DCA vs grid vs rebalance when the market refuses the story

Each Captay strategy assumes a kind of boredom. DCA assumes you are willing to buy on a clock. A grid assumes price will visit both sides of a box. Rebalance assumes you want the mix more than you want the winner. Markets are allowed to be boring. They are also allowed to trend, gap, and make one coin the whole story. This note walks through those refusals without turning any of them into a backtest.

A decline that does not bounce

Picture a coin that is lower every month for a long stretch. A DCA schedule keeps spending the quote amount you named. The average price falls, which feels like progress on a spreadsheet and is actually a larger pile of a cheaper coin. If you added the optional extra buy after each drop, the pile grows faster. Nothing in the arithmetic announces a floor. The position cap is the floor you are allowed to care about, and it only helps if you set it before the stretch, at a size you can still look at.

A grid in the same decline fills the lower rungs and leaves the upper rungs untouched. You are long. The ladder diagram did not say 'long' in large type, but the fills did. Dragging the range down to 'follow' the market is a new strategy wearing the old name. The design says stop at the edge. Stopping feels like inaction. Inaction is the point of a limit.

Rebalance in that decline depends on whether the falling coin still has a target. If it does, the rule may buy more to restore the weight, which rhymes with DCA and should scare you if you already run both. If cash has a large weight, the rule may sell other holdings into cash and buy less of the decline than a fully invested mix would. The weight is the decision. The rule will not notice that you meant the weight only for quiet months.

A range that stops being a range

Grids are drawn on a past that looked oscillatory. The next month can leave through the ceiling and not return. Sells fill, the coin is gone, and the quote balance looks like a win until you price the coin you no longer hold. The grid did the job you gave it. The job was not 'stay exposed to the upside outside the box.' If that was the job, the grid was a mismatched tool, not a tool that malfunctioned.

Fees make the pretty case worse. A completed round trip inside the box can still lose if the step is thinner than the fee. A broken range does not refund the fees of the trips that worked last week. Paper mode, when it exists, is where that arithmetic should be seen with no order transmitted. It does not exist on this site, so the arithmetic stays qualitative: narrow steps and wide fees do not compose into income.

One coin doubles

Suppose one holding doubles and the others do not. Rebalance sells some of the double to buy the laggards and to refill cash if cash has a target. You have realized a gain and reduced the position that caused it. Over a longer run of the same doubling, you will have sold many times. That can be exactly the diversification you wrote down. It can also be how you exit a trend in slices and then describe the mix as disciplined while a single untouched holding would have been larger. The rule cannot prefer your unspoken favorite.

A grid on that same coin may have sold even more of it, at the rungs, before rebalance looks at the weight. DCA, if it was still buying a different schedule, may have been adding on the way up and then stopped only at a cap. Three rules, three inventories, one coin. The collision is not a bug report. It is a configuration. The journal that would show the three streams side by side is part of the closed cabinet, so you cannot inspect a collision here. You can only decide not to stack rules you have not watched.

The brake is not inside the story

Limits sit outside the strategy so the strategy cannot talk its way past them. A DCA rule does not get to say the decline is temporary. A grid does not get to say the range just needs to be wider today. Rebalance does not get to say the drift is the point so the daily loss should be ignored. When the slice is down by the amount you named, new orders stop. The story can resume only if a person re-arms. Automatic resume would launder the limit into a pause.

Nothing in these sketches is a recommendation to run the uncomfortable case, or to avoid it, or to size it. There is no sample account, no percent gained, and no percent lost, because a made-up percent would be a result. The cabinet cannot run the sketches. If you want a rule that promises to notice the regime and switch, you want a different product, and you will not find that promise restated here under a softer verb.