Loading...
Each day the model predicts the top 10 coins most likely to outperform over the next 3 days. A "hit" = coin appeared in the real top 10 gainers 3 days later.
Backtest of these exact weights on held-out CoinMarketCap data (Oct 2025 – Sep 2026, 362 days): 17.4% hit rate vs 5.7% for random picks.
Buy all 10 picks in equal amounts at the call price, sell after 3 days, minus 0.3% trading costs per coin. Compared with the equal-weight average of the ~200 coins the model chooses from, and with Bitcoin, over the same days.
Only 7 graded calls so far: too early to judge.
| Calls | Avg return | Top-200 avg | Bitcoin | In profit | Beat market | |
|---|---|---|---|---|---|---|
| All graded calls on this site | 7 | -6.2% | -2.0% | -0.7% | 0% | 0% |
| 🟢 Made on TRADE days | 1 | -8.3% | -5.2% | -2.5% | 0% | 0% |
| 🟠 Made on CAUTION days | 6 | -5.8% | -1.5% | -0.4% | 0% | 0% |
| Backtest, held-out year (2025-10-02 – 2026-09-27) | 360 | +0.8% | -0.3% | -0.2% | 51% | – |
The median call earned -4.6%: most of the gain comes from a few big winners (best call -3.0% on 2026-10-20, worst -10.1% on 2026-10-24), so holding all 10 in equal size matters more than picking favourites. Calls before Oct 20, 2026 came from earlier model versions. How we tested this: proof page. Past results do not guarantee future returns. Not financial advice.
Click any entry to expand. RSI tags: red=overbought (>70), green=oversold (<30).