FAQs

Fair question — most backtests are. So we built MonkScore to fail the usual ways a fragile one does, and it didn't. The architecture was frozen on 2015–2024, then run on the 2000–2014 history it had never seen: the top-to-bottom gap held at +24.8pp (t = 9.4). Strip out every factor academics use to explain returns — value, momentum, quality — and +23–26% a year is still left unexplained, on two separate universes. And since we froze it, the live quarters have kept working. None of that makes it certain to repeat. It does make it something other than a curve fit to its own history.

No. MonkScore is a research signal, not a recommendation. It scores a company's fundamentals against its peers; it doesn't know your goals, your portfolio, or your risk tolerance, and it never says "buy this." Treat it as one rigorous input to your own decision, not a substitute for it. For advisors and professionals: it's impersonal research you can point to, not personalized advice you're delegating — the call, and the responsibility, stay yours. (This isn't legal advice; if you're regulated, run it past your own compliance.) For everyone else: the conviction is still yours to build — we just put a number on the fundamentals so you're building it on evidence.

One plan, everything in it: a point-in-time MonkScore on roughly 8,000 companies, unlimited screening across the global universe, the five-pillar breakdown on every name, and portfolio monitoring for score changes and thesis breaks. It's $2,400 a year — one annual plan, no monthly tier, no stripped-down free version. The cheaper way in is the 15-day full-access trial: the entire product, no credit card, so you can verify it on your own holdings before you pay a cent. One value trap avoided tends to cover the year.

Probably. The universe is roughly 8,000 companies across developed and emerging markets — not a US-only list with a global label. Every name is scored point-in-time and ranked against its regional peer group, so a Japanese bank is judged against Japanese banks, not US software. If a company is big enough (>$300M market cap) and has enough reporting history to be scored honestly, it's in; the measurability gate declines to score the ones that don't, rather than guess. Look up anything on the trial and see.

You see more than the number. Every score opens into its five pillars — Growth, Profitability, Quality, Market Conviction, Safety — and exactly where the company ranks against its peers, so you can see what's driving it. The full methodology is published in a research whitepaper: how it's built, how it was tested, and the one limitation the data can't overcome. What we don't publish is the recipe — the 149 individual ratios and their weights. We'd rather be honest that there's a proprietary core than pretend there isn't. Everything you need to judge a score, you can see; everything you'd need to clone it, you can't.

Three ways, mostly. Look up any stock before you buy — a high MonkScore says the fundamentals back the case, a low one says ask harder questions. Screen the whole universe to surface and rank your own candidates: most screeners make you pick metrics and filter by hand, while MonkScore does the opposite, scoring every company so you sort on a finished read instead of building one. And run your existing holdings through it once a quarter — about fifteen minutes — to catch score changes and thesis breaks before they cost you. It's a discipline layer on top of your judgment, not a replacement for it.

No. MonkStreet is a research and data platform. We give you the analysis to make better-informed decisions; we don't tell you what to buy or sell, and nothing on the platform is personalized investment advice. Past performance doesn't guarantee future results, and all investing involves risk.

MonkScore is a single number from 0 to 100 that rates a stock on the fundamentals that have historically separated long-run winners from losers. It blends five separately-validated dimensions — growth, profitability, quality, market conviction, and safety — into one score, assigned to roughly 8,000 companies across global markets. It's a research signal, not a tip: a disciplined, repeatable read on a business, built to replace gut feel with evidence.

Each pillar captures a distinct dimension of investment quality, and each is grounded in the academic literature rather than invented.

• Growth — the persistence and breadth of fundamental expansion, blending what a company has actually delivered with what analysts expect.
• Profitability — the level of economic return: how far returns on capital exceed the cost of that capital, confirmed across the income statement and in cash.
• Quality — trajectory and stewardship: the direction of returns on capital, earnings quality, balance-sheet resilience, and capital-allocation discipline.
• Market Conviction — the market's own valuation premium for the stock, read as a vote of confidence that only counts when the other four pillars back it up.
• Safety — the absence of three well-documented failure modes: distress dynamics, fundamental deceleration, and deteriorating earnings quality.

