Quant Memo

Everyone agreed the average loss in the tail was the better risk measure. It turned out there were several ways to define it, and only one of them is genuinely coherent.

Jul 13, 2026

A bigger model always fits your data better. Akaike found the exact price you should charge for each extra parameter, and turned model selection from an art into arithmetic.

Jul 13, 2026

Earlier execution models assumed the order book was a flat block of liquidity. Real books are lumpy. This paper solved the problem for a book of any shape at all.

Jul 13, 2026

Everyone theorised about market impact. This team took roughly 700,000 real institutional orders and just measured it, producing the model brokers still quote today.

Jul 13, 2026

Altman took five ordinary accounting ratios, weighted them into a single number, and produced the first statistical model that could tell you which companies were about to go bust.

Jul 13, 2026

Amihud built a liquidity measure out of nothing but daily returns and volume, and showed that illiquid stocks earn more, and that the whole market repriced when liquidity dried up.

Jul 13, 2026

If a stock is expensive to trade, nobody will hold it unless it is cheap enough to compensate. Amihud and Mendelson turned that obvious thought into an asset pricing theory.

Jul 13, 2026

For years GARCH looked useless because it 'explained' only about five percent of tomorrow's volatility. Andersen and Bollerslev showed the models were right and the grading was wrong.

Jul 13, 2026

Newey-West standard errors require you to pick a lag length, and everyone was picking one arbitrarily. Andrews worked out what the optimal choice actually is and how to let the data make it.

Jul 13, 2026

Ang, Hodrick, Xing and Zhang found that the stocks bouncing around the most for their own private reasons went on to earn the worst returns of all. Theory says this should not happen.

Jul 13, 2026

Bond models ignored inflation and growth. Macro models ignored arbitrage. Ang and Piazzesi built the first serious bridge, and forecasts got better.

Jul 13, 2026

Engle, Lilien and Robins let the volatility forecast feed straight back into the expected return, turning 'risk should be rewarded' from a slogan into something you can actually estimate.

Jul 13, 2026

Instead of proposing a new risk measure, Artzner and coauthors asked what any sane risk measure must do. Then they showed that Value-at-Risk, the industry standard, breaks one of the rules.

Jul 13, 2026

Asness and Frazzini showed that the standard academic value factor uses a stock price that is up to eighteen months stale, and fixing that one detail changes everything.

Jul 13, 2026

Risk parity holds a lot of bonds and levers them up. That sounds like a gimmick. Asness, Frazzini and Pedersen explain why it is actually harvesting a premium other investors cannot reach.

Jul 13, 2026

If a company is safe, profitable and growing, investors should pay up for it. Asness, Frazzini and Pedersen found they don't pay up nearly enough, and that gap is a factor.

Jul 13, 2026

Instead of matching stocks two at a time, Avellaneda and Lee stripped out every common factor from every stock at once and traded whatever wiggle was left over.

Jul 13, 2026

Most impact studies measure one moment. This one followed 400,000 real orders from the first trade to the final settling of the price, and found the market never saw them coming.

Jul 13, 2026

Real relationships do not break once. Bai and Perron built the machinery to find several unknown regime changes at once, and to test how many there really are.

Jul 13, 2026

Boring stocks beat exciting ones, and professional investors are the very people least able to do anything about it, because they are all measured against a benchmark.

Jul 13, 2026

Bali, Cakici and Whitelaw found that a stock's single best day in the past month predicts terrible returns going forward. Investors pay up for the dream of a jackpot.

Jul 13, 2026

Bansal and Yaron argued that what scares investors is not this year's recession but a tiny, nearly invisible drift in the long-term growth rate, and that a small permanent change in the future is far more frightening than a big temporary one.

Jul 13, 2026

In 1981 Rolf Banz noticed that small companies had beaten large ones by more than their risk could explain. It was the first big anomaly, and it opened the door to everything that followed.

Jul 13, 2026

Barber and Odean showed individual investors buy the stocks that grab their attention, because they can only sell what they already own, but they can buy anything.

Jul 13, 2026

If overconfidence causes excessive trading, then the group psychology says is more overconfident should trade more and lose more. Barber and Odean found exactly that.

Jul 13, 2026

Barber and Odean tracked 66,465 households and found the ones who traded the most badly lagged the market, while the ones who did almost nothing nearly matched it.

Jul 13, 2026

How can markets underreact to news at short horizons and overreact at long ones? BSV built a single investor who does both, using two well-documented psychological biases.

Jul 13, 2026

Realized volatility gave everyone a way to measure market jumpiness. This paper gave them the error bars, which turns out to matter enormously.

Jul 13, 2026

Give a trader a Value-at-Risk limit and watch what they optimize. Basak and Shapiro show they will take more risk, not less, and lose more when they lose.

Jul 13, 2026

Sanjoy Basu showed in 1977 that stocks with low price-to-earnings ratios beat expensive ones on a risk-adjusted basis. It was the value effect, found sixteen years before Fama and French made it famous.

Jul 13, 2026

Bates tested whether random volatility alone could explain option prices, found that it could not, and showed you need jumps as well.

Jul 13, 2026

Benartzi and Thaler explained why stocks pay so much more than bonds: investors hate losses and look at their portfolios far too often, so they see far too many losses.

Jul 13, 2026

When you run thousands of tests, demanding zero false positives is too strict to find anything. This paper proposed a smarter goal: control the fraction of your discoveries that are wrong.

Jul 13, 2026

The most widely used benchmark in institutional trading was invented in a 1988 paper trying to answer a simple question: how much does it actually cost a big fund to trade?

Jul 13, 2026

Good earnings surprises keep pushing a stock up for months after the announcement. Bernard and Thomas showed this is not hidden risk, it is investors failing to do simple arithmetic.

Jul 13, 2026

If you have to sell a million shares by Friday, when do you sell them? Bertsimas and Lo turned that vague worry into a solvable maths problem and got a surprisingly boring answer.

Jul 13, 2026

Best and Grauer worked out exactly how violently a mean-variance portfolio reacts to a small change in one expected return. The answer explains why nobody trusts optimizers.

Jul 13, 2026

Three of the field's architects sat down and drew the whole map of market microstructure, from why spreads exist to how markets should be designed.

Jul 13, 2026

Bikhchandani, Hirshleifer and Welch showed that perfectly rational people watching each other will all end up copying, and the whole crowd can be wrong because of two early coin flips.

Jul 13, 2026

Realized volatility mixes together the market's ordinary churn and its sudden violent jumps. Barndorff-Nielsen and Shephard found a way to measure the churn alone, and subtract to reveal the jumps.

Jul 13, 2026

Black realised that if you price an option off the forward instead of the spot, storage costs, dividends and convenience yields all vanish from the problem.

