Larry Harris · 2003 · Oxford University Press
BeginnerAny horizonBookTestable with your data
The standard practitioner introduction: who trades and why, order types, market structures, and where spreads and liquidity come from. Read first if you are new to market plumbing.
Maureen O'Hara · 1995 · Blackwell
IntermediateAny horizonBookTestable with your data
The classic textbook on inventory and information models of price formation (Kyle, Glosten-Milgrom and descendants) in one place.
Joel Hasbrouck · 2007 · Oxford University Press
IntermediateHFTIntradayBookTestable with your data
How to actually measure spreads, price impact and price discovery from tick data. Short, rigorous and practical.
Data: Trades · L1 quotes
Álvaro Cartea, Sebastian Jaimungal, José Penalva · 2015 · Cambridge University Press
AdvancedHFTIntradayBookTestable with your data
Stochastic-control treatment of market making, optimal execution, pairs trading and order-flow signals, with models concrete enough to code.
Data: Trades · L1 quotes
Jean-Philippe Bouchaud, Julius Bonart, Jonathan Donier et al. · 2018 · Cambridge University Press
AdvancedHFTIntradayBookTestable with your data
The physics-style empirical view of the order book: long-memory order flow, square-root impact, propagator models. Essential for impact modelling.
Data: Trades · L1 quotes · L2 depth (every update)
Charles-Albert Lehalle, Sophie Laruelle · 2013 · World Scientific
IntermediateHFTIntradayBookTestable with your data
Fragmented markets, dark pools, smart order routing and execution algorithms, written by people who built them.
Marcos López de Prado · 2018 · Wiley
IntermediateIntradayLFTBookTestable with your data
Information-driven bars, triple-barrier labelling, purged cross-validation and meta-labelling. Worth reading for the pitfalls even if you skip the methods.
Data: Trades · Intraday bars
Ernest P. Chan · 2013 · Wiley
BeginnerIntradayLFTBookTestable with your data
Hands-on mean-reversion and momentum strategies with honest backtesting caveats. A gentle on-ramp to stat arb.
Data: Daily bars · Intraday bars
Rishi K. Narang · 2013 · Wiley
BeginnerAny horizonBookTestable with your data
Non-mathematical tour of how a quant shop is wired: alpha, risk, cost and portfolio models, execution and infrastructure.
Jim Gatheral · 2006 · Wiley
AdvancedIntradayLFTBookTestable with your data
Stochastic and local volatility, the SVI parameterization, and how practitioners actually fit and trade the surface.
Data: L1 quotes
Sheldon Natenberg · 2015 · McGraw-Hill
BeginnerIntradayLFTBookTestable with your data
The market maker's introduction to options: greeks, volatility, spreads and risk. A good first read for index-options traders.
Richard C. Grinold, Ronald N. Kahn · 2000 · McGraw-Hill
IntermediateLFTBookTestable with your data
The fundamental law of active management (IR ≈ IC × √breadth) and the alpha-forecasting framework most quant equity shops use.
Ananth Madhavan · 2000 · Journal of Financial Markets
BeginnerAny horizonSurveyFree to readTestable with your data
Compact survey of price formation, market structure and transparency. The best short paper to start the academic literature.
Albert S. Kyle · 1985 · Econometrica
IntermediateAny horizonTheoryTestable with your data
Introduces Kyle's lambda: an informed trader hides in noise flow and prices move linearly with net order flow. The root of most impact models.
Data: Trades · L1 quotes
Lawrence R. Glosten, Paul R. Milgrom · 1985 · Journal of Financial Economics
IntermediateAny horizonTheoryTestable with your data
Spreads exist even with zero costs because market makers lose to informed traders. The core intuition behind adverse selection and toxic flow.
Data: Trades · L1 quotes
Thomas Ho, Hans R. Stoll · 1981 · Journal of Financial Economics
IntermediateHFTIntradayTheoryTestable with your data
The first inventory-based dealer model: quotes shift with inventory to control risk. Avellaneda-Stoikov's direct ancestor.
Richard Roll · 1984 · The Journal of Finance
BeginnerIntradayLFTMethodTestable with your data
Estimates the effective spread from the negative autocovariance of price changes. Handy when you only have trade prices.
Data: Daily bars · Trades
Charles M. C. Lee, Mark J. Ready · 1991 · The Journal of Finance
BeginnerHFTIntradayMethodTestable with your data
The standard algorithm for signing trades as buys or sells (quote rule plus tick rule). Needed whenever your feed lacks the aggressor side.
