How do you take profit without selling the bottom?

Most memecoin losses are not bad entries. They are good entries that were held through the whole round trip, or sold in a panic at the low. We tested the common exit rules on our own stored price data: Robinhood Chain from 13 August to 12 October 2026, and Solana from 15 September to 12 October 2026. In those windows the dip-buy entry rules we tested lost money, while the crash and old-high tests backed three exit habits.

1. Big memecoins crash fast, so always have a stop

We took every token that freshly climbed through a market cap level, meaning our first reading was under half that level, and checked how often it collapsed afterwards. Robinhood Chain covers 13 August to 12 October 2026; Solana covers 15 September to 12 October 2026.

Read the $30M row slowly. Of 14 Robinhood Chain tokens that crossed $30M, 3 were at or under $5M within 24 hours, and 57.1% were down that far within a week. The sample is small, but the rows below it say the same thing with more tokens. At $5M to $10M, 36-41% fell 83% within a day on both chains.

Look at the last column too. The same tokens doubled within 24 hours 33.3-42.9% of the time. Moves this large in both directions are why we read this test as support for stops and for selling in slices. Research on other markets agrees: stop-loss rules lower returns when prices wander at random, and add value when prices have momentum (Kaminski and Lo, MIT). Across 147 cryptocurrencies from 2015 to 2022, momentum trading with a stop-loss beat momentum without one (Sadaqat and Butt).

2. Let the stop follow the price up

A fixed stop under your entry protects you from the dump that starts right after you buy. It does nothing for the more common case: the token runs, you are up, and then it gives the whole move back.

A trailing stop solves that. It sits a set percentage under the highest price since you bought, and it moves up as the price makes new highs. It never moves down. If the token doubles and then falls by your trailing distance, you sell with most of the gain still in hand.

Two things to decide before you set one:

• The distance. Too tight and normal noise sells you. In our golden-zone test on the same Robinhood Chain and Solana data, a stop about 21% under the entry was hit about three times out of four within a day. On Robinhood Chain, 28.7% of those pullbacks fell through the whole zone between two readings about 30 minutes apart. We did not test trailing stops, but that suggests one needs more room than 21%.
• What it sells. A trailing stop on the whole position is one exit. A trailing stop on what is left after you have sold slices is a better one.

3. At the old high, sell a slice, not the bag

A popular rule says to sell when a token climbs back to its previous high. We tested it. We took tokens that set a high of at least $100K, fell 30% or 50% from it, and then climbed back to within 8% of it. Then we checked what happened next within 24 hours.

Across both chains, in the same windows, a return to the old high was rejected 40-45% of the time. At nearby non-round levels, in the round-number test below, the rejection rate was 34-37%. That is real resistance, about five to eight points of it, and it is still close to a coin flip. On Robinhood Chain, of 249 retests after a 30% drop from a $1M-plus high, 54.6% broke through first and 41.4% were rejected first.

Two more facts change how you use that level:

• Most tokens never get back there. Only 18-54% of tokens returned to the old high after a 30-50% drop, depending on the chain and size.
• When they do, it can take a while. On Robinhood Chain the median return took 5.6 to 20.2 hours. On Solana it was 0.5 to 3.8 hours.

So selling part of the position at the old high is a sensible hedge. Selling all of it there gives up the 49.4-70.0% of retests, depending on the chain and size, that broke through first.

Round numbers are discipline, not walls

The same rule often says to set take-profits just under round numbers like $1M, $5M, $10M and $20M, because "that is where tops form". In our Robinhood Chain and Solana data, in the same windows, they were not special. When a token first approached a level from below, round levels broke through first 61.4% of the time on Robinhood Chain (917 approaches) against 60.2% at nearby non-round levels (1,763 approaches). On Solana it was 63.9% (363) against 61.5% (740).

A limit sell at $10M is still fine as a discipline habit, because it is a price you chose in advance. Just do not expect the price to stop there for you.

Take profit early

The clearest result in our dip-buy test was about time. On Robinhood Chain, from 13 August to 12 October 2026, we tested buying the low after a token's first pump to $1M (196 tokens). Even with a hindsight-perfect entry, a 72-hour hold had a median of -100%. A 24-hour hold from the same entry closed above entry 55.2% of the time, with a median of +4.4%. Holding seven days was worse at every entry.

A plan you can set in two minutes

1. Before you buy, decide the stop. Write it down.
2. Set a limit sell for a first slice at the previous high, if the token has one above you.
3. Set further slices at levels you choose: round caps are fine as long as you know they are habits, not walls.
4. After the first slice sells, move the stop on what is left up to your entry or higher, and keep moving it up after each new high.

In the PumpPill Trade Bot, an Auto-Sell ladder takes rungs at a gain over your entry, at a market cap, and a stop below your entry. Market-cap rungs let you put a slice at the old high in one step. The bot's stop is a fixed one today: it does not follow the price up on its own, so step 4 is still a manual move. Everything in the ladder is your choice and you can change any rung.

What this does not tell you

• Our Robinhood Chain readings are about 30 minutes apart, so wick highs and wick lows are missed. A real stop can fill worse than the reading.
• The old high here is the highest price we recorded, not the true all-time high on the chain.
• Some tokens stopped being tracked before seven days were up, which probably understates the 7-day crash rates.
• This is about 60 days of Robinhood Chain data and 27 days of Solana data, in one stretch of market conditions.

For why the dip-buy entry is the weaker half of the trade, read Does the golden zone work on memecoins?