Radar Perene / articles

Market science

◦ Pieces dated up to 15 Jul 2026 were written under index methodology v1; the current series is v2.2. See the methodology.

How research is made — method, statistics, open data and the house’s own studies — applied to Brazil’s market. 105 essays in reading tracks: each one ends by pointing to the next. Descriptive, never a recommendation.

Start from a track

• Caixa de ferramentas do pesquisador independente2 essays
• Como nasce e avança uma pesquisa15 essays
• Dados abertos brasileiros como matéria-prima de pesquisa6 essays
• Economia e finanças brasileiras como ciência5 essays
• Estatística para quem lê mercado, não para quem programa12 essays
• Fundamentos de ciência de mercado7 essays
• Inteligência artificial e a integridade da pesquisa2 essays
• Ler ciência de mercado sem ser enganado4 essays
• O método da casa2 essays
• A disciplina dos sinais4 essays
• A disciplina dos sinais: janelas, prazos e o problema do look-ahead4 essays
• Analogia histórica: distribuição, não profecia7 essays
• Como a casa testa o que constrói9 essays
• Governança, transparência e o motor por trás9 essays
• Mercado pequeno, fórmula importada2 essays
• Mercado pequeno, fórmula importada: os limites da amostra brasileira6 essays
• Os índices publicados: o que é público e por que9 essays

Caixa de ferramentas do pesquisador independente 2

• What Zenodo is — and why the license is research tooA study published on the author's own website has the permanence of the author's own website: one unpaid hosting invoice and the "literature" disappears.
• Zotero for the researcher who studies markets aloneEvery research effort that lasts longer than a month faces the same silent adversary: the researcher's own memory.

Como nasce e avança uma pesquisa 15

• How Brazilian finance academia is organized: societies, meetings and journalsSeen from the outside, "academia" looks like a single building — a monolithic institution that approves and rejects ideas.
• How research is born: from question to hypothesisNo research project we know from the inside ever began with data.
• Submitting a paper to a Brazilian economics journalThe finished manuscript feels like the end of the line.
• The costs and timelines of publishing a scientific article: what rarely gets explainedThe public image of scientific publishing alternates between two false portraits.
• Levels of research: from curiosity to working paper, from preprint to peer reviewThe vocabulary of research reads like a catalogue of different objects: working paper, preprint, refereed article, technical note.
• What a desk reject is and what it does not say about an article's qualityThe fastest "no" in science arrives within weeks, sometimes days, and carries no referee report.
• What ORCID is and why independent researchers need itBrazilian science has a surname problem. Search for "Silva, L." in any article database and the query returns a crowd: engineers, physicians, economists, all…
• What is a preprint — and where ours liveIn January 2020, while medical journals were still organizing their review queues, the first descriptions of the new coronavirus circulated among scientists…
• Robustness testing: why a result that only works one way is not a resultThere is a question that dismantles most exciting market findings, and it fits on one line: does the result survive if something changes? Shift the window by…
• Research on demand: when research does not depend on a departmentFor most of the twentieth century, a serious research question had one path: finding a university department — or a bank's research desk — willing to adopt…
• Qualis vs. impact factor: two ways of measuring prestige that do not translate into each otherAsk two researchers whether a given journal is good and prepare for answers that do not talk to each other.
• Pre-analysis registration: the hypothesis declared firstEvery quantitative research project has an invisible drawer.
• Reproducibility in finance: why almost no one tests itThere is a simple test that separates research from well-formatted opinion: hand the study to a stranger and ask them to arrive at the same number.
• Blind review and open review: how journals evaluateAmerican orchestras changed one detail in their auditions from the 1970s onward: they hung a curtain between the musician and the panel.
• Sample size in financial series: why \"20 years of data\" can be too little"Twenty years of data" sounds like abundance.

Dados abertos brasileiros como matéria-prima de pesquisa 6

• CVM open data: what the regulator publishes beyond enforcementThe CVM appears in the news almost always in the sheriff's role: a ruling, a fine, a new rule, a formal notice.
• B3 open data: what is public and what is paidThe quote that shows up for free on a news portal arrives late — usually fifteen minutes after the trade happened.
• BCB open data: from the SGS to a Diário readingEvery business day, a handful of numbers runs through Radar Perene's Diário: the Selic rate, the CDI, accumulated inflation, the closing exchange rate.
• IBGE SIDRA: how a price-index table becomes research inputThe inflation number that opens the evening news has a house of origin almost no one visits.
• Tesouro Transparente: public debt data, open and ready for reproductionPublic debt is Brazil's most cited and least consulted macroeconomic figure.
• Ipeadata: more than 11,000 series that are not just textbook macroThe earlier articles on this trail visited four houses — Central Bank, B3, IBGE, Treasury, CVM — each the owner of its slice of Brazilian data.