Data: Trades · L1 quotes
Joel Hasbrouck · 1991 · The Journal of Finance
IntermediateHFTIntradayMethodTestable with your data
A VAR of trades and quote revisions measures the permanent price impact of trades. Still the template for impact estimation.
Data: Trades · L1 quotes
Joel Hasbrouck · 1995 · The Journal of Finance
AdvancedHFTIntradayMethodTestable with your data
Information share: which venue or instrument leads price discovery. Directly applicable to Nifty futures vs cash vs options lead-lag.
Data: L1 quotes
Yakov Amihud · 2002 · Journal of Financial Markets
BeginnerLFTEmpiricalFree to readTestable with your data
The Amihud illiquidity ratio (|return| / traded value): a dead-simple liquidity measure from daily data, plus evidence that illiquidity is priced.
Data: Daily bars
Bruno Biais, Pierre Hillion, Chester Spatt · 1995 · The Journal of Finance
IntermediateHFTEmpiricalTestable with your data
Classic empirical LOB study: thin books attract limit orders, thick books attract market orders, and order flow is autocorrelated.
Data: L2 depth (every update)
Sanford J. Grossman, Merton H. Miller · 1988 · The Journal of Finance
BeginnerAny horizonTheoryTestable with your data
Liquidity is the price of immediacy: market makers are paid for bridging the gap between the arrivals of buyers and sellers.
Maureen O'Hara · 2015 · Journal of Financial Economics
BeginnerHFTSurveyTestable with your data
Argues classic microstructure must be rethought for HFT: orders, not trades, now carry the information.
Martin D. Gould, Mason A. Porter, Stacy Williams et al. · 2013 · Quantitative Finance
BeginnerHFTSurveyFree to readTestable with your data
The best single survey of LOB mechanics, stylized facts and models. Start here for anything order-book related.
Data: L2 depth (every update)
Jean-Philippe Bouchaud, Marc Mézard, Marc Potters · 2002 · Quantitative Finance
IntermediateHFTEmpiricalFree to readTestable with your data
Limit order placement follows a power law away from the best price and the average book is hump-shaped. Key stylized facts for any simulator.
Data: L2 depth (snapshots)
Jean-Philippe Bouchaud, Yuval Gefen, Marc Potters et al. · 2004 · Quantitative Finance
AdvancedHFTIntradayEmpiricalFree to readTestable with your data
Order-flow signs have long memory yet prices stay diffusive, because impact is transient. Introduces the propagator model.
Data: Trades · L1 quotes
Fabrizio Lillo, J. Doyne Farmer · 2004 · Studies in Nonlinear Dynamics & Econometrics
IntermediateHFTIntradayEmpiricalFree to readTestable with your data
The signs of market orders are strongly long-memory (large orders get split). Any order-flow signal has to account for this.
Data: Trades
Rama Cont, Sasha Stoikov, Rishi Talreja · 2010 · Operations Research
AdvancedHFTTheoryFree to readTestable with your data
Queueing model of the LOB with Poisson arrivals and cancellations. Computes probabilities like 'the mid moves up next' or 'my order fills before the mid moves'.
Data: L2 depth (every update)
Rama Cont, Adrien de Larrard · 2013 · SIAM Journal on Financial Mathematics
AdvancedHFTTheoryFree to readTestable with your data
Models only the queue sizes at the best bid and ask, giving closed-form probabilities of the next price move from queue imbalance.
Data: L1 quotes
Weibing Huang, Charles-Albert Lehalle, Mathieu Rosenbaum · 2015 · Journal of the American Statistical Association
AdvancedHFTTheoryFree to readTestable with your data
Order arrival rates depend on current queue sizes. A realistic, calibratable LOB simulator for backtesting passive strategies.
Data: L2 depth (every update)
Sasha Stoikov · 2018 · Quantitative Finance
IntermediateHFTMethodTestable with your data
A fair-price estimator built from bid, ask, imbalance and spread that beats the mid and the weighted mid as a short-horizon predictor. Very practical.
Data: L1 quotes
Emmanuel Bacry, Iacopo Mastromatteo, Jean-François Muzy · 2015 · Market Microstructure and Liquidity
AdvancedHFTSurveyFree to readTestable with your data
Survey of self-exciting point processes for trade arrivals, order flow and microstructure noise. The toolkit for event-time modelling.
