Risk/return map: simple annual average vs CAGR

Accepted Methodology Research source: Hybrid 20 Aug 2026

System: Platform / methodology research — not tied to a single system.
Git commit: eb87c6e

Hypothesis

CAGR compares unequal test lengths more honestly than total-return-divided-by-years.

Controlled change

Scatter Y-axis switched to CAGR; caption states the fixed-lot understatement trade-off.

Dataset & conditions

Extracted balance series, all 14 systems (2–24 years).

Before / After

Before

RINA simple average (24y, 35× account growth): 179%/yr

After

RINA CAGR, same record: 20.6%/yr

Validation

Out-of-sample, walk-forward, Monte Carlo and forward results are shown only where they exist as data. None exist for this entry.

Decision

Accepted

The simple average produced figures no reader interprets correctly and compressed twelve systems into an unreadable band behind two outliers. CAGR puts 13 of 14 inside 1–28%.

AI involvement

Model: Claude (Anthropic)
What AI did: Identified the simple-average distortion on long compounding records and implemented the CAGR replacement with its stated trade-off.
Human review: Reviewed and approved by a human

AI-assisted entries are published only after human review; entries generated by AI without that review are withheld by the build pipeline. AI does not predict markets, and no entry claims otherwise.

Backtest improvement does not imply future improvement. The more experiments run, the more likely some succeed by chance — this log exists partly so that multiple-testing risk stays visible instead of hidden.