Economia e finanças brasileiras como ciência 5

• Exchange rates: 'expensive dollar' is a data questionIn March 2020, the dollar closed at R$ 4.88 and the entire country declared it expensive.
• Business cycles: describing the past is not predictingAt every quarterly GDP release, the same scene: if the number comes in negative for the second time, headlines declare a recession; if it comes in positive…
• Bank spread: what Brazilian research has already measuredAt the same counter, on the same day, the same bank practices two numbers that seem to belong to different countries: what it pays those who leave money…
• Inflation and indexation: the memory contracts keepIn early 2021, millions of Brazilian tenants received a similar letter: the rent would be adjusted by the index written into the contract, and the index…
• Real vs. nominal interest rates: inflation decides who winsBetween August 2020 and March 2021, anyone earning Brazil's base rate on their savings watched the balance grow every month.

Estatística para quem lê mercado, não para quem programa 12

• Honest backtesting: split the sample before testing anythingThere is a genre of chart that never disappoints: the equity curve of a backtest.
• The Ibovespa and the dollar: what correlation doesn't tellNo pair of Brazilian numbers gets explained as frequently — and with as little ceremony — as the stock index and the dollar.
• Efficient market hypothesis: the theory and its criticsTwo economists are walking down the sidewalk when one of them spots a twenty-dollar bill on the ground.
• Confidence interval: what a number says — and what it withholdsEvery historical market average travels through the news like a passenger with no luggage: "stocks returned such-and-such percent a year", full stop.
• Look-ahead bias: the error that lets a model \"predict the past\"A student hands in a perfect exam. The teacher grows suspicious — and discovers the answer key was sitting on the desk the whole time. No one would call that…
• What cointegration is: when two series walk on a leashA woman leaves the bar late at night with her dog on a leash.
• Regime switching: when the rules change mid-gameA market relationship works for ten years with the regularity of a clock.
• Risk premium and factors: empirical fact, not a buy tipThere is a curious journey certain words make: they are born in the table of an academic journal, surrounded by caveats, and reappear twenty years later in a…
• Spurious regression: series that move together for nothingIn 1926, the British statistician George Udny Yule published a paper with a title that still stings: Why do we sometimes get nonsense-correlations between…
• Seasonality or folklore: what survives of \"sell in May\""Sell in May and go away." The proverb is older than most of the data series used to test it — the full English version told the investor to come back only…
• Statistical significance is not economic relevanceA powerful enough microscope finds impurities in any glass of water.
• Random walk: yesterday's price does not predict tomorrow'sIn the book that popularized the subject, the economist Burton Malkiel tells of a classroom experiment: his students built the chart of a fictitious "stock"…

Fundamentos de ciência de mercado 7

• Correlation is not causation — and markets confuse the twoAt five in the afternoon of any trading day, financial news delivers a ready-made explanation: stocks fell with the dollar, rates rose because of someone's…
• What a DOI is and why every Radar Perene study has oneThe internet forgets at a speed that would embarrass any library: studies of the phenomenon estimate that an enormous fraction of the links cited in academic…
• What p-hacking is and why it undermines market modelsA researcher tests twenty versions of the same idea against the same price series.
• What peer review is — and why a refereed study counts for moreTwo studies reach the reader in the same week and claim opposite things about the same market.
• What a working paper is — and what it is notType "working paper" into a search engine in Portuguese and most of what comes back is an auditing lesson: audit working papers, inspection evidence, the…
• Overfitting in backtests is not about trading, it is about methodThere is a guaranteed way to build a model that perfectly explains the last twenty years of any market: give it enough parameters.
• Survivorship bias: why 'winning' track records deceiveIn the Second World War, the planes that returned from missions came back riddled with bullet holes in the wings and fuselage — and command wanted to armor…

Inteligência artificial e a integridade da pesquisa 2

• AI authorship and detection in science: who signs, and the real limits of detectorsIn the first months of generative AI, a few scientific articles were actually published with an AI tool listed among the authors — name on the cover, next to…
• AI in scientific research: what journals already regulateIn a few years, scientific publishing went from "not our problem" to a surprisingly convergent body of rules on artificial intelligence.

