Each account runs on its own ticker and strategy, restarting from its own seed every January, and the accounts are added up day by day. A single-ticker backtest can only tell you that a strategy looked good. What decides a real account is whether something else holds up on the days the main strategy is losing — so the table below also shows what each companion did during the main account's losing stretches.
The first row is the main account. Everything below it is a companion. Up to 6 rows.
| Role | Ticker | Strategy | Seed ($) |
|---|
Searches every split in 10% steps and scores each one together with its neighbours, so a ratio that only looks good at a single point does not win. The first and second halves of the period are also reported — if a ratio is strong in the first half and collapses in the second, it is fitted to the past.
The main account stays fixed; each candidate takes the companion slot one at a time. Two baselines are shown for reference: the companion slot left as cash, and the main account alone.
Pick one ticker and this tries every combination of two and three strategies on it, splitting the seed between them, then ranks pairs and triples together in one list. It answers "if I run several strategies on SOXL instead of one, which ones?" — the section above keeps the main account fixed, so it cannot answer this.
Every account restarts from its own seed each January (the same convention as the yearly table on the backtest page). Continuous figures chain those years together, which is the same as rebalancing back to the seed ratio every January. Fees are charged each way; taxes and slippage are not included. This is not investment advice.