A high score means a company looks strong across all five dimensions at once — not just on one. Historically, the relationship is a clean staircase: sort every stock into ten bands by MonkScore and each band has outperformed the one below it, top to bottom, with no inversions in the middle. The highest-scoring band has beaten the average stock in about 90% of quarters over 25 years. A high score is a statement of odds, not a guarantee for any single name.

Each company is ranked on every underlying ratio against its natural peer group — an industry within its region first, widening only when a peer set is too thin to be meaningful — so a Japanese regional bank is judged against other banks, not against a US software firm. The five pillars are then combined with a geometric mean rather than a simple average. That choice matters: a geometric mean punishes weakness in any one pillar harder than an average would, so a company can't buy a high score by excelling on one dimension while failing another. An expensive stock with weak fundamentals gets pulled down by those fundamentals, not rescued by its price.

Roughly 8,000 companies across North America, developed markets, and emerging markets — a genuinely global universe, not a US dataset with a label. Scores refresh as companies report, using only information that was actually filed and knowable at the time, so the score you see is built on a company’s freshest available fundamentals.

There's no hard cutoff, because the signal grades smoothly across its whole range — that's the point of the staircase. As a rule of thumb, scores of 90+ have historically been the strongest band, and the bottom band carries the classic value-trap fingerprint (low profitability, weak growth, cheap without fundamental support). But every ten points up the scale has, on average, meant a higher forward return — the score is informative across its entire range, not just at the extremes.

A screener filters on rules you set; MonkScore ranks on a tested composite. A screen can tell you which stocks have a P/E under 15; it can't tell you whether those stocks have historically done well. MonkScore folds 149 ratios across five pillars into one number whose decile ordering has held, out-of-sample, across a quarter-century and the global universe. You can still screen and filter on top of it — but the score does the heavy lifting of weighing many signals against each other at once, which is exactly the thing the human eye does poorly.

Across 436,693 company-quarters over 25 years, sorting the universe by MonkScore produces a monotone staircase: each of the ten bands outperformed the one below it. The gap between the top and bottom bands is +27.9 points a year — and that’s a spread, not a level. It clears the academic significance bar (t = 2.0) nearly seven times over, at t = 13.6. The top band beat the average stock in about 90% of quarters.

The staircase itself. A spread can be manufactured by two extreme buckets around a flat middle — this isn't that. Every step up the score scale has meant a higher average return, with no inversions across the ranks. Put as a gap: over five years, top-scored stocks returned roughly 10× the median of the worst-scored (about +97% versus +10%), in the same market over the same years.

A full 25 years, 2000–2026 — through the dot-com unwind, the financial crisis, the long expansion, the pandemic, and the rate shock. The signal is positive and statistically significant in all three sub-periods, not just on average. Notably, the top-versus-bottom gap was widest in the 2020s, which is the opposite of what a decaying or crowded signal looks like.

This is the right question to ask of any backtest, and we built the test to answer it. The architecture — which factors enter, the weights, the pillar definitions — was selected and frozen on the dense 2015–2024 data. The earlier 14 years (2000–2014) were loaded only after that freeze, which makes them a genuine out-of-sample holdout: data the model never saw when it was built. The spread holds there at +24.8 points a year (t = 9.4) with the identical 90% hit rate — roughly 82% of its in-sample size. An overfit signature doesn’t survive on data that was absent when the model was designed.

No, and we test that adversarially rather than asserting it. We took each classic factor — value, quality, momentum, size, low-risk, accruals, growth — and built the identical sort. The best single one (revenue growth) reproduces about a third of the spread; most reproduce far less, and two are flat-to-negative. We then controlled for all seven of MonkScore's own ingredients at once, and for the standard academic factor set, on two independent universes. A large, highly significant edge survives both — the combination structure is doing real work that no single factor, and no linear blend of them, reproduces. MonkScore isn't new physics; it's a disciplined combination of established signals whose value is precisely that it's irreducible to any one of them.