Ler ciência de mercado sem ser enganado 4

• How to read a finance paper's abstract in three minutesMost of what gets asserted about "studies" in market debate comes from people who read one paragraph — and read it badly.
• Conflicts of interest in bank research: who pays the analyst who signs the reportThe report arrives well designed, dozens of pages long, with a detailed model and a conclusion about a company.
• What a meta-analysis is — and why finance in Brazil has almost noneWhen two studies on the same subject reach opposite conclusions, public debate picks the one confirming what it already believed and moves on.
• Red flags of financial pseudoscienceThe most deceptive material in markets does not look like quackery — it looks like science.

O método da casa 2

• How the Atlas tests whether a precedent is a stylized fact or a coincidenceA coin lands heads five times in a row. Two explanations compete: the coin is loaded, or chance did what chance does. Both hypotheses produce the same…
• Why we never write \"you should\": the P7 filter explainedThere are sentences the reader will never find on any page of this house.

A disciplina dos sinais 4

• From the signal registry to the published index: the road not every signal travelsSeen from the outside, a published index looks like a starting point — the beginning of the story the public follows.
• How many windows fit a signal: why seven and not oneAsk whoever promotes a "signal that works" what its horizon is, and watch the answer slide.
• The signal registry that also knows when to quitMarket signals almost never die in public.
• A contrarian signal, tested and refuted: what happened when our own thesis did not holdEvery research house has a pet thesis. Ours was among the most respectable: the idea that extreme market pessimism carries, inside it, the announcement of…

A disciplina dos sinais: janelas, prazos e o problema do look-ahead 4

• What \"existing\" means for a market signal: deadline, not intuition"I saw that signal last week." The sentence circulates in any market conversation and nobody asks what it means — because the answer would embarrass the…
• Pre-registration in practice: declaring the test before running the modelPre-registration has the reputation of an academic institution: a platform, a form, a stamp — the stuff of clinical trials and university labs.
• Recomputing a signal registry: what remains when the right ruler is appliedEvery registry of market observations accumulates, over time, an uncomfortable question almost nobody asks out loud: how much of what is here would have been…
• The signal that only exists after waiting: the discipline of fixed windowsThe market produces interesting observations all day long — an extreme here, a divergence there, a configuration that resembles another.

Analogia histórica: distribuição, não profecia 7

• Why comparing with 2008 and with 2020 gives different answersTwo market collapses live in the memory of any Brazilian investor who crossed the last two decades — and almost the only thing they share is the fright.
• Range, not point: how to read the output of a historical analogy modelThere is a predictable moment in every presentation of a scenario model: someone in the audience asks for "the number".
• The error of thinking \"we have seen this before\" guarantees the next stepThere is a phrase that signals experience and delivers the opposite: "I've seen this movie before".
• k nearest neighbors: how the house chooses which crises to compare with todayWhen a commentator compares the current moment with 2008, he has made two choices he almost never declares: he chose how many episodes to compare with — one…
• From the analog engine to the signal registry: how the two converseWhoever followed the last two trails of this series met two disciplines that, at first glance, live in separate buildings.
• \"Looks like 2018\" is a distribution, not a predictionIn any tense week, the sentence appears: "this looks like 2018".
• Regime changes, analog changes: when comparing with the past stops making senseEvery method has a question it would rather not hear.

Como a casa testa o que constrói 9

• Reproducibility audit: rerunning your own scripts to see whether the result repeatsAsk anyone who works with data whether their own numbers are reproducible and the answer will be yes — with the serene conviction of someone who never…
• Beating the trivial benchmark: the test every house index must pass before it existsEvery new indicator arrives in the world with an unfair advantage: it gets compared with the void.
• Five indices, one single test: why testing jointly changes the resultEvaluating indicators one at a time looks like the careful way to work — each gets exclusive attention, its own report, an individual verdict.
• Seven criteria that do not change: what stays fixed when an index changes versionAn index that changes version faces an identity question that is almost never asked out loud: if the formula changed, in what sense is it still the same…
• The dossier that decides whether an index survives: backstage of a real decisionNew versions of indicators are usually born by acclamation: someone improves a formula, the new chart looks prettier than the old one, and the swap happens…
• An index under revision: what a not-yet-deposited working paper meansOn the house's public shelf there are studies with a DOI, verifiable by anyone at any time.
• Multiple passes before publishing: internal review when there is no journalPublishing a working paper requires nobody's permission.
• When the house kills its own hypothesis: the track that did not survive validationThere is a moment in the life of any research line when the data stops cooperating with the enthusiasm.
• A null result is also a result: the study that found no statistical differenceThere is a study underway at the house whose central finding, so far, is a sentence no marketing department would approve: we found no significant…