Jul 13, 2026

Merton let a company fail only on the day its debt came due. Black and Cox added the trip wire that lets creditors pull the plug the moment the firm crosses a line.

Jul 13, 2026

A one-factor interest rate model that lives entirely on a binomial tree, keeps rates positive, and fits both today's yields and today's volatilities. It ran real trading books before it was ever published.

Jul 13, 2026

Three future legends built the first serious empirical test of the CAPM, invented the portfolio-sorting method every quant still uses, and found that the model's central line was real but bent in a way theory could not explain.

Jul 13, 2026

Black-Derman-Toy accidentally tied mean reversion to volatility. Black and Karasinski cut the wire, giving traders three independent dials and a rate that stays positive.

Jul 13, 2026

Black argued that people trading on nothing at all are what make markets possible, and that prices are usually somewhere in the right neighborhood rather than exactly right.

Jul 13, 2026

Fischer Black noticed that the CAPM's flattest, most stubborn empirical failure disappears once you admit that most investors cannot borrow at the Treasury rate, and the fix he proposed is the intellectual root of low-volatility investing.

Jul 13, 2026

When the gap between what options charge for volatility and what volatility actually delivers gets wide, the stock market tends to go up. It beats the price-earnings ratio at forecasting returns.

Jul 13, 2026

Order flow is one of the most predictable things in finance. Prices are almost perfectly unpredictable. Bouchaud and colleagues explained how both can be true at once.

Jul 13, 2026

Boyle's insight was almost embarrassingly simple: if an option's price is an average over possible futures, simulate a lot of futures and take the average.

Jul 13, 2026

Traders had been pricing caps with a formula that no model justified. Brace, Gatarek and Musiela built the model that made the market's own formula correct.

Jul 13, 2026

Breeden showed that all of Merton's many risk factors collapse into a single one, how an asset moves with what people actually eat, giving finance a model that is both simpler and better grounded. Then the data refused to cooperate.

Jul 13, 2026

Breeden and Litzenberger showed that the whole probability distribution the market believes in is sitting inside the option chain, if you know how to look.

Jul 13, 2026

Breiman's insight: build hundreds of decision trees, cripple each one on purpose so they disagree, then average them. The result is one of the most reliable predictors ever invented.

Jul 13, 2026

Do you want a model that explains, or a model that predicts? Breiman argued the statistics profession had spent decades choosing wrong, and finance is still arguing about it.

Jul 13, 2026

The most cited statistic in asset allocation says policy explains 93.6 percent of performance. It is a real result. It does not mean what almost everyone thinks it means.

Jul 13, 2026

Implied volatility used to mean 'the number you get by inverting Black-Scholes', which is awkward when Black-Scholes is wrong. This paper extracted the market's volatility forecast without any model at all, and the VIX was rebuilt around it.

Jul 13, 2026

Simulation runs forwards, early exercise reasons backwards. Broadie and Glasserman's fix was to build two deliberately biased estimators, one too high and one too low, and trap the truth between them.

Jul 13, 2026

With data identifying which trades came from high frequency firms, the authors could finally ask whether the fast traders push prices toward fair value or away from it.

Jul 13, 2026

If the market knows you have to sell, the other traders will not help you. They will sell first, buy back cheap, and profit from your distress.

Jul 13, 2026

The high frequency arms race is not greed, it is a bug in how markets process time. Fix the clock and the race has nothing to run for.

Jul 13, 2026

Campbell and Cochrane solved the equity premium puzzle with one human observation: your fear of losing money depends not on how rich you are, but on how close you are to the standard of living you have gotten used to.

Jul 13, 2026

Campbell and Shiller turned a piece of accounting into an inescapable trap: if the market looks expensive, then either dividends must grow fast or future returns must be low, and there is no third option.

Jul 13, 2026

The expectations hypothesis says a steep curve means long yields will rise. Campbell and Shiller found long yields tend to fall instead. The sign is wrong, and it gets more wrong the longer the bond.

Jul 13, 2026

Carr and Madan borrowed a signal-processing algorithm and made it price a whole strip of options in one shot, turning exotic models from academic curiosities into production tools.

Jul 13, 2026

The market charges more for volatility than volatility ends up costing. Carr and Wu measured that gap cleanly, and found it is large, negative and remarkably persistent.

Jul 13, 2026

Order flow arrives in bursts, and one trade makes the next one more likely. Build that into your quoting model and the market maker's optimal behaviour changes completely.

Jul 13, 2026

Amateur investors writing stock articles on the internet sounds like a recipe for noise. Chen, De, Hu and Hwang found their collective view predicted returns and earnings surprises.

Jul 13, 2026

The maths of gradient boosting already existed. Chen and Guestrin made it fast, regularized and robust to missing data, and it promptly took over applied machine learning.

Jul 13, 2026

Statistical factors have no names, which makes them hard to trust. Chen, Roll and Ross tried the opposite approach: start with real economic variables like inflation and industrial production, and see which ones stocks actually pay you to bear.

Jul 13, 2026

Chinco, Clark-Joseph and Ye let a lasso hunt through every stock's recent returns for one-minute forecasts, and the odd, fleeting predictors it found turned out to be news.

Jul 13, 2026

If you can only get one input right, get the expected returns right. Chopra and Ziemba measured just how much more damage a mistake there does than a mistake anywhere else.

Jul 13, 2026

Everyone says diversification is the only free lunch. Almost nobody can define it. Choueifaty and Coignard proposed a ratio, then built the portfolio that maximizes it.

Jul 13, 2026

A risk model that breaks its limit exactly five percent of the time can still be dangerously wrong, if all the breaches arrive on the same terrible week. Christoffersen built the test that catches it.

Jul 13, 2026

Build the calmest possible portfolio of US stocks, ignoring returns entirely, and you get much lower risk with no loss of return. Finance theory says that should not happen.

Jul 13, 2026

John Cochrane's presidential address argued that almost everything that moves asset prices is not news about future profits, but changes in the return investors demand.

Jul 13, 2026

Cochrane and Piazzesi found a single combination of forward rates that predicts the returns on every Treasury bond, and it is invisible to standard term structure models.

Jul 13, 2026

Companies copy and paste last year's annual report. When they bother to change the wording, something is wrong, and nobody notices for months.

Jul 13, 2026

Collin-Dufresne, Goldstein and Martin tested every variable theory says should drive credit spreads. Almost none of the movement was explained, and what was left over marched in lockstep across every bond.

Jul 13, 2026

A token pays no dividend, so discounted cash flow is useless. Cong, Li and Wang built the model that values a token from the demand to actually use it, and found adoption follows an S-curve.

Jul 13, 2026

Optimal execution tells you how much to trade. It never tells you how. Cont and Kukanov solved the decision every trader actually faces: limit or market, and on which exchange.

