Straight answers to fair questions.

MonkStreet is a research platform. It gives about 8,000 stocks a MonkScore from 0 to 100, based on the numbers behind each business. Most tools only help you buy. MonkStreet also tells you when a stock no longer deserves a place in your portfolio.

MonkStreet costs $2,400 a year, for one plan with everything in it: every score, the screener and monitoring of the stocks you own. Try it free for 15 days first, with no credit card.

Very likely, if it is a company worth more than $300 million. We score about 8,000 companies in the US, Canada, Europe, Japan and emerging markets. Each company also needs enough data and history to score fairly. If we cannot score a company well, we leave it out rather than guess.

We designed the MonkScore on 2015–2024 data. Its five pillars, its ratios and the way it combines them were set on those years, and they have not changed since. Later versions fixed bugs, cleaned data and made the rules consistent. None of them refitted the model to the past. We then ran the finished model on every year from 1999 to 2026, including the years before 2015 that played no part in its design. The final test is the future: every quarter since October 2026, we seal our list before the market opens and publish the result, good or bad. That is no guarantee, but you never have to take our word for it.

No. It shows you how every stock scores and why, so you choose with better information. The score does not know your goals, your portfolio or how much risk you can take, so the decision stays with you. If you use it for client work, your compliance team should review it.

Yes. Every score opens into its five pillars and shows where the company ranks against its peers, with 158 ratios, each ranked against the company's peers. Our methodology page explains how we built and tested the model. The exact recipe behind the MonkScore and its weights stay private. You can see enough to judge any score, but not enough to copy it.

Anytime, in one click from your account settings. You do not need to email support or call anyone. We hate those traps too. If you stop during the free trial, you pay nothing.

MonkScore is a single number from 0 to 100 that rates how strong a company's fundamentals are next to its peers. It blends five dimensions — growth, profitability, quality, market conviction and safety — into one score, given to about 8,000 companies. It's a research signal, not a tip: a disciplined, repeatable read on a business, built to replace gut feel with evidence.

Each pillar combines several financial measures to answer one question, and each is grounded in the academic literature rather than invented. Each company is ranked against its industry peers in its region when there are enough of them; otherwise against all companies in its region.

• Growth — is the company growing faster than its peers, and can that growth last?
• Profitability — does the company earn more on its money than its peers?
• Quality — does the company turn its profits into cash and reinvest them wisely?
• Market Conviction — does the market value this company above its peers, or price it as if something is wrong?
• Safety — is the company free of the warning signs that often come before a big loss?

A high score means a company looks strong across all five dimensions at once, not just on one. Historically, the odds shift with the score. Over 27 years of backtest that keeps every delisted and bankrupt company, the lowest score group returned 6.0% a year and the 90+ group 12.0%, each held in equal amounts and rebuilt every quarter, before costs. Over ten years, 2.7% of the stocks scored 90 or above rose tenfold or more, against 1.7% of all stocks, and 4.4% lost 90% or more, against 9.5%. The middle groups are too close to rank, and we show that. A high score is a statement of odds, not a guarantee for any single name.

Each company is ranked against its industry peers in its region when there are enough of them; otherwise against all companies in its region. 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.

About 8,000 companies across North America, developed markets and emerging markets. Scores are recomputed every day from the latest filed reports and market prices, using only information that was public at the time.

There's no hard cutoff. As a rule of thumb, scores of 90+ have historically been the strongest group, and the lowest group carries the classic value-trap fingerprint (low profitability, weak growth, cheap without fundamental support). Higher scores have, on average, meant higher returns, but the middle groups are too close to rank, so read the score as a range, not a precise ranking.

