I keep meeting people who spent months planning their way into a position and roughly four seconds planning their way out. The entry got a spreadsheet, a DCA schedule, a thesis document. The exit gets a vague intention to take profits on the way up, which in practice means staring at a green candle, deciding this time is different, and selling nothing until a drawdown forces the decision at the worst possible moment. If you are holding a crypto position that is large relative to your net worth, the exit deserves the same rigor the entry got, and it needs to be written down before the market starts talking to you.
There are really only two mechanical frameworks for unwinding size, selling on a clock or selling on a price grid, and the interesting decisions are in how you combine them.
Selling on a clock
DCA out is the mirror image of DCA in. Pick the total amount to sell, a schedule, and a tranche size, then execute regardless of price. Ten percent of the position on the first of every month for ten months, or two percent every Friday for a year. Price on the sell date is irrelevant by design, and that is the entire point.
The strength of the clock is that it always finishes. There is no version of events where you end the year still holding everything because your targets never printed. You also receive the average price across your window, which sounds mediocre and is historically a perfectly respectable result, because very few people distributing a large position beat the average once you count all the times they froze and did nothing.
The weakness is that the clock is information-blind. If the market drops forty percent in month two, you keep selling into the hole, because that is the deal you signed. Some people soften this with a floor rule, something like skip any tranche below a set price and append it to the end of the schedule. That is fine, but keep it to one rule. Every conditional you add is a door that discretion walks back through, and discretion is the thing this whole exercise exists to remove.
Selling on a grid
A price ladder is a set of limit sells resting above the market. Fifteen percent of the position at one level, fifteen more at a level thirty percent higher, twenty at the next, on up the curve, so you only ever sell into strength. If the market rips through your rungs, you distributed into the euphoria instead of round-tripping it, and it is psychologically far easier to honor a limit order that fills during a pump than to market-sell on a quiet Tuesday because a calendar told you to.
The failure mode is just as clean. A ladder only executes if price cooperates. If your first rung sits twenty percent above spot and the move never comes, you sold nothing, and you ride the full position down holding a plan that technically never triggered. I have watched people do exactly this and describe it afterward as bad luck. It was a ladder with no floor under it, and the outcome was baked into the structure from day one.
Ladders also need honest spacing. Run one test before you commit: would this ladder have fully executed at any point in the asset's actual price history? If completion requires a new all-time high by a wide margin, you have placed a wish on the order book, and you should respace the rungs until the answer is yes.
For most large positions I think the honest answer is a split. Put roughly half on the clock, guaranteed to complete whatever happens, and half on the ladder for upside participation. The clock half protects you from the top never arriving. The ladder half protects you from the specific misery of selling an entire monster move at the average.
Pick your tax lots before you pick your prices
If you accumulated over years, the position is a stack of lots with different cost bases and holding periods, and which lots you sell changes the after-tax result of an identical trade. In the US, lots held longer than a year typically qualify for long-term capital gains treatment, which is meaningfully cheaper than short-term rates. Selling a lot you bought eleven months ago instead of one you bought three years ago can turn the same sale price into a very different tax bill.
Two things follow. First, use specific identification if your records and your jurisdiction allow it, and check what your exchange or tracking software defaults to, because many default to FIFO, and the lot selection generally needs to be documented around the time of sale rather than reconstructed months later. Second, match the schedule to your tax year. A distribution that straddles the year boundary splits the gain across two tax years, two sets of brackets, and two chances to offset with losses elsewhere. And if some lots are underwater, selling those first realizes losses that can offset gains from the rest of the unwind.
None of this is tax advice, and the rules move around by jurisdiction and by year, so an hour with an accountant before the first tranche is worth more than anything I can put in a blog post. The point is sequencing. Decide the lot order and the tax-year split when you write the plan, because nobody does careful basis math mid-rally.
The rip, and the plan that survives it
The hardest moment in any exit comes when the third tranche fills and then the asset doubles. The drawdown is easy by comparison. Every unit you already sold now reads as a mistake, the plan reads as a tax on your own conviction, and the urge to cancel everything and let it ride becomes almost physical. This is the exact moment the plan exists for, so decide in advance what you are allowed to change.
My rules, for what they are worth. No amendments while the position is up big on the week. Any amendment takes effect after a cooling period of a couple of days, never immediately. Amendments may stretch the remaining schedule or raise the remaining rungs, but they never touch tranches already executed and never reduce the total amount to be sold. And when the regret gets loud I go back to the arithmetic. In a rip, every remaining rung fills at better prices, and the unsold half of a hybrid plan is exactly the upside participation I paid for by accepting the clock on the other half. The plan is doing what it was designed to do, and the discomfort is part of the price.
Here is the whole thing as a checklist. Put real numbers in it and keep it somewhere you can see it.
- Total amount to sell, written in units of the asset, not dollars.
- A hard completion deadline, and the split between clock and ladder.
- Clock side: tranche size, exact dates, and the single floor rule if you want one.
- Ladder side: rungs, sizes, and a check that the ladder would have completed somewhere in the asset's real price history.
- Lot order: which tax lots go first, and which tranches land on each side of the tax year boundary.
- Amendment rule: the cooling period, and the commitment that the total to be sold never goes down.
- A destination for proceeds, because a sell without a destination is usually just a future re-entry.
On execution, resting a ladder across several venues and remembering which tranche fires on which date is exactly the kind of babysitting that erodes discipline, which is part of why we built price alerts and non-custodial execution across exchanges into Blockcircle. But the tooling matters less than the document. One page, real numbers, a deadline, a lot order. Write it on a boring day, and when the loud day arrives, do what the page says.