Jul 13, 2026

Everyone models price impact using trades. This paper shows that what actually drives short-term price moves is the imbalance of everything happening at the top of the book, including the orders that never trade.

Jul 13, 2026

Forget informed traders and strategic games. Model the order book as a set of queues with orders arriving and leaving at random, and you can compute the odds of what happens next.

Jul 13, 2026

Cooper, Gulen and Schill found that the single best predictor of poor future stock returns is not valuation or profitability, but how fast a company grew its balance sheet.

Jul 13, 2026

Copeland and Galai realised that a posted bid and ask are two free options handed to the world, and the spread is the premium you must charge to survive them.

Jul 13, 2026

Corsi forecast volatility using yesterday's, last week's and last month's averages. That is the whole model, and it embarrasses far more sophisticated competitors to this day.

Jul 13, 2026

Cortes and Vapnik asked which of the infinitely many lines separating two groups is the right one, and their answer, the widest gap, dominated machine learning for a decade.

Jul 13, 2026

Corwin and Schultz noticed that the daily high is almost always a buy and the daily low is almost always a sell, and built a spread estimator out of that one observation.

Jul 13, 2026

CIR rebuilt the yield curve from an actual economy rather than a hedging trick, and got a short rate that keeps itself above zero and gets calmer as it falls.

Jul 13, 2026

Cox and Ross noticed that if an option can be hedged, its price cannot depend on how much investors fear risk, so you may as well assume they do not fear it at all.

Jul 13, 2026

Cox, Ross and Rubinstein found a way to price options using only addition and multiplication, and in doing so made the logic of Black-Scholes visible to everyone.

Jul 13, 2026

Attention was always assumed and never measured. Da, Engelberg and Gao realised that when a retail investor is interested in a stock, they type its ticker into Google.

Jul 13, 2026

Dai and Singleton organised every affine term structure model into a family tree and found a painful trade-off: you can model volatility, or you can model correlations, but not both.

Jul 13, 2026

Daniel, Hirshleifer and Subrahmanyam showed that investors who overrate their own research, and take credit for wins while blaming luck for losses, will produce momentum and long-run reversal all by themselves.

Jul 13, 2026

Momentum earns steady profits for years and then loses a decade of them in a couple of months. Daniel and Moskowitz showed exactly when those crashes happen, and that you can see them coming.

Jul 13, 2026

Fama and French said value stocks earn more because they move together with a risk factor. Daniel and Titman ran a clever test and concluded it is the characteristic itself that pays, not the risk.

Jul 13, 2026

They tested nearly 8,000 technical trading rules on 100 years of Dow data, and asked whether the winners were skill or just the best of a very large lottery.

Jul 13, 2026

De Bondt and Thaler bought the market's most hated stocks, shorted its most loved ones, and found that the crowd's enthusiasm reliably runs out.

Jul 13, 2026

Zhang, Zohren and Roberts built a deep network that predicts short-term price moves from raw limit order book data, and found it worked on stocks it had never seen.

Jul 13, 2026

De Long, Shleifer, Summers and Waldmann showed that irrational traders create a risk all their own, one that scares off the smart money and lets mispricing survive.

Jul 13, 2026

Demeterfi, Derman, Kamal and Zou showed that a specific weighted basket of options gives you pure exposure to volatility, with no view on where the market goes. That result is why the VIX exists.

Jul 13, 2026

DeMiguel, Garlappi and Uppal tested 14 sophisticated portfolio models against the dumbest possible rule, put an equal amount in everything, and the dumb rule held its own.

Jul 13, 2026

Derman and Kani took the binomial tree and grew it backwards out of market prices, producing a tree that reproduces the volatility smile by construction.

Jul 13, 2026

Diebold and Li turned Nelson-Siegel's static curve-fitting formula into a forecasting machine, by treating its three parameters as things that move through time.

Jul 13, 2026

Model A has a lower error than Model B. Is that a real edge, or just luck? Diebold and Mariano built the test that answers the question everyone was eyeballing.

Jul 13, 2026

Everyone modelled squared returns because the maths was convenient. Three researchers checked whether the data agreed, and found the market's memory is strongest when you do not square anything at all.

Jul 13, 2026

Bitcoin in 2014 had no institutional market makers, no regulation and barely any arbitrageurs. The square-root law of market impact held anyway, which tells you something profound.

Jul 13, 2026

The most sophisticated interest rate models in finance predicted future yields worse than assuming nothing changes. Duffee found the culprit and fixed it.

Jul 13, 2026

Vasicek and CIR both produce closed-form bond prices. Duffie and Kan explained why, and mapped out the entire family of models that share the magic.

Jul 13, 2026

Duffie and Singleton found a way to price risky bonds using exactly the same machinery as safe bonds, just by bending the discount rate to swallow the default risk.

Jul 13, 2026

Dupire showed that the volatility smile does not just have to be tolerated, it can be inverted: the market's own option prices tell you exactly what volatility must be at every price and every date.

Jul 13, 2026

One number, printed under every regression for seventy years, that tells you whether your errors have a memory. Ignoring it is how researchers discover spurious relationships.

Jul 13, 2026

Easley, Kiefer, O'Hara and Paperman found a way to count informed traders you can never see, using nothing but the daily tally of buys and sells.

Jul 13, 2026

How do you measure the uncertainty of an estimate when the maths is impossible? Efron's answer was to resample your own data, over and over, and just look.

Jul 13, 2026

Unit root tests are famously bad at detecting slow mean reversion. Three authors worked out how much better any test could possibly be, then built one that gets close.

Jul 13, 2026

Fit GARCH to almost any market and the parameters nearly add up to one, which means today's shock still matters a year from now. Engle and Bollerslev took that seriously and built IGARCH.

Jul 13, 2026

Modelling volatility for one asset is easy. Modelling how a hundred assets move together used to be impossible. Engle's two-step trick made it routine.

Jul 13, 2026

Every volatility model implies a picture of how today's surprise moves tomorrow's volatility. Engle and Ng drew that picture, gave it a name, and used it to referee the whole GARCH family.

Jul 13, 2026

Estrella and Hardouvelis showed that the slope of the yield curve forecasts real economic activity better than the professional forecasters do, which is why an inverted curve terrifies everyone.

Jul 13, 2026

Fama and Bliss showed that when forward rates are high, it is mostly not because the market expects rates to rise. It is because bonds are paying you extra to hold them.

Jul 13, 2026

Twenty-two years after inventing the three-factor model, Fama and French admitted it was missing something and added two more ingredients: how profitable a company is, and how fast it is growing its assets.

Jul 13, 2026

Fama and French took size, value and momentum to four regions of the world. Value and momentum showed up nearly everywhere, Japan was the awkward exception, and global models failed.

