Where DCI could go
The strategy rebuilt from its 17 assets, re-weighted every day by its own rule, over thousands of random futures. Or DCI's own daily moves since the switch to the composite strategy, replayed at random.
- DCI
- Strategy backtest
- Median (P50)
- P10 – P90
- Sample paths
The model leaves out funding and trading fees: since 8 Apr 2026 it gained +5.0% in USD, DCI +2.6%, mostly before the index moved to spot holdings in October, which cut its funding.
Past performance is not indicative of future results. These paths only replay the assets' past daily moves; they are not a forecast or a promise of any return.
In 5 years
| Year | P10 | P50 | P90 | P50 a year |
|---|---|---|---|---|
| Nov 2027 | — | — | — | — |
| Nov 2028 | — | — | — | — |
| Nov 2029 | — | — | — | — |
| Nov 2030 | — | — | — | — |
| Nov 2031 | — | — | — | — |
How it works
The strategy is rebuilt from the daily closes of its 17 assets (Yahoo Finance), weighted by its own rule. Each simulated day picks one real day since 30 Nov 2024, when all 17 assets trade, and moves every asset as it moved that day, so their correlations are kept. The weights are then reset the way the rebalancer sets them: class share × price ÷ 200-day average, scaled down when they add up to more than 1.5×, the rest in cash. As in the oracle, DCI keeps 94% of an up day's gain (the 6% fee) and the whole of a down day. Funding, trading fees and slippage are left out. 2,000 paths are drawn; P10, P50 and P90 are the prices that 10%, 50% and 90% of them stay below on each day.
Each move is drawn on its own, so streaks and changes of regime are not modelled. A short history means a rough picture. This is past behaviour replayed, not a forecast: past performance is not indicative of future results, and DCI can lose value.