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 a company’s fundamentals across five pillars into one number, tested over 27 years of point-in-time history with every dead company included. 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 365,193 company-quarters from 12,105 US companies above our size floor, from 1999 to 2025, with every delisted and bankrupt company kept in, higher scores did better. Every quarter, the companies were sorted into ten groups by score and held in equal amounts until the next quarter: the lowest group returned 6.0% a year and the 90+ group 12.0%, before costs. The average stock returned 8.5% a year, and the S&P 500 index 8.2%, dividends reinvested. Our main test asks whether the 90+ stocks returned more than the stocks they are chosen from after adjusting for their lower risk. The answer is a t of 4.1, above the bar of 3 we set because we tried many versions of the model. The other tests are in our performance document, available on request. In the backtest, the 90+ stocks returned more than the average stock in 86 of 89 rolling five-year periods and in 69 of 69 rolling ten-year periods. One middle group did slightly worse than the one below it, and we show that rather than smooth it. The score sorts the top and the bottom of the market more clearly than the middle. Backtested and hypothetical, not a promise.

The staircase itself. Every quarter, all scored companies were sorted into ten groups by score and held in equal amounts until the next quarter. Return climbs from 6.0% a year in the lowest group to 12.0% in the 90+ group, before costs. The middle groups are too close to rank, so we describe the climb and do not claim a perfect order. Over five-year periods the picture is the same: about +40% in total at the bottom against about +82% at the top.

A full 27 years, 1999–2026: through the dot-com unwind, the financial crisis, the long expansion, the pandemic and the rate shock. Every delisted and bankrupt company is kept in: 365,193 company-quarters from 12,105 US companies above our size floor, from 1999 to 2025. In the backtest, the 90+ stocks returned 12.0% a year against 8.5% for the average stock, before costs, and returned more than the average stock in 69 of 69 rolling ten-year periods. After adjusting for risk, the t is 4.1. It is not smooth: the record includes multi-year stretches behind the S&P 500, and we publish those too.

It's the right worry, and the honest answer is: partly, unavoidably. Every model is shaped by the data its designers lived through. So we do three things about it. First, adversarial tests. Only 13 of 500 random copies of the Safety pillar did as well as the real one. A look-ahead cheat we injected on purpose made the result explode to +339% a year (t = 7.7), the signature of a leak and nothing like our real numbers. And after controlling for six academic factors, +1.9 points a year remain (t = 2.3), before costs. Second, a correction discipline: when we re-ran everything on cleaner, delisting-inclusive data and the numbers came back smaller, we published the smaller numbers. Third, and this is the real answer, each portfolio is sealed before its quarter starts, and the track record is public. Curve-fitting can fake a backtest. It can't fake the next ten years, timestamped.

No, and we test that rather than assert it. Control for the six academic factors (market, size, value, profitability, investment and momentum) and MonkScore still adds +1.9 points a year they don't explain (t = 2.3), before costs. And the disaster filter is where the combination earns its keep: over ten years, the share of stocks that lost 90% or more falls at every step of the score, from 16.3% in the lowest group to 4.4% in the 90+ group. MonkScore isn't new physics; it's a disciplined combination of established signals, and the combination is where the value lives.

We show the droughts instead of hiding them. The edge is not smooth: the 27-year record includes multi-year stretches of weak results. From January 2022 to March 2026, the 90+ stocks gained 25%, well ahead of the 17% of the average stock: the score kept its lead. The S&P 500 gained 46%, carried by its largest companies. From 2022 to 2025, it beat its own equal-weighted version by 5.2 points a year. Profitability and investment explain about 5% of the lead. The largest known-factor tilt is toward smaller companies. After all six standard factors, +1.9 points a year remain before costs. We won't argue you into comfort: the public track record is the live answer to this question, updated every quarter.

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.

It's a backtest, and no amount of discipline changes that. The record is hypothetical: no real money was at risk, and real trading has costs and frictions (our figures are before trading costs). Our performance claims are for US stocks only. And the edge arrives unevenly, with multi-year droughts a real investor has to sit through. What we can say is that the classic biggest limitation, survivorship, no longer applies: we rebuilt the record on 365,193 company-quarters from 12,105 US companies above our size floor, from 1999 to 2025, with every delisted and bankrupt company included, and retired every number that didn't survive the rebuild.