Data: Trades · L2 depth (every update)
Ciamac C. Moallemi, Kai Yuan · 2017 · Working paper (SSRN)
AdvancedHFTTheoryFree to readTestable with your data
Puts a value on queue position. On large-tick instruments, being early in the queue is worth a meaningful fraction of the spread.
Data: L3/MBO (order-level)
Rama Cont, Arseniy Kukanov, Sasha Stoikov · 2014 · Journal of Financial Econometrics
IntermediateHFTEmpiricalFree to readTestable with your data
Order flow imbalance (OFI) at the best quotes explains short-horizon price changes linearly, with a slope inversely proportional to depth. The go-to HFT feature.
Data: L1 quotes
Ke Xu, Martin D. Gould, Sam D. Howison · 2018 · Market Microstructure and Liquidity
IntermediateHFTEmpiricalFree to readTestable with your data
Extends OFI beyond the touch. Deeper book levels add explanatory power for mid-price changes.
Data: L2 depth (every update)
Álvaro Cartea, Ryan Donnelly, Sebastian Jaimungal · 2018 · Applied Mathematical Finance
AdvancedHFTTheoryFree to readTestable with your data
Volume imbalance predicts the arrival of market orders and price moves, and is built into optimal market-making and execution strategies.
Data: L1 quotes
Alexander Lipton, Umberto Pesavento, Michael G. Sotiropoulos · 2013 · arXiv preprint
IntermediateHFTEmpiricalFree to readTestable with your data
Quote imbalance at the touch predicts the direction of the next mid-price move. A short practitioner paper.
Data: L1 quotes · Trades
Petter N. Kolm, Jeremy Turiel, Nicholas Westray · 2023 · Mathematical Finance
AdvancedHFTEmpiricalFree to readTestable with your data
Deep nets fed stationary order-flow-imbalance features beat nets fed raw book states, with predictability strongest at the shortest horizons.
Data: L2 depth (every update)
Rama Cont, Mihai Cucuringu, Chao Zhang · 2023 · Quantitative Finance
AdvancedHFTIntradayEmpiricalFree to readTestable with your data
Multi-level OFI integrated across levels; cross-asset OFI adds little contemporaneously, but lagged cross-asset OFI helps forecast returns.
Data: L2 depth (every update)
Justin Sirignano, Rama Cont · 2019 · Quantitative Finance
AdvancedHFTEmpiricalFree to readTestable with your data
One deep model trained on hundreds of stocks predicts price direction better than stock-specific models: price formation looks universal.
Data: L2 depth (every update)
David Easley, Marcos M. López de Prado, Maureen O'Hara · 2012 · Review of Financial Studies
IntermediateHFTIntradayEmpiricalFree to readTestable with your data
VPIN: a volume-bucketed measure of order-flow toxicity, claimed to rise before the 2010 Flash Crash. Debated, but a useful lens for market makers.
Data: Trades
Steven L. Heston, Robert A. Korajczyk, Ronnie Sadka · 2010 · The Journal of Finance
IntermediateIntradayEmpiricalFree to readTestable with your data
Returns in a given half-hour interval persist at the same time of day for weeks. A simple intraday seasonality signal.
Data: Intraday bars
Lei Gao, Yufeng Han, Sophia Zhengzi Li et al. · 2018 · Journal of Financial Economics
BeginnerIntradayEmpiricalFree to readTestable with your data
The market's first half-hour return predicts its last half-hour return. Easy to test on Nifty futures.
Data: Intraday bars
Guido Baltussen, Zhi Da, Sten Lammers et al. · 2021 · Journal of Financial Economics
IntermediateIntradayEmpiricalFree to readTestable with your data
Late-day index momentum is linked to gamma hedging: short-gamma hedgers amplify moves into the close. Relevant wherever index options dominate volume.
Data: Intraday bars
Marco Avellaneda, Sasha Stoikov · 2008 · Quantitative Finance
IntermediateHFTTheoryTestable with your data
Quote around a reservation price that skews with inventory; the spread depends on risk aversion and order-arrival intensity. The starting point for model-based market making.
Data: L1 quotes · Trades
Olivier Guéant, Charles-Albert Lehalle, Joaquin Fernandez-Tapia · 2013 · Mathematics and Financial Economics
AdvancedHFTTheoryTestable with your data
Solves Avellaneda-Stoikov with inventory limits and gives closed-form approximations. The version practitioners usually implement.
