# ROT Sizer Task 1: same basket v3 picks, different mechanics (v1, 7 Oct 2026, 12:10 Bangkok)

## Read this first
- **The published v3 basket lines have bad prints in them.** Receipt: majors basket on 2024-04-30 shows +14.6% while all 10 members fell (BTC 63,841 to 60,637). Same on 2024-03-07 (+20.7%) and 2024-02-07 (+19.8%). One bad coin print at +200% lifts an equal weight basket of 10 by 20%. So the old v3 receipts (PF 2.65, 118 trades) are too good.
- Rebuilt clean from A1's CMC daily files (477 one day bad prints dropped; a coin that leaves the list priced at the list cutoff). Clean v3 trades 2023 to Sep 2026: 126 trades, +5.8% avg, median +0.3%, 50% wins, PF 2.02.
- **Clean v3 book, 1 day delay, 1% each way: +2.0% a month (2019 to 2022), +0.3% a month (2023 to today minus 90 days), worst drop -64%, 2025 alone -54%.** The 10% a month figure does not survive clean data and a 1 day delay.
- **No mechanic rescues it.** Every variant checked by year loses 37% to 54% in 2025. The gap is in the picks, not the sizing.
- What helps a little, both eras: "sell when the sector stops leading" (basket 14 day return below majors): +2.4% / +1.3% a month, worst drop -22% / -42% (vs -37% / -64%). Only variant that passes the bar in both eras.
- Top 3 coins per basket: +1.6 points a month in 2023 to 2026, nothing in 2019 to 2022. Weekly re-check: smaller drops (-29% / -50%), about the same return (+1.7% / +0.7%).
- KILLED: top 1 coin (2023 to 2026: -0.7% a month, worst -91%); 1 basket all in on its top coin (-12.5% a month, -100%). Cuts and trails on basket members: no effect. Pyramid: drops cut to -16% / -27% but only 9% of money is ever in.
- Data check vs published lines (daily return correlation): sol_pump 0.999, sol_memes 0.989, base_ai 0.988, eth_memes 0.983, bnb_memes 0.942 (total +1,712% vs +1,796%). Weak: majors 0.85, eth_ai 0.65, bnb_ai 0.65, sol_bonk 0.33. Main cause found: bad prints in the published lines (also eth_ai 2025-02-09: published +16.7%, its 15 members all between -11% and +7%). The platform map (CMC's current platform per coin vs the old snapshot field) may explain the rest.

