How to evaluate a stock scanner you can actually trust (5 criteria most fail).

Choosing a stock scanner is mostly an exercise in detecting what isn't said. Every product claims to "beat the market" or "outperform" or "deliver signals." Almost none publish the data that would let you verify those claims. The asymmetry is the whole story: the products with real records publish them prominently; the products without real records hide behind proprietary algorithms and selected case studies.

This post is the buyer's checklist — five tests we'd put any scanner through, with the questions to ask and what each answer tells you. Tapeline is one of the products you might evaluate with these criteria; we score ourselves at the end so you can compare us against the same yardstick we'd apply to a competitor.

Test 1: Can you see every pick, including the losers?

This is the single most important test and the one most products fail. "Picks of the day" or "today's recommendations" are easy to publish; what's hard is publishing every pick that was wrong, with the same prominence, on the same page, with the original score and signal still attached.

The right answer looks like: "Yes, every top-10 daily pick we've ever published is logged at this URL, sorted by date, with the next-day return and the original score visible. The page is append-only — we can't go back and edit it. You can see every win and every miss."

The wrong answer looks like: "Our algorithm has a 67% win rate based on internal testing" with no link to the raw record. That's a marketing claim, not a verifiable statement.

Test it: ask for the URL of the daily picks log. If they hesitate, you have your answer.

Test 2: Is the benchmark named?

"Outperforms the market" is a meaningless claim without a named benchmark. Outperforms which market? The S&P 500? Russell 2000? Equal-weighted universe? Sector ETF? The choice of benchmark changes the answer by 5-10 percentage points annually.

Real answer: "Picks are back-checked against SPY (S&P 500 ETF) on a same-day, next-trading-day basis. Alpha is the pick's 1-day return minus SPY's same-day return."

Vague answer: "Outperforms the broader market" with no specific index named.

Tapeline's choice: SPY same-day-pick to next-trading-day-close vs SPY same window. Documented at /scorecard.

Test 3: Is the scoring methodology published?

This is the test most products fail through omission rather than misdirection. They simply don't disclose the formula at all. "Proprietary algorithm developed over X years" is the standard formulation.

The right answer publishes the inputs, the weights, and any transformations explicitly. Tapeline's example: Trend 25% / Relative Strength 20% / Fundamentals 15% / Smart Money 15% / Macro 15% / Momentum 10%, summed to a 0-100 composite. Each sub-score's input data sources and computation steps are documented at /how-it-works.

What this enables: if the product underperforms in a particular regime, you can look at the score breakdown and see which factor is dragging. You can disagree with the weights and reason about that disagreement. With a black box, you can't do any of that — you can only stop using it.

Test 4: How fresh is the data?

"Real-time" means different things at different price tiers. On a free tier, "real-time" often means 15-minute delayed, because that's the IEX exchange delay limit on free consolidated feeds. On a paid tier, "real-time" usually means direct exchange data with sub-second latency.

Test it by checking the timestamp on a single quote. If it's behind the broker quote you'd see at your trading platform, you have your answer.

For composite scanners (like Tapeline), the question is slightly different: how often does the SCORE refresh, not just the underlying price? Tapeline's worker ticks every 60 seconds during market hours, recomputing the composite from fresh snapshot data. Free tier is 24-hour-delayed (intentional gating); Pro is 60-second freshness.

Test 5: What's the unsubscribe / cancel friction?

This is the dirtiest test, but the most diagnostic. Products that are confident in their value make it trivial to leave. Products that depend on dark-pattern retention make it hard.

Check before you sign up:

Can you cancel from a settings page in one click, or do you need to email support and wait?
Is the trial auto-converting? Do you have to add a credit card?
What's the refund window?

Tapeline's policy: trial doesn't require a card, cancel from /app/billing in one click, 7-day refund window on monthly subscriptions. We'd rather lose subscribers cleanly than retain them via friction.

How Tapeline scores against the checklist

Test	Tapeline
Public daily picks log with losers visible	Yes — /scorecard
Named benchmark	SPY, same-day-pick to next-trading-day-close
Public scoring formula with weights	Yes — /how-it-works with exact weights
Data freshness	60s composite refresh on Pro+; 24h delayed on Free (intentional)
Cancel friction	One-click cancel, 7-day refund, no card required for trial

What this checklist does NOT test

To be honest: none of these criteria tell you whether the product will work for YOUR specific trading style. You might need filter breadth Tapeline doesn't have. You might need international coverage we don't offer (US-only for now). You might prefer the analyst-rating focus Zacks does better than anyone. The checklist tells you whether the product is honest about what it does, not whether what it does matches your needs.

The two questions you need to answer separately:

Does this product publish enough evidence for me to verify their claims? (Checklist above.)
Does the thing they're publishing evidence FOR match what I actually want to do? (Read the product page + try the free tier + read the scorecard.)

Tapeline's pitch: we'd rather lose your business to a product that fits you better than win it via misleading claims. If the scorecard convinces you, the 14-day Premium trial is the no-card way to see the rest. If it doesn't, that's useful information too.