Probability in Trading

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
enigma
Roundtable Fellow
Roundtable Fellow
Posts: 99
Joined: Tue Apr 29, 2003 6:56 am
Location: UK

Probability in Trading

Post by enigma » Wed Apr 30, 2003 10:22 am

Hi everyone,

I have some difficulty in understanding the uses of probability in trading, particularly in some rules which state: never take a trade unless the expected profit is 3 times the expected loss and also with the concept of positive expectancy.

My problem stems from the fact that one cannot precisely know the true probability distribution of the underlying price process even if we have an extremely huge database of historical prices, since we don't have a clear definition for 'huge'. In addition to this, I would think most of the forum members would agree that very few traders (if any at all) has the uncanny ability to predict price movements with an acceptable level of accuracy. Given these points, does this not imply that the rule of taking trades with 3-1 profit potential is like gambling in reality since one is only guessing that the trade is profitable?

Secondly, correct me if I'm wrong, but trading systems are tested for positive expectancy based on historical tests (including out-of-sample tests). Again, even if the test results show a significant positive expectancy, how can we possibly know this will be the case in the future?

At the present, I have come with a simple theory: Confidence in having a positive expectancy in one's system should not really come from historical (statistical) testing, but rather from the philosophy of the strategy.

For example, by following the golden rule of cutting losses and letting profits run, one provides a sound philosophy that 'maximises the chances' of having a positive expectancy for one's system. Of course, one can go into details as to where to cut losses and take profits, but the general idea is there. So in the end, we can never know if our system has positive expectancy, but we can be comfortable and confident if we belief we have a reasonable philosophy that underlies the strategy. I suppose this is the whole business of speculation.

In short, I'm not too sure if trading is in fact a game of probabilities. If this is in fact the case, the benefits of testing the statistical properties of trading systems may be slightly overrated. I hope this issue does not sound too much like academic gibberish. Would appreciate the opinions or criticisms of other forum members on this. Thanks.

damian
Roundtable Knight
Roundtable Knight
Posts: 815
Joined: Tue Apr 15, 2003 8:43 pm
Location: dusseldorf

Re: Probability in Trading

Post by damian » Wed Apr 30, 2003 10:53 am

Hi, here are some of my thoughts.
enigma wrote: in some rules which state: never take a trade unless the expected profit is 3 times the expected loss
I would never try and guess the expected profit target. Sometimes I have seen it done where the trader is trading counter trend. when i was a trainee in a bank I was taught (told) that you determine your stop (if you even have one) by first setting a profit objective. Then your stop was 1/3rd of this amount. How easy, you have a 3:1 RR. The position size was determined by how gutsy you were feeling that day. I decided this was a bad strategy.
but trading systems are tested for positive expectancy based on historical tests ..... even if the test results show a significant positive expectancy, how can we possibly know this will be the case in the future?
You don't, unless your....
Confidence in having a positive expectancy in one's system [comes from] the philosophy of the strategy.

the benefits of testing the statistical properties of trading systems may be slightly overrated.
I am often only casually interested in the details of back testing. It indicates if an idea is (was) good or not-good. Not much more.

enigma
Roundtable Fellow
Roundtable Fellow
Posts: 99
Joined: Tue Apr 29, 2003 6:56 am
Location: UK

Post by enigma » Wed Apr 30, 2003 11:11 am

damian,

Thanks for your thoughts. Can I assume then that you don't think the 3/1 RR rule is of much use?

As an afterthought, having read through certain threads, I am quite surprised/impressed to hear that that there are some systems that produce real-time results that are similar to backtested results. [/quote]

Post Reply