Jul 13, 2026

Fama and MacBeth set out to test the CAPM and invented a two-step regression procedure so useful that it outlived the theory it was designed to test.

Jul 13, 2026

Market impact is concave and then decays. This paper argues that is not a coincidence of plumbing but a logical consequence of markets being unpredictable.

Jul 13, 2026

Hundreds of factors have been published. Feng, Giglio and Xiu built a statistical test that asks whether a new one adds anything at all, and most of them do not.

Jul 13, 2026

Volatility shocks do not vanish in a week, but they do not last forever either. Baillie, Bollerslev and Mikkelsen built the model that lives in the awkward middle.

Jul 13, 2026

Foley, Karlsen and Putnins used police seizures and darknet market addresses as seed data, then traced the blockchain outward to estimate what share of Bitcoin activity was criminal.

Jul 13, 2026

Tversky and Kahneman proved that simply rewording a decision, with no change to the actual outcomes, can flip people's preferences completely.

Jul 13, 2026

Textbook finance says riskier stocks should earn more. Frazzini and Pedersen showed the opposite has happened for decades, and pinned the blame on investors who cannot borrow.

Jul 13, 2026

Friedman showed that if you keep training small models to fix the mistakes of the model before them, you end up with something remarkably powerful.

Jul 13, 2026

Two stocks entitled to the same cash flows traded at persistently different prices for decades, and each one drifted with the mood of the country it happened to be listed in.

Jul 13, 2026

Currency options broke Black-Scholes because a currency pays interest while you hold it. Garman and Kohlhagen found the fix, and it is simpler than you would guess.

Jul 13, 2026

Wall Street had been quietly trading pairs of similar stocks for decades. Gatev, Goetzmann and Rouwenhorst finally tested the folklore on forty years of data, and it worked.

Jul 13, 2026

Pick the wrong combination of impact shape and impact decay and your model implies you can print money by trading in circles. Gatheral worked out which combinations are legal.

Jul 13, 2026

Gatheral, Jaisson and Rosenbaum measured how jagged volatility actually is, found it far rougher than every model assumed, and accidentally explained the smile that nothing else could.

Jul 13, 2026

Geske priced options on options, and in the process explained why volatility rises when stocks fall, twenty years before anyone called it the leverage effect.

Jul 13, 2026

Testing whether twenty-five portfolios all have zero alpha by checking twenty-five t-statistics is a recipe for fooling yourself. Gibbons, Ross and Shanken built the single joint test that finance still uses to grade every asset pricing model.

Jul 13, 2026

Glosten, Jagannathan and Runkle added one simple on-off switch to GARCH so that only losses get an extra volatility kick, and stumbled onto an uncomfortable finding about risk and reward.

Jul 13, 2026

The bid-ask spread is not one fee. Glosten and Harris built a way to separate the part that is pure overhead from the part that is the fear of trading against someone smarter.

Jul 13, 2026

In 1994, Glosten argued that an open electronic limit order book is not just one market design among many. It is the one that competitors cannot undercut.

Jul 13, 2026

Granger sidestepped the philosophy of causation entirely and asked a question data can actually answer: does knowing A's past help you predict B's future?

Jul 13, 2026

Two random walks with nothing to do with each other will produce a beautiful, highly significant regression. Granger and Newbold showed this, and half of empirical economics had to be re-examined.

Jul 13, 2026

Griffin and Shams followed the blockchain and found that freshly issued Tether flowed into Bitcoin right after price dips, propping the market up. One entity did most of it.

Jul 13, 2026

You can be a great investor with a terrible hit rate, as long as you make enough independent bets. Grinold turned that idea into the one formula every active manager should know.

Jul 13, 2026

If prices already reflect all information, nobody has any reason to gather information, and then prices can't reflect it. Markets can only be efficient if they are a little bit inefficient.

Jul 13, 2026

They fed nearly a century of stock data into every major machine learning model and asked a blunt question: which one actually predicts returns?

Jul 13, 2026

Making markets is easy until you are stuck holding the thing nobody else wants. This paper turned the market maker's nightmare into closed-form quotes you can actually compute.

Jul 13, 2026

Most market making models let you quote any price you like. Real markets have ticks, queues, and the brutal fact that being second in line means you do not get filled.

Jul 13, 2026

Three Fed economists published the daily US Treasury yield curve back to 1961, and gave it away. It became the shared foundation of nearly all modern bond research.

Jul 13, 2026

Hagan and colleagues showed that local volatility models predict the smile will move the wrong way, which corrupts every hedge, and built SABR to fix it.

Jul 13, 2026

Hamilton wrote a paper whose title is the entire argument, and then proved it: the most popular detrending tool in macroeconomics manufactures cycles out of nothing.

Jul 13, 2026

Markets do not behave the same way all the time. Hamilton built a model where the world secretly flips between states, and taught it to figure out which state we are in without ever being told.

Jul 13, 2026

Most estimation methods demand that you specify the entire probability distribution of your data. Hansen showed you only need a few things you believe should average out to zero.

Jul 13, 2026

Hansen and Jagannathan found a way to test economic models without committing to any particular one: markets themselves reveal a minimum amount of volatility any valid model must have, and most models fail it badly.

Jul 13, 2026

Hansen and Lunde tested 330 volatility models against the simplest one in the family. The plain vanilla version held its own, with one instructive exception.

Jul 13, 2026

White's Reality Check had a flaw: adding useless strategies to your search made it harder to detect a genuinely good one. Hansen showed how to repair it.

Jul 13, 2026

Harrison and Kreps proved that 'no arbitrage' and 'there exists a fair-odds probability' are the same statement, turning option pricing into a branch of probability theory.

Jul 13, 2026

The paper that gave derivatives pricing its mathematical foundations, and proved that being able to hedge anything is exactly the same as having only one fair-odds measure.

Jul 13, 2026

Your backtested Sharpe ratio is inflated because you searched for it. This paper tells you exactly how much to cut it down.

Jul 13, 2026

Researchers had published hundreds of factors that 'predict' stock returns. This paper showed the statistical test they all used was far too easy to pass.

Jul 13, 2026

When one security trades in many places, which venue is actually discovering the price and which is just copying? Hasbrouck built the standard scoreboard.

Jul 13, 2026

Nobody labels their orders as high-frequency. Hasbrouck and Saar found a clever way to detect the fast traders in raw message data, then asked what they do to the market.

Jul 13, 2026

Hasbrouck built a way to ask, of any single trade, how much of the price move it caused was permanent news and how much was noise that faded.

Jul 13, 2026

HJM proved that once you choose how volatile forward rates are, no-arbitrage forces everything else. You do not get to pick the drift. It picks itself.

Jul 13, 2026

Everyone argued about whether algorithmic trading helped or hurt liquidity. These three found a clean experiment that could actually answer it.

