Every review on this site scores the same five criteria, weighted the same way, on the same 1 to 5 scale. No review gets a custom criterion invented to flatter it, and no score rests on what other review sites say.
This page is the full system. If a rating anywhere on the site can’t be reproduced from the rules below, that’s a defect, and you should tell us.
The five criteria
Every product, whether it’s a $150-a-month scanner or a free broker app, is scored on these five, each from 1 to 5 in half-point steps.
Core capability (40% of the overall). What the product category exists to do, and nothing else. A journal doesn’t earn core-capability points for pretty charts; a broker doesn’t earn them for a nice mobile login screen. The exact definition changes by category (table below) because “good at its job” means something different for a scanner than for a short-locate desk.
Value (20%). Price against capability, and against the documented prices of real alternatives. Value is scored for the product’s intended user, the one named in the verdict box’s “Best for” line. A professional scanner isn’t marked down on Value because a beginner shouldn’t buy it; that mismatch lives in the “NOT for” line, where it actually helps you, and is never double-counted into the score.
Ease of use (15%). Learning curve, interface, onboarding. Some of the most capable tools in this market are genuinely hard to learn, and the score says so without pretending difficulty cancels capability.
Trust and transparency (15%). Billing and cancellation terms, marketing honesty, data practices. A platform that auto-renews an annual plan with no refund window is handled differently than one with a 14-day window, and the review tells you which protection play to run (billing cadence, renewal reminder, written cancellation confirmation) either way.
Support and education (10%). Documentation, help channels, training material, and whether you can get a human when the software breaks at 9:28 premarket.
Reviews never add a sixth criterion. If something matters and doesn’t fit these five, it goes in the body of the review as analysis, not into the math.
What core capability means per category
| Category | Core capability is scored on |
|---|---|
| Scanners | Scan speed and alert quality |
| Brokers | Execution, routing, and locates |
| Trading journals | Analytics depth and import coverage |
| AI tools | Signal quality and transparency |
| Charting platforms | Charting and automation |
One consequence worth spelling out: a direct-access broker is scored on fills, routes, and borrow inventory, not on whether its app looks like Robinhood. A scanner is scored on how fast it surfaces stocks in play and how usable its alerts are, not on its charting. Cross-category griping (“the journal’s charts are worse than TradingView’s”) is exactly the kind of noise this table exists to filter out.
The weights, and why a flat average fails
The overall rating is a weighted mean, published to one decimal:
| Criterion | Weight |
|---|---|
| Core capability | 40% |
| Value | 20% |
| Ease of use | 15% |
| Trust and transparency | 15% |
| Support and education | 10% |
A flat five-way average punishes powerful professional tools and rewards bland simple ones. A barebones app that does little, costs little, and confuses nobody would coast to a decent flat average while a serious platform gets dragged down for having a learning curve. Weighting core capability at 40% fixes that: the thing the product exists to do counts most, and everything else is context.
A worked example. A scanner scores Core capability 4.5, Value 3.5, Ease of use 3.0, Trust and transparency 4.0, Support and education 4.0. The flat average is 3.8. The weighted overall is (4.5 × 0.40) + (3.5 × 0.20) + (3.0 × 0.15) + (4.0 × 0.15) + (4.0 × 0.10) = 1.80 + 0.70 + 0.45 + 0.60 + 0.40 = 3.95, published as 4.0. The half-point Ease deficit costs it less than its category-leading scan engine earns it. That’s the intended behavior.
Every review lists all five sub-scores in its verdict section, so you can re-run the math yourself and reweight it for your own priorities. If support matters more to you than it does to us, the raw numbers are right there.
Scale anchors
The numbers mean the same thing on every review:
| Range | What it means |
|---|---|
| 4.5–5.0 | Category-defining. No significant flaws |
| 4.0–4.4 | Excellent, with real trade-offs |
| 3.0–3.9 | Good, with serious caveats |
| Below 3.0 | Hard to recommend |
There’s no grade inflation lever here. Most products land in the 3s and low 4s because most products involve real trade-offs, and a wall of 4.8s would tell you nothing.
Where the scores come from
Every sub-score, pro, con, and judgment on this site must be a stated logical conclusion from documented facts: features and policies verified on the product’s official pages, verifiable price and feature comparisons across products, or original calculations we show in full. The fact is cited, the logic is shown, and the conclusion is ours.
What that rules out:
- First-person testing theater. We don’t claim hands-on trials we didn’t run. The voice here is the analyst who has read every official document, not the influencer performing a demo. Where a review states what a feature does, that statement traces to the product’s own documentation, checked on the date in the byline.
- Review-site sentiment as evidence. Star ratings on aggregator platforms are gameable in both directions, and most articles in this niche copy each other rather than verify anything. A criticism that recurs across ten blogs is a lead to check against official documentation, not a fact. If we can’t confirm it at the source, it doesn’t appear, hedged or otherwise.
- Norm-blind nitpicking. A documented policy counts as a con only if it sits verifiably below the category norm, which we establish by checking peer policies during research. If nobody in a category refunds monthly plans, no product gets dinged for not refunding monthly plans; the review states the category fact, your stakes, and the protection play instead.
- Treating product structure as a flaw. Tiering, feature gating, and add-on pricing are architecture. The review’s job is to tell you which tier you actually need, and structure enters a judgment only through verifiable cross-product comparison, such as one platform gating features by tier while a rival includes everything and gates capacity.
The same standard applies to praise. “Best AI scanner” is somebody’s marketing copy until the documented feature set and the price math support it.
What never moves a score
Market share and popularity are not criteria. Neither is hype. A product half the trading internet recommends gets the same five-criteria treatment as one nobody covers, which matters in a niche where most day traders lose money and the loudest recommendations often come from people selling courses.
Money doesn’t move scores either. This site launched with no paid relationship with any product or broker it covers; the full statement is on the editorial independence page, and if that status ever changes, the disclosure changes with it, on every affected page. How the site is run day to day is on the about page.
One structural guarantee you can audit yourself: within a category, review ratings are monotonic with our ranking order. The #1 pick on a best-of list never carries a lower rating than a pick ranked below it. If you ever catch a violation, something is wrong and we want to know.
When ratings change
Scores change when documented facts change: a price moves, a policy is rewritten, a feature ships or gets cut. They don’t change because sentiment shifted or a competitor published a hit piece. Every review carries a “Last verified” date reflecting when its pricing and policy sources were actually checked against the official pages, and pricing and policy pages for covered products are monitored between updates. When a monitored fact moves, the affected sub-score is recomputed under the weights above and the overall updates with it.
See the system in action
The fastest way to judge a methodology is to read its outputs. Start with our full day trading app comparison, where every rating in the tables links to the review that derives it, or go straight to a flagship example like our Trade Ideas review, where all five sub-scores and the logic behind each are laid out in the verdict.