Demo
No-login workflow demo
A static product walkthrough showing how one idea moves from backtest evidence to comparison and paper tracking before account creation.
Workflow Guide
ETF Rotation Backtesting
The hard part of rotation research isn't ranking tickers — it's deciding whether a ranking rule survives frictions, benchmark context, and forward drift after the backtest stops flattering it.
Why Alphrex fits rotation research
Benchmark isn't an afterthought
Compare against buy-and-hold, a single proxy, or a weighted basket like SPY:0.60 IEF:0.40 — not one blunt reference series.
Turnover cost stays visible
Frictions are in the core backtest contract. High-turnover mythology is harder to mistake for durable edge.
Validation attached to the result
Holdout and walk-forward summaries sit close to headline metrics — matters when small ranking changes shift the sample boundary.
A realistic rotation workflow
1. Start simple
Begin with a momentum, mean-reversion, or factor thesis. Keep ranking logic interpretable before inventing a large allocation story.
2. Pressure-test in Backtest
Check against realistic costs, a defensible benchmark, and validation splits. Rotation logic that only wins in one uninterrupted sample window usually isn't worth promoting.
3. Use Compare to reject flattering variants
Place competing ideas side by side. Look for differences that survive once active return, drawdown, benchmark fit, and validation quality share the same surface.
4. Track survivors in Paper
Only the variants that survived a real validation screen belong in paper. That's where you learn whether the rotation logic keeps its edge outside the historical sample.
What this does not solve
Execution realism
Alphrex models costs but isn't a live execution stack. Intraday constraints, tax lot logic, or broker routing need separate planning.
Benchmark design judgment
Weighted baskets help — the platform doesn't choose the right economic benchmark for you. Framing the opportunity set is still your call.
Regime certainty
A good holdout result is model output, not proof of regime invariance. Forward paper evidence remains necessary.
Related proof pages
Example
Saved result example
A concrete saved-result memo showing how validation, benchmark fit, and decision framing should sit next to the headline backtest.
Example
Benchmark-aware compare example
A side-by-side comparison example showing why the leader is the strategy with the strongest combined validation and benchmark-relative read, not just the highest CAGR.
Example
Paper and scoreboard example
A forward-evidence example linking saved research, paper tracking, alerts, and scoreboard standing without confusing live ranking with historical backtests.
Start
Run the workflow on your own idea.
Open the no-login demo to see the research path, then create a free workspace.