Jul 13, 2026

Heston let volatility be random, let it fall when the market rises, and still managed to get a formula you can compute in microseconds. That combination made it the industry standard.

Jul 13, 2026

Ho and Lee flipped interest rate modelling on its head: take today's yield curve as a fact, then ask how it is allowed to move. Every model desks actually use now does this.

Jul 13, 2026

The first proper answer to the market maker's daily question: given the position I am stuck with, where exactly do I put my bid and my ask?

Jul 13, 2026

Neural networks used to forget everything more than a few steps back. Hochreiter and Schmidhuber built a memory cell that could hold on, and it became the backbone of financial deep learning.

Jul 13, 2026

How do you split a wiggly economic series into 'the long-run trend' and 'the cycle around it'? Hodrick and Prescott gave an answer so convenient that it took over macroeconomics.

Jul 13, 2026

Hoerl and Kennard proved something heretical in 1970: a biased estimator can beat the unbiased one. Half a century later it is the foundation of every regularized model in finance.

Jul 13, 2026

Hong and Stein got momentum and reversal from two groups of ordinary, limited investors, without needing anyone to be psychologically broken.

Jul 13, 2026

Instead of hunting through data for patterns, these three authors started from a textbook theory of how firms decide to invest, and derived a four-factor model that digested most of the anomaly zoo.

Jul 13, 2026

Hou, Xue and Zhang rebuilt every published stock return anomaly they could find using careful methods. Nearly two thirds of them simply were not there.

Jul 13, 2026

If your model of how trading moves prices is even slightly the wrong shape, it hands you a machine for making free money. Huberman and Stanzl proved which shape is the only safe one.

Jul 13, 2026

Vasicek had good economics but the wrong prices. Ho-Lee had the right prices but bad economics. Hull and White welded them together, and desks have used the result ever since.

Jul 13, 2026

Hull and White let volatility wander randomly on its own, and found that Black-Scholes systematically misprices options, in a pattern that looks exactly like the smile.

Jul 13, 2026

Banning short sales is theoretically wrong: it can only make the portfolio worse. Jagannathan and Ma showed it makes portfolios better, and revealed the hidden reason why.

Jul 13, 2026

An option on a coupon bond looks hopelessly complicated. Jamshidian found the change of view that turns it into a bundle of simple options you already know how to price.

Jul 13, 2026

Instead of modelling why a company fails, Jarrow and Turnbull treated default as a random bolt of lightning and calibrated its odds straight from bond prices.

Jul 13, 2026

Michael Jensen invented a way to separate genuine skill from mere risk-taking, pointed it at 115 mutual funds, and found that after fees the average manager destroyed value.

Jul 13, 2026

Engle and Granger could test whether two wandering series were tied together. Johansen showed how to find all the invisible ropes binding a whole basket of them, in one shot.

Jul 13, 2026

Historical average returns are terrible forecasts. Jorion showed that dragging them all toward a common anchor, deliberately making them less accurate individually, produces better portfolios.

Jul 13, 2026

Your risk system reports a Value-at-Risk of 10 million dollars. Jorion asks the question nobody was asking: how confident are you in the 10?

Jul 13, 2026

Term structure models were famously impossible to estimate reliably. JSZ found a way to rewrite them that makes estimation almost instant, and revealed something surprising about no-arbitrage along the way.

Jul 13, 2026

Tversky and Kahneman catalogued the rules of thumb people use to judge probability, and showed exactly how each one fails in predictable, exploitable ways.

Jul 13, 2026

Hedge funds bought satellite photos of retailer parking lots, counted the cars, and knew the earnings before anyone else. The paper that measured exactly how much that was worth.

Jul 13, 2026

Two funds trade the same stock and pay wildly different costs. Keim and Madhavan showed the difference is not luck, it is what kind of investor you are.

Jul 13, 2026

Do cheap stocks earn more because they are riskier, or because the market is wrong? A machine learning twist on PCA suggests the answer is risk, and that the two camps were arguing past each other.

Jul 13, 2026

In one week of August 2007, quant equity funds that had nothing to do with subprime lost enormous amounts of money and then made most of it back. Khandani and Lo reconstructed what happened.

Jul 13, 2026

If you are graded against the volume-weighted average price, the safe play is to trade like the volume curve. Konishi worked out when you should not.

Jul 13, 2026

Every unit root test asks 'can you prove this series is stable?' Four authors asked the opposite question, and discovered that the honest answer to both is often 'I don't know.'

Jul 13, 2026

Everyone concluded that equal weighting beats optimization. Kritzman, Page and Turkington argued the studies had been feeding optimizers absurd forecasts, and that with sane inputs optimization wins.

Jul 13, 2026

A sleepy utility and a frenzied tech stock look nothing alike. Kyle and Obizhaeva claim that if you measure time in the right units, they are the same market.

Jul 13, 2026

Lakonishok, Shleifer and Vishny argued that cheap stocks beat expensive ones not because they are riskier, but because investors keep projecting the recent past into the distant future.

Jul 13, 2026

For months, the market valued 3Com's stake in Palm at more than all of 3Com, implying the rest of the company was worth less than nothing. Nobody could fix it, because you couldn't borrow the shares to short.

Jul 13, 2026

The historical covariance matrix is a noisy mess, and optimizers eat the noise. Ledoit and Wolf showed how to pull the extreme numbers back toward sanity, on purpose, by exactly the right amount.

Jul 13, 2026

The tape tells you a trade happened but not who was the aggressor. Lee and Ready wrote the rule of thumb that half of empirical microstructure quietly depends on.

Jul 13, 2026

Closed-end funds trade at a discount to the stocks they hold, the discounts all move together, and they move with small-cap returns. Lee, Shleifer and Thaler said that is what sentiment looks like.

Jul 13, 2026

Black-Scholes says hedge continuously. Leland pointed out that continuous hedging with real trading costs would cost you an infinite amount of money.

Jul 13, 2026

Big stocks and small stocks look nothing alike, until you rescale them. Then every price impact curve collapses onto a single universal shape.

Jul 13, 2026

Sharpe usually gets the credit, but John Lintner derived the same risk-return law independently, and he did it while trying to answer a practical question: which corporate projects are actually worth funding?

Jul 13, 2026

Litterman and Scheinkman pointed a statistical microscope at bond returns and found that almost everything the yield curve does is three simple movements.

Jul 13, 2026

Liu and Tsyvinski tested cryptocurrency returns against every standard risk factor in finance. None of them worked. What did work was momentum and pure attention.

Jul 13, 2026

A good time-series model should squeeze all the predictable structure out of the data, leaving pure noise. Ljung and Box built the standard check for whether it actually did.