We used to refuse to publish figures like that. Our early tests used data with only the companies that survived. That bias inflates absolute numbers the most, so we published only the gaps between groups. Then we rebuilt the record on point-in-time data, with every delisted and bankrupt company included. That is what makes an absolute figure honest. $10,000 in the 90+ stocks (every stock scored 90 or above, rebuilt each quarter) grew to about $213,000 over 27 years before costs. The average stock, held the same way in equal amounts, grew to about $91,000, and the S&P 500 index to about $83,000, both with dividends reinvested. These results are backtested and hypothetical, but missing failures no longer flatter them.

Two things. The average stock is every US company above our size floor, held in equal amounts, the same way our portfolio holds its stocks. It tells you whether the score picks better than chance, so the statistical tests use it. The S&P 500 index, with dividends reinvested, is what most investors would otherwise own: from 1999 it returned 8.2% a year, and $10,000 grew to about $83,000. The live track record uses the S&P 500 equal-weight, the closest public index held the same way.

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 use each company's size at the time. Picking past companies by their size today would inflate returns massively: the average stock would show 18.8% a year instead of 8.5%.

If you're asking which pillar we'd defend least hard: none of them is built to win alone. Safety, for instance, is a veto, not a return engine, and the geometric-mean structure deliberately keeps any single pillar from carrying the result. The weakness we can actually show you is elsewhere. On the quarterly ladder, one middle group lands slightly below the one before it, and nothing we were willing to do fixed that honestly, so we publish the wobble as it is. And the edge takes multi-year vacations. The score's strength is clearest at the top and the bottom of the range. In between, the ordering is real but noisy, and we'd rather show you the weak step than smooth the chart.

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 portfolio held 309 names on 30 June 2026, with a median market value of $11.8bn and 83% of the money in companies worth more than $2bn: a broadly investable portfolio, not a micro-cap curiosity. Turnover is modest for a fundamental strategy, about 30% a quarter. The long side is the figure we lead with precisely because it's the one you can 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 colour inverted so a dangerous reading still shows red. The 4 detectors flag warning signs that historically come before trouble: value traps, fading growth, profits not backed by cash, and reliance on new outside money.

Less than we used to claim, and we would rather show you the record than a slogan. In 2008 the portfolio fell 36.3%, and in 2022 it fell 22.2%: very little shelter. Measured peak to trough rather than by calendar year, it fell less than the average stock in the dot-com unwind (−21% against −32%) and in the crash from September 2007 to March 2009 (−43%, against −49% for the average stock and −46% for the S&P 500). The pattern: the score leans defensive against company-specific trouble, and gives you very little when the whole market falls at once. It is not a safe harbour.

In the backtest, over ten years, 2.7% of the stocks scored 90 or above became ten-baggers, against 1.7% of all stocks, and 4.4% lost 90% or more, against 9.5%. At a MonkScore of exactly 100, 6.5% became ten-baggers, but 5.3% lost 90% or more. Concentrating into the very best scores buys upside and gives back some of the safety, so the portfolio we track uses the wider 90-and-above group. A few documented illustrations, each measured from the quarter the score put the company in the 90+ group. Amazon scored 100 in December 2008; holding it from there to March 2026 returned about +8,000%. NVIDIA entered the 90+ group in June 2016, before AI was the obvious story; holding to March 2026 returned about +15,000%, and the quarterly rule never sold it. NVIDIA's earlier signal, a score of 99 in June 2007, was followed by a 64% fall, and we publish that too. Monster Beverage first entered the 90+ group in March 2004. Holding it from there to March 2026 returned about +50,000%, while the quarterly rule held it until September 2015 for about +15,300%. Selected illustrative examples, not representative; they show what a top score can precede, not how often.

MonkStreet was built by Alberto Echevarría, a finance professional, as an independent project. The original motivation was a personal one: a costly loss on a stock that, in hindsight, scored only 67 the quarter before it peaked, with a Safety score of 19. That 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 check from what's disclosed that the method is principled; you can't rebuild MonkScore from it, and that's by design. A performance document with the full results and every test is available on request.

No. One big gain shows what a top score can lead to, not how often it does. Most high-scored stocks do far less, and some lose money. The evidence that the score works is the full record: every stock, every quarter, with every bankrupt and delisted company included. You can read it in the evidence questions above and on the methodology page. From now on there is also the public track record, which anyone can check.

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.