The backtest headline is not enough
A strong CAGR or Sharpe gets attention, but it is not the research conclusion. The saved result needs to show whether validation, frictions, and benchmark context still point in the same direction.
Saved Result Example
A saved result should read like a compact research memo, not like a screenshot of a flattering equity curve. This example shows the type of evidence Alphrex is built to keep attached to one run before it becomes a paper candidate.
Example fit
Use this page when you want to see what a public-facing result page should actually communicate: why the run looked interesting, what the benchmark-relative read says, where validation weakens the story, and what the next action should be.
What this result is trying to prove
The benchmark has to be defensible
Outperforming one soft proxy is weaker evidence than surviving a harder fitted basket. A saved result becomes acquisition-grade only when the benchmark logic is legible.
The next step must be explicit
A result page should end with a decision: compare it against peers, promote it into paper, or reject it as historically flattering but operationally weak.
Example evidence stack
1. Headline read
12.1% CAGR · 1.28 Sharpe · -15.4% max DD. Interesting, but still only a first-pass summary.
2. Validation contract
Walk-forward 5 folds · avg Sharpe 0.84 · worst fold 0.29. The idea survives holdout pressure well enough to stay alive, but not well enough to skip peer comparison.
3. Benchmark-relative read
Active return +3.6% · beta 0.71 · tracking error 8.4% against the stated benchmark, plus a constructed basket that is only modestly tighter. That suggests the edge is not purely a bad baseline artifact.
4. Research action
This is a compare candidate first and a paper candidate second. The saved result is useful because it says what still needs to be disproved.
What would fail this example
A holdout collapse
If the full-sample Sharpe stays high while walk-forward support turns negative or unstable, the result becomes a warning signal rather than an acquisition asset.
A benchmark that is too easy
If a tougher constructed basket closes most of the active-return story, the saved result is no longer demonstrating a durable edge. It is demonstrating benchmark-selection risk.
No path into forward evidence
A saved result without a compare step, paper step, or clear rejection rule is still just a static artifact. Alphrex matters when that memo pushes the workflow forward.