Jul 13, 2026

Everyone assumed contrarian trading made money because investors overreact. Lo and MacKinlay proved you could earn those profits even if no single stock overreacted at all.

Jul 13, 2026

Longstaff, Mithal and Neis used the young credit default swap market as a clean thermometer for default risk, and finally measured how much of a corporate bond's extra yield is about default and how much is about illiquidity.

Jul 13, 2026

Longstaff and Schwartz solved American option pricing by simulation with an idea so simple it is almost cheeky: run a regression to guess what holding on is worth.

Jul 13, 2026

Longstaff and Schwartz built the first credit model where interest rate risk and default risk talk to each other, and found the surprising result that credit spreads shrink when rates go up.

Jul 13, 2026

Markowitz's optimizer treats every asset as a potential substitute for every other one, which is why it goes haywire. HRP groups similar assets into a tree first, then allocates down the branches.

Jul 13, 2026

Researchers were measuring the tone of company filings with a psychology dictionary that thinks the words 'cost', 'liability' and 'tax' are negative. In finance, they are just Tuesday.

Jul 13, 2026

Robert Lucas built the simplest possible economy, one tree that drops fruit, and used it to show that an asset's price depends not on how risky it is but on whether it pays off when you are hungry.

Jul 13, 2026

A bond's maturity date tells you when the last payment arrives, not when your money actually comes back. Macaulay invented a better clock, and it still runs the bond market.

Jul 13, 2026

Madan, Carr and Chang got fat tails and skew not by adding jumps to a random walk, but by letting the clock itself tick at a random rate.

Jul 13, 2026

A structural model of the trade-by-trade price that separates real news, the cost of immediacy, and the mechanical fact that order flow is predictable.

Jul 13, 2026

A 60/40 portfolio puts 60 percent of the money in stocks and around 90 percent of the risk. Maillard, Roncalli and Teiletche worked out the portfolio where every asset contributes the same amount of risk.

Jul 13, 2026

Bitcoin traded at wildly different prices on different exchanges for weeks at a time. Makarov and Schoar showed the gaps were real, huge, and explained almost entirely by where you live.

Jul 13, 2026

Margrabe priced the right to swap one asset for another, and discovered that when both sides are risky, the interest rate vanishes and only relative risk matters.

Jul 13, 2026

McLean and Pontiff tracked 97 published return predictors and found their profits fell by more than half after publication. Writing the paper destroys the trade.

Jul 13, 2026

Mehra and Prescott showed that the extra return stocks have paid over safe bonds is so large that no sensible model of human risk aversion can explain it, and forty years later nobody has convincingly solved it.

Jul 13, 2026

Menkveld got the trading records of one large HFT and showed exactly how it made money: earn on the spread, lose on the position, repeat a thousand times a day.

Jul 13, 2026

Merton showed that owning stock in an indebted company is exactly like owning a call option on the company's assets, and that one insight turned credit risk into option pricing.

Jul 13, 2026

The CAPM assumes the world ends after one period. Merton asked what happens if it doesn't, and discovered that investors will pay up for assets that protect them against a worsening future, which is where multi-factor models come from.

Jul 13, 2026

Black-Scholes assumes prices glide. Merton asked what happens when they gap, and found that the perfect hedge quietly stops working.

Jul 13, 2026

Merton solved the lifetime investing problem in continuous time and got a shockingly simple answer: hold a constant fraction of your wealth in stocks, and spend a constant fraction of your wealth each year.

Jul 13, 2026

Portfolio optimizers are mathematically flawless and practically distrusted. Michaud explained the paradox: they systematically fall in love with your worst estimates.

Jul 13, 2026

The models are not the problem. Lopez de Prado lists the ten ways quant teams misuse machine learning, and most of them are about how research is organised, not which algorithm you picked.

Jul 13, 2026

Firms spend fortunes to shave microseconds off their trading. This paper asks the obvious question nobody had answered: how much is that speed actually worth, in dollars?

Jul 13, 2026

Four authors turned volatility from a hidden quantity you infer with a fancy model into a number you simply observe, and then beat the fancy models with an almost embarrassingly simple one.

Jul 13, 2026

A young Norwegian economist wrote the shortest, tidiest derivation of the CAPM, and in doing so nailed down the one thing Sharpe and Lintner had left slightly vague: what 'equilibrium' actually means.

Jul 13, 2026

Digital cash had one unsolvable flaw: you can copy a file. Nakamoto solved it by making the ledger public and making lying more expensive than telling the truth.

Jul 13, 2026

Standard GARCH treats a 5% crash and a 5% rally as identical shocks. Nelson showed markets do not, and built the model that finally lets falling prices scare volatility more than rising ones.

Jul 13, 2026

If recessions are temporary dips below a trend, they eventually heal. Nelson and Plosser found that most economic data says they do not heal at all.

Jul 13, 2026

Nelson and Siegel found a formula with a handful of parameters that draws almost any yield curve the market can produce. Central banks have used it ever since.

Jul 13, 2026

Instead of assuming a model of the market and solving it, this paper let an algorithm learn how to execute orders by trial and error against a year and a half of real order book data.

Jul 13, 2026

Your regression's coefficient is probably fine. Your t-statistic is probably a lie. Newey and West wrote the short fix that made overlapping-return studies honest.

Jul 13, 2026

Novy-Marx showed that a boring line near the top of the income statement, gross profits, predicts stock returns about as well as the famous value ratio, and that the two work beautifully together.

Jul 13, 2026

Liquidity is not a fixed cost you pay, it is a resource that regenerates. Once you model the refilling, the best way to trade a big order stops being a smooth trickle.

Jul 13, 2026

Odean got hold of real brokerage records and proved investors sell winners far more readily than losers, and that the winners they sell go on to beat the losers they keep.

Jul 13, 2026

Odean checked whether investors' trades actually improved their portfolios. On average, the stocks they bought went on to lag the stocks they sold, before even counting commissions.

Jul 13, 2026

O'Hara argues that machine markets are not just faster human markets. The concepts we inherited, the trade, the day, even the price, need rebuilding.

Jul 13, 2026

Pastor and Stambaugh argued the real danger is not that a stock is hard to trade, but that it becomes hard to trade exactly when everything else does too.

Jul 13, 2026

Grade volatility forecasts with the wrong scoring rule and you can crown the worse model as the winner. Patton worked out exactly which scoring rules are safe, and the list is short.

Jul 13, 2026

Your backtest made money. Your fund did not. Perold explained exactly where the difference went, and gave the industry the number it still uses to measure it.

Jul 13, 2026

Rebalancing to a fixed mix and buying portfolio insurance are opposite strategies with opposite payoffs. Perold and Sharpe showed there is no free lunch in choosing between them.

Jul 13, 2026

Economists had concluded that almost every macro series was a random walk. Perron showed that a single ignored event, like the 1929 crash, was enough to manufacture that conclusion out of thin air.

