The clear box
No black box.
This is the rubric.
Every score BlueChip52 has ever published came from the fixed 100-point system on this page: the same arithmetic, every Saturday, with no overrides and no judgment calls. If a stock scores 74, it scores 74 whether we like it or not.
The premise
Quality, temporarily out of favor.
BlueChip52 hunts in one narrow pond: large-cap stocks trading near their 52-week low whose fundamentals are still sound. Big companies rarely sit at their lows for long without a reason; the interesting question is whether the reason is temporary. The rubric below is our attempt to answer that question the same way every single week. Upcoming catalysts (earnings and ex-dividend dates) are mapped and shown as context, but they earn no points and gate nothing.
The cadence is fixed: every Saturday the scanner sweeps about 500 names across the DOW, NASDAQ 100, and S&P 500, filters for proximity to the low, and scores whatever survives. Weekly, rubric-driven, zero discretion. The list is whatever the arithmetic says it is.
The rubric
100 points. Three pillars. Fixed weights.
Each pillar is the sum of graded factors. A stock earns points by landing in a fixed band for each factor: no partial credit, no adjustment, and no way for us to nudge a favorite upward. The factors are named below; how they are weighted and the exact grading bands are part of what members pay for.
Momentum & timing
35/100How washed-out is the price, and is the business quietly improving underneath it? The score is built from factors such as distance from the 52-week low, where closer to the low earns more, and the quarter-over-quarter gross margin trend, which checks whether the operations agree with the pessimism in the price.
Value
35/100Cheap relative to earnings, assets, and growth. The score is built from the classic trio: the price-to-earnings ratio, price to book, and year-over-year revenue growth, each graded on fixed bands where cheaper multiples and healthier growth earn more.
Safety
30/100A stock at its low with a stressed balance sheet is a different animal from one that is merely unloved. Safety asks whether the company can wait out its own bad news, built from factors such as profitability across recent quarters, debt to equity, and the current ratio.
What the total means
The total maps to a band: a label for how strongly a name fits the screen, not an instruction to do anything. What you do with a shortlist is entirely yours.
| 80–100 | Strong candidate | Fits the screen on nearly every axis |
| 70–79 | Candidate | Solid fit with a soft spot or two |
| 60–69 | Watch list | Interesting, not yet compelling |
| Below 60 | Not listed | Doesn't make the email at all |
Shown, never scored
Context signals stay out of the math.
Some signals are genuinely useful to see but too noisy, or too unproven, to earn points. They appear in every report as context, and none of them can move a score by a single point.
Catalyst timing
How far out the next earnings or ex-dividend date sits. A dated reason for the market to re-look at the name. Shown as a window, never scored.
Street consensus
What analysts as a group call the name, how many of them cover it, and where their median price target sits. Useful orientation, never an input.
Sector divergence
How far the stock has fallen relative to its own sector ETF: is it the sector's problem or the company's?
Stabilization
Volatility contraction, volume dry-up, and a base that keeps holding above the low. Seller exhaustion, in short. Used only as a tie-breaker between equal total scores.
Analyst grade momentum
Recent upgrades and downgrades from the street. Reported as-is; analysts get no vote in the rubric.
Insider activity
Whether insiders have been net buyers or net sellers over the last three months, from SEC Form 4 filings. People with the best information, voting with their own money.
Where AI fits
The model writes. The arithmetic decides.
Every score on this site is deterministic arithmetic: fixed component bands, added up. No language model assigns, adjusts, or ranks anything.
Where an LLM does appear, it has exactly one job: writing the plain-English explanations (the "why it surfaced" notes and earnings-call summaries), and it is grounded in each stock's own published numbers when it does. Those notes are descriptive, never a recommendation: the model explains what the score already said, and it cannot say anything the numbers don't support.
What this can't do
The honest part.
A rubric like this has real failure modes, and pretending otherwise would be its own kind of black box. Value traps exist: a stock can be near its low, score well on every band above, and keep falling because the market understood something the trailing numbers didn't show yet. Catalysts disappoint: a scheduled earnings date guarantees a re-look, not a recovery; the news on that date can be bad. And backtests are not live results: any historical performance of these rules was computed with hindsight, without slippage, and without the discipline it takes to hold a falling name in real time.
That is why the score stops at "candidate" and never becomes an instruction, and why we publish every pick's outcome, winners and losers alike, from the moment it goes out.
One more honest thing: the rubric is not frozen. We revise the scoring system as the graded record teaches us what predicts and what doesn't, and we will keep changing the platform, the screens, and the weights as we learn. What never changes is the accounting: every published week stays on the record, graded under the rubric that produced it, and no revision ever rewrites an old grade.
What the list is for.
A weekly shortlist is a starting point for research, not an answer. Here is what that means in practice.
A starting point, not a conclusion
The screen surfaces names worth a closer look and does the first pass of the arithmetic for you: where a stock sits against its 52-week range, what the fundamentals say, what is scheduled next. The research after that is yours.
Not a portfolio
The list is a ranked set of candidates that each cleared the same bar, not an allocation and not a basket meant to be bought together. There is no position sizing anywhere in BlueChip52 because sizing depends on things we do not know about you.
Built to sit alongside your own approach
Most readers already have a process, an adviser, or a plan they like. This is meant to complement that by widening the set of names in front of you, not to replace it.
A high score is not a green light
Scores rank candidates against each other in a single week. They are not a probability, not a price target, and not a signal to act. A name can score well and keep falling, which is why every pick stays on the public record either way.
The short version: do your own research, decide for yourself, and treat a place on this list as a reason to look rather than a reason to act. If something here would be more useful in a different shape, tell us.
BlueChip52 is an informational screening tool, not investment advice. Nothing on this page is a recommendation to trade any security. Always do your own research. Full disclaimer.