The trade I keep coming back to when I think about this is one where I got the entry exactly right and still walked away with less than a quarter of what was on the table. I was long something that ran, I felt smart, I added into it three times on the way up, and then the trend rolled over and gave back most of the paper gains before I could react. My entry was fine. My sizing was the problem. Every add-on I stacked on top pushed my total open risk higher than the day I first put the trade on, so when the giveback came it hit a much bigger position than the one I originally sized for.
Pyramiding done well is supposed to do the opposite. The whole point is to turn one good entry into a large position while keeping the amount you can actually lose from here roughly flat, or even shrinking. If you find yourself risking more of your account near the end of a trend than you did at the start, you have it backwards, and most people who blow up a good trend trade have it exactly that way.
The rule that keeps total risk flat
Here is the mechanic that makes this work. Open risk is not your entry price. Open risk is the distance between where price is now and where your stop sits, multiplied by your size. When you move your stop up behind a rising trend, you are shrinking the open risk on the shares you already hold. That freed-up room is what you spend on the next add. You are not adding new risk to the account. You are recycling risk that the trailed stop just took off the table.
So the sequence looks like this. You enter with your normal position and set a real stop, the one that says the idea is wrong. Price moves in your favor and you trail the stop up to at least breakeven on the first tranche. Now that first tranche can lose nothing, which means you have budget to add a second tranche whose stop, combined with everything already locked in, keeps the total possible loss at or below your original one-trade risk. You keep doing that. Each add is only allowed if raising the stop underneath the whole stack has banked enough prior gain to pay for it.
The number I anchor to is a fixed fraction of the account per trade, something modest, and I never let the sum of open risk across all tranches climb back above it. When I do the arithmetic honestly, that constraint almost always forces the add-ons to get smaller as I go up. Which turns out to be the right instinct anyway.
Why the adds have to shrink
Late pyramids are where the pain lives. The higher you add, the closer you are statistically to the end of the move, and the more the average cost of your whole position drifts up toward the current price. A big add near the top does two ugly things at once. It fattens your position right when the trend is most likely to reverse, and it drags your blended entry up so a normal pullback can flip the entire trade from green to red in a hurry.
A decreasing schedule fixes both. If your first add is smaller than your base position, your second smaller than your first, and so on, your average cost stays well below current price and your largest chunk of size sits at the lowest, safest prices. When the trend finally quits, the giveback lands mostly on the small late tranches, not on the fat early one. A rough schedule I have used and trust looks like this.
- Base entry: full unit, this is your largest piece.
- First add, after the trend confirms and the base stop is at breakeven: about two thirds of a unit.
- Second add, after the next higher-timeframe pullback holds: about half a unit.
- Third add and beyond: a third of a unit or less, and only if the trail has locked in real gains underneath.
The exact fractions matter less than the shape. Every add is smaller than the last, and every add is gated by the stop having moved up first. No stop move, no add. That single gate is what keeps you from doing the thing I did, which was adding on emotion because the chart looked strong rather than because the math had made room.
Timing the adds to structure, not to feeling
The trigger for each add should be a piece of price structure you can point at, not a green day. I like adding when price makes a new leg high and then pulls back to a level that holds, a prior resistance that becomes support, or a moving average the trend has been respecting. That pullback gives you two gifts. It gives you a close, logical stop for the new tranche, which keeps the added risk small, and it filters out the blow-off spikes where you would be adding into the exact bar that marks the top.
When you add on the pullback, your new stop sits just under the level that held. If the level breaks, you are out of that tranche cheaply and the older, lower tranches are still fine because their stops trailed up long ago. If the level holds and the trend resumes, you have more size working at a good price with tightly controlled downside. Compare that to adding on a vertical breakout candle, where your only sensible stop is miles below and the trade is already extended. The pullback add is slower and less exciting and it is the one that survives.
A worked example
Say you risk a fixed slice of your account per trade and you buy a base position with a stop that represents that full slice. The market trends. After the first leg up you trail the stop to your entry, so the base position now risks nothing. You add a smaller second tranche on a pullback that holds, with its own tight stop, and you nudge the shared stop up so the combined open risk still sits at or under that same original slice. Another leg, another trail, a third and smaller tranche on the next held pullback, stop up again. By now you might be carrying three or four times your starting size while your worst case from here is still no more than one normal loss, because every tranche below the current stop is locked in profit.
Then the trend ends, as they all do. Price breaks the level, your shared stop triggers, and you exit the whole stack at once. Because the big size was cheap and early and the expensive size was small and late, you keep the bulk of the run. The giveback stings a little on the top tranche and barely touches the rest. That is the trade I wish I had run the first time.
If you want one thing to carry away, make it the gate. Before every add, ask whether raising your stop has already paid for the new risk. If it has not, you do not get to add yet, no matter how good the chart looks. Size the adds down as you climb, tie each one to a level that held, and let the trail do the accounting for you.