Governança, transparência e o motor por trás 9

• The checklist the house applies before citing its own sourceThe easiest place to slip is not where one steps in fear — it is where one steps with confidence.
• Conflict of interest in reverse: what changes when whoever calculates the index sells no product on itThe most useful question to ask of any market analysis is not "is this right?" — it is "what happens to whoever wrote this if I believe it?".
• From self-assessment to open data: closing the loop between what the house audits and what it publishesSeen up close, the pieces of this trail look like separate subjects: a regulatory vacuum here, a stopping rule there, a published hit rate, a citation check…
• Licensing the engine: what it means when another provider wants the same disciplineAcross this trail, one thing became visible by accumulation: what sustains the house's indices is not a formula — it is a working regime.
• Publishing your own hit rate, even when it is not goodNotice what market analysis material tends to display: the call that anticipated the drop, the study that "predicted" the turn, the chart where the analyst's…
• Who audits the auditor: the problem of an index no one regulates from outsideA bank has a supervisor. A fund has a regulator, an administrator, an independent audit. A listed company answers to a swarm of instances. Now consider an…
• Who pays the bill if a market index errs: the institutional case for public self-assessmentWhen a medicine fails, there is pharmacovigilance, recall, civil liability.
• Reproducibility in action: what to do when a central statistic does not repeatThere is a moment in the life of anyone who works with data that no manual romanticizes: the number holding up the analysis is recalculated under a…
• Voluntary transparency has a cost: what the house opened and what it kept closed"Be transparent" is the cheapest advice in the world — for whoever gives it.

Mercado pequeno, fórmula importada 2

• From 12 to 82 stocks: why American-style market breadth does not fit BrazilAmerican market-analysis manuals devote entire chapters to a family of indicators called market breadth: how many stocks rise against how many fall, what…
• How many stocks are enough to measure \"the market\" as a whole"The market went up." The sentence is said thousands of times a day, and almost no one asks how many companies it takes for it to mean anything.

Mercado pequeno, fórmula importada: os limites da amostra brasileira 6

• Adapt or import: the decision that precedes any Brazilian sector indexThere is a moment, in every piece of research applied to a small market, that appears in no final paper: the moment someone decides whether the foreign…
• Small sample, big conclusion: the risk of generalizing from scarce data in BrazilThere is a statistical irony that haunts anyone researching the Brazilian market: the smaller the market, the larger the conclusions usually published about…
• Sector baskets in a concentrated market: when a few companies decide the whole sectorAsk an American analyst what "the financial sector" is and they will point to a list with dozens of banks, insurers, asset managers and payment processors.
• What the American finance literature assumes that Brazil does not haveEvery paper carries two layers: what it claims and what it assumes.
• Sector rotation: what changes when the sector itself is too small to rotateSector rotation is one of the most elegant ideas in applied finance: capital would migrate between sectors according to the phase of the economic cycle…
• Three incidents, one lesson: what fails when a concentrated sector basket breaksMethodologies do not mature in seminars. They mature when something breaks in production — when the series that had been running for months returns a number…

Os índices publicados: o que é público e por que 9

• 16 spreadsheets, one hash: the public data package behind two indicesA chart persuades; a downloadable file invites distrust — and that is why the file is worth more.
• One index, many regimes: why Structural Ânima and Tactical Ânima are not the same indexA city's climate and today's weather share the same object — the atmosphere — and nobody confuses the two.
• Structural Ânima: measuring a kind of risk that does not show up in the priceThe closing price is the most photographed piece of information in the market — and one of the most incomplete.
• Two versions, one revision: the ILI finding that did not surviveThere is a document almost no indicator provider wants you to find: the old version.
• Housing liquidity has an index: what the ILI measures and what it does not promiseWhoever sells a liquid stock discovers the price in seconds.
• Market relationships that do not survive: what remained after testing everything againEvery market accumulates a folklore of gears: when A rises, B falls; sector X leads the index; the ratio between Y and Z "always" reverts.
• Perene Risk: the index with a DOI and no public formula — and why that is the right choiceThere are two ways to hide a method. The first is the classic secret: nothing published, nothing verifiable, trust us. The second is subtler and more common…
• The only index with public weights: why the house opened one formula and not the othersAnyone leafing through the house's working paper series finds an asymmetry that looks like incoherence: five indices published with concept, DOI and…
• Utilities and the rest of the market: the relationship that survived the retestEvery market has its folklore of pairs: sectors that "move together", asset classes that "hedge each other", gears investors repeat in conversation as if…

The house’s laboratory lives at Research; each day’s reading, in the Daily; the precedents, in the Atlas.