Data: L1 quotes · Trades
Álvaro Cartea, Sebastian Jaimungal, Jason Ricci · 2014 · SIAM Journal on Financial Mathematics
AdvancedHFTTheoryFree to readTestable with your data
Market making with self-exciting order flow, adverse selection and short-term alpha. Shows how to skew quotes on signals.
Data: L1 quotes · Trades
Albert J. Menkveld · 2013 · Journal of Financial Markets
BeginnerHFTEmpiricalFree to readTestable with your data
Anatomy of one HFT market maker across Chi-X and Euronext: tiny per-trade margins, flat inventory, profit from the spread minus adverse selection.
Data: Trades · L1 quotes
Sasha Stoikov, Mehmet Sağlam · 2009 · Review of Derivatives Research
AdvancedHFTTheoryTestable with your data
Extends inventory-based market making to options, where delta and vega exposure drive the quote skew.
Thomas Spooner, John Fearnley, Rahul Savani et al. · 2018 · AAMAS 2018
IntermediateHFTEmpiricalFree to readTestable with your data
An RL agent with a carefully shaped reward (asymmetrically dampened PnL) learns a competitive quoting policy on LOB data.
Data: L2 depth (every update)
André F. Perold · 1988 · The Journal of Portfolio Management
BeginnerIntradayLFTMethodTestable with your data
Defines implementation shortfall, the gap between paper and real portfolio returns. The benchmark most execution algorithms target.
Dimitris Bertsimas, Andrew W. Lo · 1998 · Journal of Financial Markets
IntermediateIntradayTheoryTestable with your data
Dynamic programming for splitting a large order to minimise expected cost. The first formal optimal-execution model.
Robert Almgren, Neil Chriss · 2001 · The Journal of Risk
IntermediateIntradayLFTTheoryTestable with your data
The mean-variance trade-off between impact cost and timing risk gives an efficient frontier of execution schedules. The model behind most implementation-shortfall algos.
Robert Almgren, Chee Thum, Emmanuel Hauptmann et al. · 2005 · Risk (published as "Equity market impact")
IntermediateIntradayLFTEmpiricalFree to readTestable with your data
Calibrates permanent and temporary impact on a large broker's US equity orders. Temporary impact grows roughly with participation rate to the power 0.6.
Data: Trades
Anna A. Obizhaeva, Jiang Wang · 2013 · Journal of Financial Markets
AdvancedIntradayTheoryFree to readTestable with your data
Execution against a resilient order book: trade an initial block, then continuously, then a final block, with the mix set by how fast the book refills.
Data: L2 depth (snapshots)
Bence Tóth, Yves Lempérière, Cyril Deremble et al. · 2011 · Physical Review X
IntermediateIntradayLFTEmpiricalFree to readTestable with your data
Evidence and a mechanism for the square-root impact law (impact ∝ σ√(Q/V)) from a fund's own metaorders.
Data: Trades
Andrea Frazzini, Ronen Israel, Tobias J. Moskowitz · 2018 · Working paper (SSRN)
IntermediateLFTEmpiricalFree to readTestable with your data
Real execution costs from a large asset manager's live trades are far below the academic estimates, so more anomalies survive costs than commonly assumed.
Data: Trades
Rama Cont, Arseniy Kukanov · 2017 · Quantitative Finance
AdvancedHFTIntradayTheoryFree to readTestable with your data
How to split an order across venues and between limit and market orders given fill probabilities. A blueprint for a smart order router.
Data: L2 depth (every update)
Yuriy Nevmyvaka, Yi Feng, Michael Kearns · 2006 · ICML 2006
IntermediateIntradayEmpiricalTestable with your data
Early large-scale RL on NASDAQ order book data. Learned policies cut execution costs compared with simple submit-and-leave strategies.
Data: L2 depth (every update)
Terrence Hendershott, Charles M. Jones, Albert J. Menkveld · 2011 · The Journal of Finance
BeginnerHFTEmpiricalTestable with your data
Using the NYSE autoquote rollout as a natural experiment, algorithmic trading narrows spreads and makes quotes more informative.
Data: Trades · L1 quotes
Jonathan Brogaard, Terrence Hendershott, Ryan Riordan · 2014 · Review of Financial Studies
IntermediateHFTEmpiricalFree to readTestable with your data
HFTs trade in the direction of permanent price changes and against transitory errors. Their market orders are informed; their limit orders get adversely selected.