## Table (A = 2019 to 2022 minus holdout, B = 2023 to today minus 90 days, B2 = 2025 to today minus 90 days; "B vs C0" = points a month vs equal weight all members)
| variant | A 19-22 /mo | A worst | B 23-26 /mo | B worst | B money in | B2 25-26 /mo | B vs C0 |
|---|---|---|---|---|---|---|---|
| L0 v3 basket line, no delay, 0.5% (published method) | 1.9% | -34% | 1.3% | -58% | 25% | -2.9% | +0.9 |
| L1 v3 basket line, 1d delay, 1% | 2.0% | -39% | 0.3% | -64% | 25% | -3.5% | -0.1 |
| C0 v3 coins, equal weight all members (BASE) | 2.0% | -37% | 0.4% | -64% | 25% | -3.3% | +0.0 |
| 1a top 5 coins per basket | 1.9% | -39% | 0.6% | -72% | 24% | -3.5% | +0.2 |
| 1b top 3 coins per basket | 1.9% | -41% | 2.0% | -65% | 24% | -2.6% | +1.6 |
| 1c top 2 coins per basket | 1.6% | -39% | 1.5% | -63% | 24% | -4.0% | +1.1 |
| 1d top 1 coin per basket | 1.0% | -47% | -0.7% | -91% | 24% | -11.0% | -1.1 |
| 1e 1 basket, all members | 5.7% | -35% | -0.7% | -79% | 30% | -7.4% | -1.1 |
| 1f 2 baskets, top 1 coin | 2.7% | -47% | -1.4% | -96% | 26% | -15.2% | -1.8 |
| 1g 1 basket, top 1 coin (all in) | 0.6% | -74% | -12.5% | -100% | 30% | -30.1% | -12.9 |
| 2 pyramid 25% +25% at +20/+50/+100 [all] | 1.1% | -16% | 1.0% | -27% | 9% | -1.2% | +0.6 |
| 3 cut -15% [all] | 1.5% | -36% | 0.2% | -63% | 21% | -2.4% | -0.2 |
| 3 cut -20% [all] | 2.0% | -37% | 0.2% | -64% | 22% | -2.7% | -0.2 |
| 3 cut -30% [all] | 2.0% | -37% | 0.3% | -64% | 24% | -2.9% | -0.1 |
| 4 trail 25% [all] | 1.5% | -37% | -0.1% | -63% | 20% | -2.5% | -0.5 |
| 4 trail 30% [all] | 1.5% | -37% | -0.1% | -66% | 22% | -2.6% | -0.5 |
| 4 trail 35% [all] | 2.0% | -37% | 0.2% | -65% | 23% | -2.7% | -0.2 |
| 4 trail 40% [all] | 1.9% | -37% | 0.5% | -61% | 24% | -2.6% | +0.1 |
| 4 half at 2x [all] | 1.7% | -37% | 0.0% | -64% | 24% | -3.3% | -0.4 |
| 4 time stop 14d under +10% [all] | 1.9% | -34% | 0.6% | -61% | 22% | -2.8% | +0.2 |
| 4 sector stops leading (14d < majors) [all] | 2.4% | -22% | 1.3% | -42% | 18% | -1.5% | +0.9 |
| 5 baskets weekly re-check [all] | 1.7% | -29% | 0.7% | -50% | 21% | -2.3% | +0.3 |
| 2 pyramid 25% +25% at +20/+50/+100 [top3] | 1.3% | -15% | 1.6% | -28% | 9% | -0.8% | +1.2 |
| 3 cut -15% [top3] | 1.8% | -41% | 1.2% | -70% | 22% | -1.9% | +0.8 |
| 3 cut -20% [top3] | 1.8% | -41% | 1.4% | -69% | 23% | -2.8% | +1.0 |
| 3 cut -30% [top3] | 1.8% | -41% | 1.4% | -69% | 24% | -2.7% | +1.0 |
| 4 trail 25% [top3] | 1.7% | -41% | 1.1% | -71% | 22% | -2.6% | +0.7 |
| 4 trail 30% [top3] | 1.6% | -41% | 1.5% | -72% | 23% | -2.6% | +1.1 |
| 4 trail 35% [top3] | 1.9% | -41% | 2.1% | -70% | 24% | -2.3% | +1.7 |
| 4 trail 40% [top3] | 1.9% | -41% | 2.1% | -65% | 24% | -1.7% | +1.7 |
| 4 half at 2x [top3] | 1.7% | -41% | 1.8% | -64% | 23% | -2.5% | +1.4 |
| 4 time stop 14d under +10% [top3] | 2.0% | -40% | 1.8% | -66% | 24% | -2.6% | +1.4 |
| 4 sector stops leading (14d < majors) [top3] | 2.3% | -25% | 1.1% | -61% | 18% | -2.6% | +0.7 |
| 5 baskets weekly re-check [top3] | 1.0% | -31% | 2.4% | -49% | 21% | -1.5% | +2.0 |
| 5 coins daily re-rank [top3] | 1.4% | -46% | 1.6% | -69% | 25% | -2.5% | +1.2 |
| 5 coins pick at entry only [top3] | 1.7% | -47% | 1.9% | -69% | 25% | -2.0% | +1.5 |

Version: v1, 7 Oct 2026 12:15 Bangkok. Code: rot-box /opt/rot/sizer/ sz1.py (md5 f70c01ecdff936f20b541d046c8f5d90). Raw: http://159.223.88.168/hx-c9ois/rot/sizer-t1_report.json . Latest: http://159.223.88.168/hx-c9ois/rot/sizer-t1.md