Copytrading
How do I set up a copytrade bot?
Configure it in the opposite order to the one that feels natural: size first, exits second, filters third, and the trader last. The trader is the part everybody agonises over and the part that matters least, because a good trader copied at the wrong size still ends the account. Start with one bot, one trader, and an amount you would shrug at losing twenty times in a row.
1. Decide the size, as a number
Pick the SOL amount per trade before you look at a single trader, and pick it by asking what twenty consecutive losses would do to you. That is not a pessimistic scenario in memecoins, it is a Tuesday. If the answer is uncomfortable, the number is too big. Write it down, because the temptation to raise it arrives immediately after the first good week and that is precisely when it is least justified.
2. Set the exits before the first fill
- A take profit, so a win becomes money rather than a story about a position you watched round trip.
- A stop loss, sized so that hitting it is survivable and not so tight that ordinary volatility triggers it.
- A maximum hold time, which is the exit that saves you from the token nobody is talking about anymore.
3. Set the filters that run before you do
A liquidity floor is the single highest-value filter, because the token you cannot exit is a worse outcome than the token that goes down. A market cap band is the second: it encodes what kind of trade you are actually trying to make, rather than leaving it to whatever the trader felt like that day. Filters run server side before a card is ever considered, so they cost nothing at fill time.
4. Then pick the trader
Now, and only now, choose who to follow. Use realised profit over a window that matches your horizon, read their closed trades rather than their best one, and prefer somebody with many small results over somebody with one enormous one. A trader whose edge is a single outlier is not an edge you can copy.
5. Let it run, then read your own fills
Do not adjust after three trades. You need enough closed positions for the average to mean anything, and the honest number is closer to fifty than to five. When you do review, review your fills and your slippage, not the trader's published results. Their numbers were never yours: you entered later, at a different price, on different exits.
Common questions
- How much should I set per trade?
- Small enough that twenty consecutive losses would be annoying rather than damaging, because that sequence is entirely normal in memecoins. Fixing one number and leaving it alone matters far more than picking the perfect one.
- Should I use a stop loss on a memecoin?
- Yes, with the caveat that a stop on a thin token can fill well below where you set it. That is an argument for a liquidity floor on entry rather than for trading without a stop, since the alternative is holding a position with no exit condition at all.
- How many trades before I know if it works?
- More than feels reasonable. Under about fifty closed positions you are mostly reading noise, and the instinct to change settings after a bad run of five is the most reliable way to never accumulate a sample worth judging.
- Can I change the settings while a bot is running?
- You can, and the discipline is to decide in advance what would justify it. Changing size or exits mid-run resets whatever sample you were building, so it is worth separating a real correction from a reaction to the last three results.
Last reviewed