Yes — if anything, wider. Re-run on the global ex-North-America book, a different set of names with a heavier emerging-market weight, the staircase still holds and the top-versus-bottom gap is larger. A signal that survives a wholesale change of universe is measuring something real about companies, not something incidental to one market’s history.

The two practical tells point the other way. First, the top-versus-bottom gap was largest in the most recent decade, not the earliest — it isn't fading with time. Second, MonkScore's predictive power holds out to long horizons rather than evaporating in months the way crowded, short-lived anomalies do. And structurally, the specific way these signals are combined has never been published, while the edge concentrates in mid- and smaller-cap names that arbitrage capital can't cheaply reach — the corner where surviving mispricing is supposed to live.

Less reliably, and we'd rather you knew that. MonkScore is a fundamental signal; its edge builds with holding period rather than fading. It's designed for investors thinking in years, not days. If you're looking for a short-term trading trigger, this isn't the right tool.

Survivorship, and we put it in the executive summary of the whitepaper rather than the appendix. The historical universe is built from companies trading today, with their history back-computed — so firms that delisted, failed, or were acquired before now are absent. This inflates absolute return levels, which is exactly why we never headline an absolute return and lead with gaps instead. A gap (top versus bottom, or versus the average stock) is far less sensitive to the bias than a level. And the failed names that are missing would have clustered in the low-scoring bands, so their absence, if anything, makes the gap we report understated, not flattering. The bias we can’t remove from this data, we name, bound, and lean against — and a survivorship-free re-run is the validation we’d most like to put it through.

Because that's the figure survivorship bias most distorts. An absolute compounding number is precisely the kind of headline the missing-failures problem inflates, so quoting one would overstate what the signal earned. We publish spreads, hit rates, and base-rate gaps instead — measures whose bias either cancels out or runs against us. It's a deliberate discipline, not an omission.

The benchmark is the average stock — the equal-weight return of the investable universe MonkScore scores. Every comparison on the site is against that. We don’t benchmark against the S&P 500 because the relevant question is whether the score beats a random pick from the same universe, which is exactly what "the average stock" measures. It’s the honest yardstick for a stock-selection signal.

Two disciplines. First, point-in-time scoring: at every historical date, a company is scored using only data that had actually been filed and was knowable then — late or restated filings can’t leak backward. Second, we tested the one look-ahead we could measure: applying the size floor at each historical date rather than at today’s values changes the headline spread by only 0.7 of a point. Hindsight on size isn’t what produces the result.

Quality, in this particular dataset, and we say so plainly in the whitepaper. Its sub-factors are each correctly signed but individually weak over this sample, and removing the pillar marginally improves the headline. There are good structural reasons — capital-allocation discipline mostly pays off by helping you avoid the bankruptcies and value-destroying blow-ups that a survivors-only universe can't contain, so the test is partly blind to the pillar's core job. We keep it because the composite's strength is that no single pillar carries the result: drop any one of the five and a large, significant spread remains. We'd rather show you the weak link than hide it.

No. MonkScore is research — a systematic read on a company’s fundamentals — not personalized investment advice, and not a directive to trade any security. It’s an input to your process, not a substitute for it. You bring the portfolio context, the risk tolerance, and the final decision.

Most users use it three ways: as a ranking to surface candidates from a universe too large to read by hand; as a fast fundamental gut-check on a name already on the radar (the pillar breakdown shows why the score is what it is); and as a discipline that flags when a stock you like screens poorly, so you confront the disagreement deliberately rather than by accident. It compresses a great deal of fundamental work into one defensible number and the reasons behind it.

The backtest rebalances quarterly, and the signal's edge strengthens the longer you hold. It's built for investors operating on a multi-quarter to multi-year horizon, which is where the evidence is strongest. It is not a day-trading or short-horizon timing tool.

Yes, on the long side. The top band holds roughly 500 names at a representative recent date, centered on mid-caps, with a median market capitalization near $4 billion and most of the book at or above $2 billion — a broadly investable portfolio, not a micro-cap curiosity. Turnover is modest for a fundamental strategy (about 35% a quarter), so costs take well under a point and a half off the top band's return. The long side is the figure we lead with precisely because it's the one you can actually act on.

