Every score on this site comes from the same process. This page is the contract — every review links back here, and if we ever break one of these rules, tell us and the correction jumps every queue.
How scoring works
Each tool is scored out of 10 as a weighted average of three parts:
| Component | Weight | What it measures |
|---|---|---|
| Correctness | 50% | Does it actually do the job? Outputs are executed and checked, not eyeballed. |
| Quality | 30% | How good is the result — clean, usable, maintainable. |
| Completeness | 20% | Edge cases, documentation, and the details that bite later. |
Scoring bands
| Score | Meaning |
|---|---|
| 9.0–10 | Exceptional — the pick in its category, few caveats. |
| 8.0–8.9 | Excellent — recommendable for most people, with noted trade-offs. |
| 7.0–7.9 | Good — right for specific situations we spell out. |
| 5.0–6.9 | Mixed — real strengths, real problems; read before buying. |
| Below 5 | Not recommended at current pricing. |
The five rules
- No number without a source and a date. Every price, score, and stat says when it was verified. “Last updated” on a page means it — the date is machine-readable, too.
- We pay retail. Tools are bought with our own card at the tier we review. Some outbound links are affiliate links, always disclosed — they never touch a score, and a tool cannot pay to be reviewed or re-scored.
- Investing pages carry no ratings, targets, or advice. Ever. We publish a bull case and a bear case on every covered name and let the reader weigh them. Risk labels describe historical volatility, not predictions.
- Corrections jump every queue. A reader-reported error gets fixed before any new content ships, and the fix is noted on the page.
- No invented statistics, testimonials, or endorsements. If we didn’t test it or can’t cite it, it doesn’t ship.
Re-testing
Scores aren’t forever. Tools are re-tested on major version releases and prices are re-checked monthly; the score history on each review shows every change. When a score moves, the page says why.