Thoughts on Compounding/Exponential systems

Hi there,

I have a bit of a problem. Compounding/Exponential systems are good etc, but not if you open more trades at the same time.

Example:
Starting with 10 000$. Risking 5% each trade.

Winning
3 trades starting on the same time will give you 3 x 5% = 1 1500$
Above is not exponential that is linear, exponential is 10 000 x 1.05^3 = 1 1576$

As long as you are winning it is totally fine, but when you are losing it strikes you badly.

Losing
3 trades starting on the same time will give you 3 x -5% = 8 500$
Above again is not exponential it is linear, exponential should be 10 000 x .95^5 = 8 573$

Maybe you will say: mwah, that's not that big of a deal. But in the long run it really hurts you. Because the draw down is always higher then the gain.

What are you thoughts?

-Jysticim

Assume each open trade is a loss when calculating the size of the subsequent trades.

Then you decrease your profits by alot and then its even better sometimes to just use linear compounding.

If you want to have your account very sensitive to this compounding, - and more importantly if you find the time for it, while placing all your orders - you can use your current equity to do the math.
Personally I would find this kind of compounding too painful.
I don't work with long running positions, so, have the luxury to start with 'clean slate' after the orders are closed. I don't normally carry over positions to the next day.
But unless there happen to be a very profitable run during the day, I don't bother much.
Adjust the position sizes day to day works OK for me. During the day I don't compound the profits, nor the losses. Can miss out something here and there, but not having your mind on your entries and exits can cost more than trying to squeeze out the last cents from your account.
So, if 'linear' then let it be, it is still compounding, and that matters in the end.

This compounding is multiplicative so the result doesn't depend on the order of the trades. It depends only on the number of winners and losers. If you start the three trades at the same time you don't know the number of winners. You also want to risk the same amount on each trade bacause none is a priori better or worse than the others. This leads to this system of equations:
Where E is the starting equity and x is the risk. This system has no solution. You can only approximate the solution. The solution requires to know the number of winners and losers. We know this in average: winners = 3 * winrate and losers = 3 * (1 - winrate). You expect the equity after three trades to be multiplied by
You need solving:
The solution is I let you check that if the winrate is 1 (holy grail) the result is 11576 and if the winrate is 0 (always lose) the result is 8573 as you would expect.

This being said and as a side note, the x% compunding is not good with a 1:1 system because one winner and one loser don't cancel out. Whatever x the result is always smaller than the equity before the two trades: the system becomes negatively biaised. And this doesn't depend on the winrate.

@ PipMeUP

Yes that is what I ment to say.