The Safety pillar's risk detectors are stored on an inverted scale from the main pillars: a higher number means more risk. On a company page, the platform flips them so they read intuitively alongside the pillars (higher is better); on the screener, you see the raw value with the color inverted so a dangerous reading still shows red. The detectors flag three failure modes — valuation-distress dynamics, fundamental deceleration, and accruals-based earnings-quality deterioration — that historically precede trouble.

The signal separates winners from losers more sharply when markets drop, not less — the gap is wider in the worst quarters than the calm ones. In practical, long-only terms: through 2008, top-scored stocks fell about half as much as the market (roughly −21% against −37%); in 2022, they were roughly flat-to-up while the market fell around 18%. We're careful not to oversell this — the long book still lost money in crisis quarters, just materially less. It leans defensive; it isn't a safe harbor.

Historically, the top band has been about 2.4× more likely to contain an eventual 10-bagger than the average stock. A few documented illustrations, ascending: Amazon scored a perfect 100 at the 2002 dot-com bottom, when the market was pricing bankruptcy, and returned about +450% over the next five years; NVIDIA scored a perfect 100 in 2016, right at its AI inflection, and returned roughly +1,647%; and Monster Beverage held a perfect 100 in 2002 — one of 32 separate quarters it scored 100 — and returned about +7,038% over the following five years. These illustrate what a top score can precede — they don’t prove a hit rate, and we can only point to names still in the data. Treat them as examples, not as evidence.

MonkStreet was built by Alberto Echevarría, a finance professional with more than twenty years in the industry, over five years as an independent project. The original motivation was a personal one: a costly loss on a stock that, in hindsight, scored near the bottom of the MonkScore scale the quarter before it peaked — which is what prompted the search for a systematic alternative to gut-feel investing.

No. MonkStreet publishes research and tools; it does not provide personalized investment advice, manage money, or recommend that any specific person buy or sell any specific security. Nothing on the platform should be read as individualized advice. Consult your own advisers before making investment decisions.

The pillars draw on a well-established body of work: Fama-French’s factor models, Novy-Marx on gross profitability, the Asness-Frazzini-Pedersen quality-minus-junk literature, Sloan on accruals, Piotroski on financial-statement signals, and Campbell-Hilscher-Szilagyi on distress risk, among others. MonkScore is an independent composite — these references establish the empirical foundations its pillars rest on, not any affiliation. Every candidate factor was sourced from the literature first and tested second, never mined from the data.

We disclose the architecture and withhold the recipe — and we're explicit about the line. Public: the five pillars and what each measures, the academic basis for each, the geometric-mean combination, the peer-ranking approach, the point-in-time discipline, and the full results with their limitations. Withheld: the identities of the individual ratios, the exact weights, and the specific construction of the Safety detectors. You can verify from what's disclosed that the method is principled and robust; you can't rebuild MonkScore from it, and that's by design. A whitepaper covering the construction, results, and an adversarial self-test is available on request.

No, and we mark them as illustrations every time. They’re survivor-selected — we can only point to companies still in the data — so they show what a top score can precede, not how often it does. The evidence that the model works is the staircase, the out-of-sample holdout, and the factor tests in the evidence section, not any individual name.

MonkStreet is a single plan at $2,400 per year. One tier, full access — no feature gates between paying users, no upsells.

Yes — 15 days of full access, no credit card required. You'll create an account first (there's no anonymous browsing of the data), and the trial gives you the complete product so you can judge it on its merits before deciding.

No. MonkScore is a long-horizon signal, and the platform is priced and built for investors using it that way, so the plan is annual only.

Independent RIAs and serious individual investors who want a systematic, evidence-based read on company fundamentals across global markets — people who'd rather anchor a decision in a tested signal than in a story. It's a fit if you value disclosed methodology and owned limitations over hype; it's not a fit if you're looking for short-term trade alerts or hot tips.