Jul 13, 2026

Dickey and Fuller told you to model the noise in your data before testing it. Phillips and Perron found a way to stop modelling the noise and just correct for it afterwards.

Jul 13, 2026

The plain bootstrap destroys the time structure in financial data. Politis and Romano found a way to resample in blocks of random length that keeps it intact.

Jul 13, 2026

Poon and Granger read two decades of volatility forecasting research so you do not have to, and reported which methods win, which lose, and why so many comparisons were meaningless.

Jul 13, 2026

A method that answers the question every quant should ask: what are the odds my best-looking strategy is just the luckiest of my failures?

Jul 13, 2026

Kahneman and Tversky showed people don't judge outcomes by final wealth. They judge gains and losses against a reference point, and losses sting far more than equal gains feel good.

Jul 13, 2026

Four mathematicians proved that with enough tries you can produce a stunning backtest from random data, and that most published strategies are probably exactly that.

Jul 13, 2026

Everyone agreed average-loss-in-the-tail was a better risk measure than VaR. Nobody could optimize it. Rockafellar and Uryasev found the trick that turned it into a simple linear program.

Jul 13, 2026

Richard Roll pointed out something devastating: every test of the CAPM ever run is really a test of whether your chosen index is efficient, and nobody has ever observed the thing the theory is actually about.

Jul 13, 2026

Richard Roll showed that the bid-ask spread leaves a fingerprint in the price series itself, so you can estimate trading costs from closing prices alone.

Jul 13, 2026

Ross's arbitrage pricing theory said returns are driven by a handful of factors but refused to say what they were, so Roll and Ross let a statistical algorithm find them, and it found three or four.

Jul 13, 2026

Eight years before Fama and French, three practitioners published the book-to-price effect and the short-term reversal effect, and titled the paper with unusual confidence.

Jul 13, 2026

Split volatility into the market's steady grind and its sudden shocks, and you learn something useful: the grind persists for months, the shocks are forgotten almost immediately.

Jul 13, 2026

A.D. Roy published a portfolio theory in the same year as Markowitz, built on a far more human idea: investors do not want to optimize a tradeoff, they want to avoid disaster.

Jul 13, 2026

Rubinstein's version of the implied tree starts from the end and works backwards, choosing the smoothest distribution that fits the market rather than forcing an exact fit to noisy prices.

Jul 13, 2026

The original Dickey-Fuller test only worked if you knew the exact structure of your data's noise. Said and Dickey proved you can approximate any noise well enough, and made the test usable on everything.

Jul 13, 2026

Everyone believes young investors should hold more stocks because they have time to recover. Samuelson proved that under standard assumptions, your horizon should not change your allocation at all.

Jul 13, 2026

Scharfstein and Stein showed that managers copy each other on purpose, ignoring their own research, because being wrong alongside everybody else is far safer than being wrong alone.

Jul 13, 2026

Schwert gathered stock market data back to 1857 and tried to explain why volatility rises and falls. The most striking result is how badly the obvious explanations fail.

Jul 13, 2026

Every factor test estimates betas first and then uses them as if they were facts. Shanken showed that ignoring the noise in that first step makes your final results look far more convincing than they really are.

Jul 13, 2026

Before 1966, funds were ranked by returns alone, which rewarded anyone willing to gamble. Sharpe proposed dividing return by risk, and accidentally created the most quoted number in all of finance.

Jul 13, 2026

Shefrin and Statman named the most expensive habit in retail investing: cashing in gains far too eagerly while clinging to losses until they rot.

Jul 13, 2026

Shiller compared the wild swings of the stock market against the calm, steady dividends those prices are supposed to be forecasting, and found the market moves many times more than any rational forecast could justify.

Jul 13, 2026

Shleifer and Vishny explained why the smart money fails to fix mispricing: arbitrageurs trade with other people's capital, and that capital runs away exactly when the opportunity is best.

Jul 13, 2026

Sims looked at the giant macro models of his day, declared their assumptions 'incredible,' and proposed something radically humbler: just let every variable depend on the past of every other variable.

Jul 13, 2026

They trained a deep network on billions of market events and discovered that a single model, trained on all stocks together, beats a model tuned to each stock individually.

Jul 13, 2026

Sloan showed that earnings made of cash are far more durable than earnings made of accounting estimates, and that the market cheerfully fails to notice the difference.

Jul 13, 2026

Regressing returns on dividend yield seems harmless. Stambaugh showed that the very structure of that regression manufactures predictability that is not there.

Jul 13, 2026

Stambaugh, Yu and Yuan showed that almost all the profit in stock market anomalies comes from the stocks you should sell, and only after periods when investors are euphoric.

Jul 13, 2026

You have hundreds of possible predictors and only a few hundred observations. Stock and Watson showed how to squeeze them all into a handful of factors and forecast better than anyone.

Jul 13, 2026

Hans Stoll argued that a market maker is not selling stock. They are selling immediacy, and the spread is the price of that service.

Jul 13, 2026

Svensson added one more bump to Nelson-Siegel and turned a curve-fitting formula into a tool for reading what the market expects from monetary policy.

Jul 13, 2026

Realized volatility should improve as you sample prices more finely. In practice it falls apart, because tick data is mostly noise. Three authors found a way to use the noise against itself.

Jul 13, 2026

GARCH says today's volatility is calculable from yesterday's returns. Taylor said volatility has a mind of its own. Two philosophies, and the second one is harder but arguably more honest.

Jul 13, 2026

Tetlock counted the negative words in one Wall Street Journal column each day and found that media pessimism pushed prices down temporarily before they bounced back.

Jul 13, 2026

Thaler showed people file money into separate mental folders and refuse to move it between them, which breaks the most basic assumption in economics.

Jul 13, 2026

Ed Thorp took a formula from information theory, used it to beat casinos, then used it to run a hedge fund. His survey explains what the Kelly criterion really promises, and what it does not.

Jul 13, 2026

Tibshirani found a way to make a regression automatically set the coefficients of worthless predictors to exactly zero, solving variable selection and overfitting in one stroke.

Jul 13, 2026

Forget comparing stocks to each other. Just ask whether an asset went up over the last year, and if so, buy it. Moskowitz, Ooi and Pedersen found this works on nearly every futures market on earth.

Jul 13, 2026

Tobin proved that choosing which risky assets to own and choosing how much risk to take are two completely separate decisions, which is why your cautious grandmother and a hedge fund should hold the same stocks.

Jul 13, 2026

Why does a large order cost the square root of its size? Because the order book you can see is a tiny sliver of a hidden, V-shaped pool of latent liquidity.

