# Supertrend(10, 3) against a moving average, on crypto: report 2026-11-05

Pre-registration `apps/research/docs/preregistration-supertrend-vs-ma-2026-11-05.md` · code at commit `f71db9e1b23b293b80137fa42566008e7024d5fc`.

## Acceptance test

**PASS.** Snapshot `fd1d67415b8f8408…`; the harness reproduces the published Rule v1 headline exactly: period 2019-01-01 → 2026-10-09 (2839 days), CAGR 42.5%, max drawdown -46.1%, Ulcer 26.69, 682 round trips.

## Outcome

**Ladder N.** In past data Supertrend(10, 3) and a moving average that traded as often were not clearly different in how far they fell. Supertrend's return was higher.

Depth difference (moving average minus Supertrend): 27.3 points. Conditions for "clear": at least 5 points: yes · bootstrap interval excludes zero: no · same sign in both halves: yes · same sign at stress costs: yes.

## The matched moving average

Rule v1 made 682 Core round trips. The grid length with the closest count is **250 days** (871 round trips, a gap of 189) — **matched at the edge of the grid**. Selection used trade counts only.

| Length (days) | Round trips |
|---|---|
| 10 | 5018 |
| 20 | 3217 |
| 30 | 2531 |
| 40 | 2078 |
| 50 | 1833 |
| 60 | 1609 |
| 70 | 1506 |
| 80 | 1428 |
| 90 | 1359 |
| 100 | 1343 |
| 110 | 1275 |
| 120 | 1197 |
| 130 | 1181 |
| 140 | 1114 |
| 150 | 1100 |
| 160 | 1025 |
| 170 | 1002 |
| 180 | 983 |
| 190 | 947 |
| 200 | 926 |
| 210 | 956 |
| 220 | 968 |
| 230 | 953 |
| 240 | 914 |
| 250 **(matched)** | 871 |

## Headline (point-in-time Core basket, equal weight, base costs 0.30%/side)

| | Supertrend(10, 3) | SMA 250 | Buy and hold |
|---|---|---|---|
| CAGR | 42.5% | 13.8% | 38.5% |
| Max drawdown | -46.1% | -73.5% | -85.1% |
| Ulcer index | 26.69 | 57.13 | 59.01 |
| Calmar | 0.92 | 0.19 | 0.45 |
| Sharpe | 1.04 | 0.52 | 0.82 |
| Time in market | 95.2% | 95.0% | 100.0% |
| Longest days under water | 922 | 1980 | 1980 |

## Experience (Core round trips)

| | Supertrend(10, 3) | SMA 250 |
|---|---|---|
| Round trips | 682 | 871 |
| Losing share | 58.2% | 88.7% |
| Average win | 48.7% | 108.7% |
| Average loss | -14.2% | -6.9% |
| Lost within ten days | 12.0% | 68.1% |
| Longest losing run | 17 | 64 |
| Average days held | 36.5 | 27.1 |
| Profit factor | 2.46 | 2.01 |

## Agreement

On 66.6% of 56,819 Core asset-days the two rules held the same position (both in 29.1%, both out 37.4%).

## Episodes (Bitcoin drawdowns ≥ 30 %)

| Peak | Trough | Basket fall | Supertrend avoided | SMA 250 avoided | Supertrend missed of rebound | SMA 250 missed of rebound |
|---|---|---|---|---|---|---|
| 2019-06-26 | 2020-03-12 | -53.8% | 95% | -5% | 72% | 80% |
| 2021-04-13 | 2021-07-20 | -41.9% | 41% | -5% | 49% | 75% |
| 2021-11-08 | 2022-11-21 | -80.7% | 49% | 63% | 46% | 72% |
| 2025-10-06 | 2026-06-30 | -50.2% | 69% | 50% | 38% | 61% |

## Robustness

- Stress costs (0.60%/side): max drawdown -48.0% vs -77.9%; CAGR 38.7% vs 10.0%.
- Execution one day later: max drawdown -49.1% vs -66.9%; CAGR 41.5% vs 20.0%.
- first half (from 2019-01-01): max drawdown -43.6% vs -64.7% (difference 21.1 points); CAGR 69.2% vs 28.6%; holding -84.6% and 49.1%.
- second half (from 2022-11-20): max drawdown -45.9% vs -53.6% (difference 7.8 points); CAGR 20.1% vs 0.8%; holding -58.0% and 28.7%.
- Date-block bootstrap (B=2000, Supertrend minus SMA 250, 95 % intervals): max-drawdown difference (positive = Supertrend shallower) 15.1% [-6.5%, 36.7%]; CAGR difference 28.6% [1.6%, 59.2%]; Sharpe difference 0.52 [0.05, 0.99].

### Neighbours of the matched length

| Rule | Round trips | CAGR | Max drawdown | Ulcer | Calmar | Time in market | Losing share |
|---|---|---|---|---|---|---|---|
| SMA 240 | 914 | 13.0% | -75.2% | 58.12 | 0.17 | 95.2% | 88.0% |
| SMA 250 | 871 | 13.8% | -73.5% | 57.13 | 0.19 | 95.0% | 88.7% |

## The conventional lengths and the crossover (disclosure, not a search)

| Rule | Round trips | CAGR | Max drawdown | Ulcer | Calmar | Time in market | Losing share |
|---|---|---|---|---|---|---|---|
| Supertrend(10, 3) | 682 | 42.5% | -46.1% | 26.69 | 0.92 | 95.2% | 58.2% |
| SMA 20 | 3217 | 32.5% | -59.4% | 30.59 | 0.55 | 95.2% | 76.2% |
| SMA 50 | 1833 | 37.6% | -48.8% | 28.61 | 0.77 | 94.0% | 79.4% |
| SMA 100 | 1343 | 20.7% | -78.0% | 57.19 | 0.27 | 93.4% | 86.6% |
| SMA 200 | 926 | 13.9% | -76.2% | 59.55 | 0.18 | 96.0% | 88.0% |
| SMA 50/200 crossover | 180 | 14.4% | -72.4% | 56.37 | 0.20 | 94.6% | 71.7% |

## Limitations

Crypto history is short with few full cycles; the assets move together, so the effective sample is far smaller than the asset count; assets that delisted are not in the data; results describe the past, and real execution, fees and discipline differ from a simulation. One kind of moving average (simple, price above it) at lengths fixed in advance, and one crossover; exponential and other averages are not tested; matching by trade count is one definition of a fair comparison, not the only one; the two rules hold the same position on most days, so any difference comes from a minority of days; one market and one period; nothing here says which rule will fall less in future.
