Proof Page

Benchmark-Aware Compare

Several candidates look plausible in isolation. This shows how benchmark-relative context and validation pick the cleaner strategy over the loudest headline return.

Example candidate set

Strategy CAGR Sharpe WF avg Verdict
Trend sleeve 11.4% 1.11 0.73 Solid — advances to paper
Fast crossover 13.2% 1.05 0.22 Drop — thin holdout, turnover costs underrepresented
Regime-conditioned winner 10.6% 1.18 0.81 Promote — strongest fold stability

Why the leader wins

1. Validation survives the split
Walk-forward average and worst-fold behavior are materially stronger than the fast crossover. Lower headline CAGR, cleaner holdout contract.
2. Benchmark context stays coherent
Active return remains positive without a soft benchmark. Raw outperformance against one easy proxy is not an argument for promotion.
3. Fewest unresolved objections
Compare is useful because it puts objections in the same table. The leader is the strategy that survives that table, not the one that avoids it.
4. Outcome is a workflow action
Leader moves to paper or a saved-result review. Weaker candidates are dropped — not carried forward out of optimism.

What compare does not settle

Forward drift Cleaner historical case doesn't replace the paper layer. Promote and track.
Benchmark choice A technically tighter basket can still be economically wrong. You pick the opportunity set.
Execution feasibility Turnover and cost checks don't replace broker-side implementation design.

Live artifacts

Related proof pages

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.

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
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.