How the picks are made
The universe is the 60 Binance USDT coins with the most trading volume over the last 30 days, at least 200 days of history and $2M a day traded. Stablecoins, fiat, wrapped and gold tokens, leveraged tokens and memecoins are left out. Each coin gets a 0-1 rank on nine factors, and a weighted sum of the ranks is its score:
- 30-day momentum: the return over that window (coins that kept rising tend to keep rising for a while).
- 90-day momentum: the return over that window (coins that kept rising tend to keep rising for a while).
- 180-day momentum: the return over that window (coins that kept rising tend to keep rising for a while).
- 1-year momentum: the return over that window (coins that kept rising tend to keep rising for a while).
- Trend vs 200-day average: how far the close is above or below its 200-day average.
- Near its 52-week high: how close the price is to its highest close of the last year.
- Low volatility: calmer coins rank higher (90-day volatility of daily returns).
- Liquidity: bigger, more traded markets rank higher; they are easier to exit and less likely to die.
- Years listed: coins that have survived longer rank higher (capped at about four years).
Each horizon starts with its own weights: short holds lean on 30- and 90-day momentum, long holds on 180- and 365-day momentum, liquidity and age. A coin is only picked if it closes above its 200-day average; high-ranked coins below it go on the watch list. The five best become the week's picks.
Entry, take-profit and stop-loss
Entry is the Monday 00:00 UTC daily close, the moment the week is locked. The stop is 1.5 expected moves below the entry (1.5 × the coin's daily volatility × √days held), between 8% and 50%, and it is checked on daily closes only. The take-profit is 2× the stop distance above, a resting sell order. The pick ends at the take-profit, at the first daily close at or below the stop, or at the close on its end date. This is version 2: the first version used a tighter stop that any intraday touch triggered, and it was stopped out far more often (see how v2 was chosen). Results include a 0.1% fee on each side and 0.02% slippage. During the week the cards show the live price, what is left to each level from now, and whether the price is still in the buy zone.
How it learns
When a week's holding period ends, the model scores every coin of that week's universe with the same take-profit and stop rules and measures how well each factor ranked the winners. Factors with a good record gain weight, poor ones lose it, recent weeks count more, and the record never sets more than 70% of the weights. The model started with 374 finished weeks of history from the backtest and keeps learning from every live week. The current weights and what changed are on the track record page.
Is it worth following?
The honest answer from the test is: not instead of Bitcoin. The picks made money on average at 1 month, 2 months, 3 months, 4 months, 6 months, 1 year, but holding BTC for the same days made more at every horizon, mainly because most altcoins lost ground to Bitcoin over these years. The model is published with that result because the record keeps growing and the learning keeps adjusting; check the validation before using any list, and compare with the Trend Breakout setup, the only DanioX setup that passed its own test, and with trend holds, the same rules used as a holding rule on this universe.
Use it responsibly
These picks are automated and educational, not financial advice. Size every holding so its stop costs no more than you planned, spread across several picks, and never invest money you cannot afford to lose. Read the disclaimer.