I have a strategy which has a specific entry, stop, and exit (for a gain). Backtesting it, when I average for the same share amount at a specific percentage away from the entry during a loss, the system becomes more profitable. And profitability increases the lower the percantage is away from the entry point (ie. average at 20% away from entry is better than 80% away). Here's the question. At 0% away which means doubling the initial position, I get twice the profit obviously. If none of the averaging techniques greater than 1% away are more profitable than 0% away, would that mean its pointless to average? Because if I just take twice the shares to start with, I'll make more money than any averaging technique. Also, the ratio of standard deviation of daily profit to average daily profit for averaging is larger relative to that of a double initial share amount. Although that does not determine how profitable a system is, I prefer to have a small ratio (hopefully less than 1) Is this obvious or am I am thinking too much? Ya, I could take double the initial shares AND average, but then I might as well take 4 times the initial shares. Then I could average that, but I might as well take 8 times the shares. You get the point. edited: to add some details and correct some things I missed
There is no free lunch. If you're going to average losers, you have to take 1/2 the position on your first entry, and then take the second 1/2 if oppotunity allows you to do so. This way you will always stay within the appropriate risk parameters of your trading model. Please don't take my word for it, here's a guy who believes in averaging losers also. Why You Shouldn't Average Losers JJ P.S. Do not think of profit when trading, think of managing risk, this will ensure your survivial for a very long time. Then, provided your system has good R:R and Expectancy, you cannot help but succeed as a trader ... remember 90% of the initial retail traders who try, FAIL. You wanna guess why?
You're right. It is more a question of risk management than profitability. That is what I originally thought but didn't wanna make my OP even longer. And having thought about it further, since I enter using limit orders only I figure its better to post 10,000 shares at one level rather than 5,000 here and 5,000 somewhere else cause it could help to create somewhat of a resistance/support level. I only do lots of 1000 now but I have to think about the future when my risk level allows me to take 10,000.
Definitely agree with JJ. Averaging down is simply an excuse for refusing to take a small loss (or whatever the size but if bigger than small then you got two problems now), nothing else. Smart money management is not. Anek
Why stick that in my face? I have told you it is a habit I picked up I am fighting to get rid of. Why do you say I believe in it?
Defensive rolling up or down is a form of averaging down. Also, I find it impossible to enter the trade at the absolute best price. Trading in rungs seems to help this. Both seem to be a form of averaging down. Just like everything else is trading... managing the position, anticipating and planning for adding the additional load is risk management. I always hate the absolute NEVER do this or ALWAYS do that. Each system has different component necessary to make it work in total. Anticipating and preparing to add 10 rungs to a position is part defense and part offense... even if you can't get all 10 on. Just one man's thoughts. Peace and gtty, Lar
I can look at it as adding onto a position at a better price. Knowing that any one of my system has a 75-80% win rate, adding on an additional lot of shares at below (or above if I'm shorting) my entry does indeed increase profitability. But not enough so that it generates more profit than initially doubling the position. I will experiment by backtesting using 2 times the shares, or 3 times the shares. As well as averaging in 1/2, 1/3, 1/4 shares at multiple prices. Not that I would do this in realtime, but it would be interesting to see if if a 2x average can generate more than taking 3 times the initial position, and so on. So would you guys consider an average down strategy to be more profitable if it generated higher net than an original position with the same amount of shares you could possibly risk? (that may not make sense the way I said it) I guess thats how I would and should think of it. But I have yet to make a system that shows that result. I don't want people to think I'm set on averaging down. I'm just looking at for ways to increase profitability. I'm not worried about risk management here because these are short term day trades on equities which have ranges of +/- 10 to 40 cents at most and I make over 50 trades a day (so far)... unless I had 50 bad trades in a row... but that's what a daily stop loss is for.
i developed a high freq emini strategy that relies on martingale. without averaging it'd be worthless as most initial entries are piss poor. very good results so far in sim....only catch you need to stomach heavy running losses.
yeah, i hold a similar view of it all... instead of playing all your cards in one hand you play 'em one by one seeking to place them each at its most advantageous spot over time. it's rare when playing high freq, thus with a loose attitude, to enter the mkt at the right price yet this weakness can be exploited and turned it into a strength by improving the avg price and reducing immediate exposure to losses. i have to thx mschey for opening my eyes to this and offering one powerful and priceless suggestion. edit; by the way, why give a damn about what folks think of it. if it works... it works, period: some of the most despised methods can prove to be the most revealing.
You nailed it. There is a very meaningful difference in backtesting an average down strategy than in actually trading it. For example, a strategy that backtests buying weakness, and shows profitablity, will often show increased profits by averaging down on entries. But trading them is another story, particularly when you have a day of overall market weakness.