The clone · how it was made

It learned from every KOL wallet, then from what ran.

Every coin of the first 30 minutes, every second at first, then less often, with only what was known at that moment: the coin's own trades, its dev's record, which KOLs were in, and which wallets had a habit of buying right before KOLs, or right before a run. The question for each moment: if the clone buys now, does the coin reach its take-profit before its stop-loss? A gradient-boosted tree model learned the answer from 30 days of every trade.

Tested on a week it never saw, with the live rule

1 SOL became
—
Trades
—
Won
—
Max drawdown
—
Buying blind
—
Copying KOLs
—
KOL in after
—
The week before
—

 

Every setting on the test week

Exits (%, time)LineTradesWon1 SOL becameMax drawdownWeek before

 

What it looks at most

How it was tested