect your own slippage, then you need to be careful how you implement this. If slippare IS an adverse issue as the size is increased, then a potential fix would be to enter your position using a number of individually small 'increments'. Each increment would be small enough to avoid creating slippage. If a system has been perforning very well, and has a large "x" value, then you can choose to increase the maximum number of increments that you allow that system to hold as it's full position size.
My goal is to ensure that 100% of the system is mechanical in nature. With this sort of system, the systems that are active, as well as the size being traded in any system, are all handled 100% mechanically. Less day-to-day decisions for me equals less stress, fewer worries, and less overall work, IMO.
-Eric
I noticed the distribution table you posted is a standard t-distribution, but I'm not familiar with the above equation as it is not the standard t-test. Where is this formula from?
It seemed to me that this is a test of profitablility and not necessarily a test for conformance to past results, as the above formula is based on a sample of the most recent trades. While I think this is a very valuable formula, I think it poses some problems for certain systems that are working according to design but tend to have a high standard deviation in nature.
For instance, my current "pet" that I am looking to start trades three markets. Each trade starts with the assumption of a day trade with a scaling exit. However, occasionally the trade has such favorable conditions that is converts to a swing trade. Thus, the sample has a very high standard deviation becuase some trades end up holding for several days, while occasionally an unfavorable gap causes a very large loss.
Perhaps a tradiional t-test is the best alternative for testing out of sample data versus historical results.
Any thoughts?
Edit: I suppose what I am saying is the system has "fat tailed" distribution. Therefore this test gives a somewhat poor result.