I trade as if price is random. So far I'm really only comfortable doing this on a long term basis (eg. not intraday). Let's discuss methods of trading that don't involve prediction. I hear people say a lot "cut your losers short and let your winners run." When I trade, I can't do that because I don't know if a winner is going to keep being a winner nor do I know when a loser is going to turn around and become a winner, but I discuss my trading system in another thread. Does anyone have any concepts or ideas for non-predicting systems? For example, something that if price happens to go your way turns into a winner, and if not it's not big deal (eg. cutting your losses short) with positive expectancy? I was thinking about something like going long after every green bar. This would fit "adding to winners" that everyone seems to think is such an important rule (although it does not work for me in my own trading system), and if you happened to catch a big trend you'd make a ton of money. But of course your average cost may be so close to price that a small move against you will turn your winner into a loss. If you get it wrong, whatever, you get stopped out for a few points with one contract. If you get right, tons of contracts in the correct direction (unless price reverses at bad point relative to your average price in which case your winner turns into a loser). This still seems like it requires being able to predict if price is going to trend or not, and if you were able to do that it would make more sense to just short with a big position from the beginning. But maybe this will lead to something useful.
At some point, to make money, you are going to have to make a prediction, whether it's on direction, lack of direction, volatility, or both. However, with that said, I like the topic of discussion.
If you follow through with your "random entry, add to winners" idea, it is indeed a prediction. You are predicting that the market will be auto-correlated enough, and the equity run smooth enough, that the system will yield a profit. All action implies a prediction of some kind or another. Not considering the expected value of your initial entry point does not change this fact.
As long as you understand that you are always trading probabilities, I don't think the idea of prediction matters for any one trade. You can't predict that you will have 50% of a win for a given trade, even if your overall winning percentage is 50%. Individual trades are binary, but sets of multiple trades are probabilistic.
My way of trading involves no prediction but just a long or short BIAS. I need to have a bias just to know if I should be long or short. I try to be in the market from open to close most of the time reversing the previous trade. If I make money its just because I happened to be in the market before the price moves. I'm still struggling with the concept of letting your winners run. I don't have a problem in cutting the losers soon. As Jake Bernstein says 90% of the profits will come from 10% of the trades.. Exit is the hardest part of this puzzle.
RIDICULOUS! All trading relies upon "prediction" to some degree and over some time frame. (BTW... not all price is random.)
i cant understand how can one trade without prediction. Bias is prediction - expectation that market will go one side more than other. It is based on past experience. If this and that then market more likely moves up / down. T/L break is a prediction based on stat analysis of T/L breaks indicator trading also prediction, based on some stat. news trading is also prediction - if number xx pts above expectation then blahblah complex discretionary trading based on prediction - stat of past experiences. ???