Quote from riskfreetrading:
Here are some points that seem not to have been covered (at least explicitly):
1. One does not pyramid to make more (less) or lose less (more) but rather to make a compromise between the two extremes, which is in essence to produce a smoother equity curve.
Pyramiding has the great impact on the mean/standard deviation of returns. This is the point that most posts seem to have missed. It is the most important way, in my view, that one should measure the effect of pyramids!
2. If one decides to pyramid, there is an implicit assumption in this: there is a trend and the trend is choppy. So one goes with the trend, but should not commit all the roll as one cannot distinguish between the beginning of a chop and the end of the trend. That is why one should add to the position only if and when the choppy part is found to be not the end of the trend.
3. Given the above I pyramid not in price (but believe it or not): I pyramid in time! I divided my roll in N parts (they can be equal
in size or decreasing in size as a function of time). My understanding of 8s-accountant comment is close to mine on this.
My equity curve will have a nice (mathematically proven) relationship between simple and exponential moving averages.
I cannot say more (sorry trade secrets), on this but the author welcomes comments and PMs from experts or potential users of such tools to discuss them further.
Best regards
Interesting post. Here's my pyramiding philosophy and approach:
IMO pyramiding is done because early in a potential trend, the picture is quite unclear. Because of the lack of clarity, you don't have as much confidence in the trend - it could be just a false breakout, or a temporary blip. So you want a conservative position size to test the waters.
Now let's say the market goes your way, it breaks out to a new yearly high and extends further. You now have more evidence that a potential bull market is unfolding. So, that can justify a bigger position size, since you have more evidence to back your view.
Next, you want to see how the market handles a pullback. Will it bottom out at a higher low, then move back strongly and break to new highs? That's what you want to see. So, if the market can handle the first pullback, then recover to new highs, that's more evidence of a strong trend in place.
After that, I have my maximum conviction in the trend. So, from then on I want to be buying each pullback, as I feel justified in assuming that the trend remains intact. Buying short-term pullbacks exploits the market noise, to my benefit. As long as it keeps making higher highs, and then higher lows on pullbacks, I will keep adding on each dip.
This works only for markets that have the potential to make very large long-term moves. Markets that are potentially going to rally 30, 40, 50%+.
Pyramiding is very bad for smaller moves, since you are highly likely to be adding near the end of the move. E.g. if a market is going from 100 to 85, you don't want to be pyramiding at 92, 88, 85. You will end up with max size at the lows, and get hammered once the move reverses.
Overall, I think it's better not to think in terms of pyramiding, but rather in terms of sizing the position to the risk/reward equation, and your confidence level in the trade. As a breakout turns into a trend, and the trend gets tested and passes those tests, my confidence level increases, thus I add size. Whereas a smaller move, once it's made say 75% of the expected move, the risk/reward is not as attractive as at the start of the move, so you should be scaling out.
Pyramiding works for some situations, and scaling out works for others. If you look at trade odds, expectation, risk/reward, and confidence level, then you will realise which approach is best for any given situation, rather than sticking to one approach dogmatically without and regard for the specifics of the market situation.