Jul 13, 2026

Your analysts have opinions on twenty stocks. The market has thousands. Treynor and Black worked out exactly how to blend a handful of strong views into an index portfolio without wrecking it.

Jul 13, 2026

Jack Treynor derived the CAPM before Sharpe, circulated it as a typescript, never published it, and watched three other men win a Nobel Prize for the idea.

Jul 13, 2026

A genuine market timer should be aggressive in rallies and defensive in crashes, which means their returns should curve upward. Treynor and Mazuy went looking for that curve and mostly found a straight line.

Jul 13, 2026

Asness, Moskowitz and Pedersen took value and momentum out of the US stock market and ran them on bonds, currencies, commodities and foreign equities. Both worked almost everywhere, and they hedge each other.

Jul 13, 2026

Vasicek showed that if you model just the overnight rate and forbid free money, the entire yield curve follows. It was the Black-Scholes moment for bonds.

Jul 13, 2026

Standard theory says who buys a bond cannot affect its price. Vayanos and Vila built the model in which it can, and gave central bankers the theory behind quantitative easing.

Jul 13, 2026

When markets got too fast for the clock, Easley, Lopez de Prado and O'Hara rebuilt the informed-trading gauge to run on volume instead of time.

Jul 13, 2026

Ordinary regression assumes every data point is equally noisy. Financial data laughs at that. White showed how to get honest uncertainty without ever modelling the noise.

Jul 13, 2026

A bootstrap procedure that tests whether the winner of your strategy search actually beats the benchmark, or whether someone was always going to win.

Jul 13, 2026

Money has to do three things. Yermack ran Bitcoin through each test in turn and concluded it works as a speculative asset, not as a currency.

Jul 13, 2026

The square-root law of market impact fits beautifully over a narrow range of order sizes. Zarinelli and colleagues zoomed out and found something that fits far better.

Jul 13, 2026

Perron fixed unit root testing by telling the test when the break happened. Zivot and Andrews pointed out that choosing the break date after looking at the chart is data mining, and built the version that finds it for you.

Jul 13, 2026

The lasso picks one variable from a group of correlated ones and arbitrarily kills the rest. Zou and Hastie found the fix, and it matters enormously in finance.

Jul 13, 2026

Sell a giant block of stock quickly and you crush the price; sell it slowly and the market might move against you. This paper found the sweet spot.

Jul 6, 2026

A market maker has to quote a buy and a sell price at once, and manage the pile of stock it accidentally accumulates. This 2008 paper solved how.

Jul 6, 2026

Markowitz's portfolio math was brilliant but wildly oversensitive to your inputs. Two Goldman Sachs researchers found an elegant way to tame it.

Jul 6, 2026

The 1973 paper that showed how to put a fair price on an option, and accidentally launched a trillion-dollar industry.

Jul 6, 2026

A 1986 tweak to Engle's idea that made volatility forecasting simple enough for everyone, and it's still the everyday workhorse decades later.

Jul 6, 2026

Carhart added momentum to the Fama-French model and used it to ask a brutal question: do winning mutual funds keep winning because of skill, or luck and cheap fees?

Jul 6, 2026

The 2014 paper explaining why a stunning backtest is usually a statistical illusion, and how to discount it once you admit how many strategies you tried.

Jul 6, 2026

The 1979 paper that gave us a way to tell whether a price series drifts off forever like a random walk, or gets pulled back toward home like a stretched spring.

Jul 6, 2026

The 1982 paper that noticed markets have quiet spells and panicky spells, and built the first model that could see a storm coming.

Jul 6, 2026

The 1987 paper that explained how two wandering, unpredictable prices can still stay tethered together, the statistical bedrock of pairs trading.

Jul 6, 2026

The paper that argued market prices already bake in everything we know, so consistently beating the market is far harder than it looks.

Jul 6, 2026

The CAPM said only market risk matters. Fama and French showed that small companies and cheap 'value' stocks earn extra, and rewrote how we measure risk.

Jul 6, 2026

The bid-ask spread isn't greed, it's a tax the market maker charges to protect itself from people who know more than it does.

Jul 6, 2026

The paper that showed stocks which have gone up recently tend to keep going up, a stubborn pattern that shouldn't exist if markets were perfectly efficient.

Jul 6, 2026

A Bell Labs scientist used information theory to answer the oldest question in gambling and investing, how big should your bet be?

Jul 6, 2026

How does someone who knows a secret cash in without giving the game away? Kyle's 1985 paper cracked the hidden logic of informed trading and market depth.

Jul 6, 2026

The 1988 paper that ran a clever statistical test and found the stock market isn't a pure coin-flip after all, there's a faint memory in the noise.

Jul 6, 2026

The 1952 paper that turned 'diversify' from folk wisdom into math, and started the whole field of quantitative portfolio management.

Jul 6, 2026

Published the same year as Black-Scholes, Merton's paper found the deeper logic and general laws that make option pricing hold together.

Jul 6, 2026

If a stock's return depends on more than just 'the market,' what should it earn? Ross's answer built the foundation of modern factor investing.

Jul 6, 2026

The paper that gave us a single number, beta, to measure how much market risk a stock carries, and a rule for what return that risk should earn.

Jul 6, 2026

Forget the movies. A grounded, jargon-free look at what a working quant's job really involves, and the myths worth dropping before you chase one.

Jul 5, 2026

Everyone obsesses over what to buy. The survivors obsess over how much. A gentle, math-light intro to why bet size matters more than bet choice, and the Kelly intuition behind it.

Jul 4, 2026

Your strategy crushed it on paper. Then you went live and it fell apart. That gap is real, predictable, and has a name for each part of it. Here's what changes when the money is real.

Jul 3, 2026

Markets are mostly random noise with a faint signal hidden inside. Telling the two apart is the whole job, and your own brain is wired to get it wrong. A fishing story.

Jul 2, 2026

No PhD from a top school? You're not locked out. A realistic look at the paths in, what actually matters to the people hiring, and how to build proof you can do the work.

Jun 30, 2026

Beginners obsess over returns. Survivors obsess over risk. A plain-English intro to drawdowns, ruin, and why the first question a pro asks is always about the downside.

Jun 28, 2026

You're not going to out-compute a hedge fund from your kitchen table. But that's not the whole story. An honest look at what edges are actually realistic for a DIY quant.

Jun 26, 2026

Scared off by talk of stochastic calculus? Good news: the math that actually gets you started is smaller and friendlier than the internet makes it sound. Here's the honest list.

Jun 24, 2026

You found something that works. Then it stops. That's not bad luck, it's the market doing exactly what it's designed to do. A plain-English look at crowding and alpha decay.

Jun 22, 2026

A backtest is the easiest place in all of finance to lie to yourself. Here are the traps, lookahead, survivorship, costs, overfitting, explained with everyday examples.

Jun 20, 2026