Money management based on normal distribution

Ok so I was thinking about some money management strategies today and i had the following thoughts. Please let me know if my logic or mathematics could be flawed.

If we build a distribution of returns, we will generally find that they are approx normal. If the mean return is say 3% with a S.dev of say 0.5%, and we get a return of 8% on 3 consecutive trades, there may be 0.0005% chance of this happening, according to the CLT. The assumption is that returns are INDEPENDENT, and each event is purely random and does not have any influence on the events prior or proceeding. My question; is this in fact an accurate assumption? When we trade a particular strategy, and for hypothetical sake, the system works on the existence of some sort of trend or momentum, than these returns, CANNOT be independent. A return on one trade could be completely dependent on the prior event.

For example:

We take a trade long on the AUD/USD, then exit the trade on a trailing stop, the process is then repeated and the same result is attained. These two trades may have spanned over a total time period of 3 months. Going forward, any long trade, may now have a reduced probability of success, as the momentum dynamics are no longer intact. When you think about it, this is not such a far fetched proposition. If we were to run a scan on the average price movement of the AUD/USD before it experiences a significant change in the reverse direction, we would probably find that the trades are dependent on this horizon.

SO, if we have a winning long trade on the AUD/USD and have not seen a major reversal in price action, our expectation of returns should drop, for the long side. If this is the case, we should be scaling our money management based on the type of strategy we have, and the phase of the pair we are looking at.

So should we reduce our risk from say 1% to .6% to represent this change in expectancy, and then increase it given the more losing trades we have. So we have a dynamic MM strategy, based on the expectancy of the returns?

I hope that all made sense lol.

this will help us to calculate the risk :slight_smile:

http://forums.babypips.com/other-trading-tools/53775-spidypips-lot-calculator.html#post487702

I have a lot calculator, thanks anyway :smiley:

But do you have any views on what i posted?

If I understand you correctly, we should increase our risk after consecutive losses, based on the assumption that winners and losers are normally distributed, and so we can expect the likelihood of a winner will increase after a loss?

I also recommend Spidypips Lot calculator