Data: Trades · L1 quotes
Eric Budish, Peter Cramton, John Shim · 2015 · The Quarterly Journal of Economics
IntermediateHFTEmpiricalFree to readTestable with your data
The ES-SPY correlation breaks down at millisecond horizons, creating mechanical sniping profits. Proposes frequent batch auctions as the fix.
Data: L1 quotes
Matteo Aquilina, Eric Budish, Peter O'Neill · 2022 · The Quarterly Journal of Economics
IntermediateHFTEmpiricalFree to readTestable with your data
Exchange message data from the LSE shows latency races are frequent, last microseconds and impose a measurable tax on liquidity.
Data: L3/MBO (order-level)
Andrei Kirilenko, Albert S. Kyle, Mehrdad Samadi et al. · 2017 · The Journal of Finance
IntermediateHFTEmpiricalFree to readTestable with your data
Account-level E-mini data from 6 May 2010: HFTs did not trigger the crash but amplified it by passing inventory back and forth.
Data: Trades
Joel Hasbrouck, Gideon Saar · 2013 · Journal of Financial Markets
IntermediateHFTEmpiricalFree to readTestable with your data
Identifies 'strategic runs' of linked messages in NASDAQ order data; more low-latency activity goes with tighter spreads and more depth.
Data: L3/MBO (order-level)
Albert J. Menkveld · 2016 · Annual Review of Financial Economics
BeginnerHFTSurveyTestable with your data
Short survey of what research says HFTs do, earn and cost. The best entry point into the empirical HFT literature.
Jonathan Brogaard, Björn Hagströmer, Lars Nordén et al. · 2015 · Review of Financial Studies
IntermediateHFTEmpiricalFree to readTestable with your data
When NASDAQ OMX Stockholm sold a premium colocation upgrade, buyers used speed to manage inventory and avoid adverse selection, and liquidity improved.
Data: L3/MBO (order-level)
Andriy Shkilko, Konstantin Sokolov · 2020 · The Journal of Finance
IntermediateHFTEmpiricalFree to readTestable with your data
When rain disrupts microwave links and removes the speed edge, adverse selection falls and liquidity improves.
Data: Trades · L1 quotes
Chen Yao, Mao Ye · 2018 · The Review of Financial Studies
IntermediateHFTEmpiricalFree to readTestable with your data
With a binding tick size, speed buys queue priority, so HFTs provide more liquidity in large-tick stocks. Key for tick-constrained Indian names.
Data: L3/MBO (order-level)
Ciamac C. Moallemi, Mehmet Sağlam · 2013 · Operations Research
IntermediateHFTTheoryTestable with your data
Puts a number on what latency costs a trader using limit orders, and shows how that cost scales with volatility.
Gregory Laughlin, Anthony Aguirre, Joseph Grundfest · 2014 · Financial Review
BeginnerHFTEmpiricalFree to readTestable with your data
Measures the Chicago-New York latency race: microwave networks pushed information transmission close to the speed-of-light limit.
Data: Trades
Evan Gatev, William N. Goetzmann, K. Geert Rouwenhorst · 2006 · Review of Financial Studies
BeginnerLFTEmpiricalFree to readTestable with your data
The distance method: match stocks by normalised price paths, trade two-sigma divergences. The baseline every pairs paper compares against.
Data: Daily bars
Marco Avellaneda, Jeong-Hyun Lee · 2010 · Quantitative Finance
IntermediateIntradayLFTEmpiricalFree to readTestable with your data
Model each stock's residual against PCA factors or sector ETFs as a mean-reverting (OU) process and trade on s-scores. A blueprint for factor-neutral mean reversion.
Data: Daily bars
Robert J. Elliott, John van der Hoek, William P. Malcolm · 2005 · Quantitative Finance
IntermediateIntradayLFTTheoryTestable with your data
Models the spread as a mean-reverting process observed with noise and estimates it with a Kalman filter.
Christopher Krauss · 2017 · Journal of Economic Surveys
BeginnerIntradayLFTSurveyTestable with your data
Maps the pairs-trading literature into distance, cointegration, time-series, stochastic-control and other approaches, with strengths and weaknesses of each.
Christopher Krauss, Xuan Anh Do, Nicolas Huck · 2017 · European Journal of Operational Research
IntermediateLFTEmpiricalTestable with your data
Daily cross-sectional ML on lagged returns: strong early returns that fade in recent years. An honest look at alpha decay.
Data: Daily bars
Amir E. Khandani, Andrew W. Lo · 2011 · Journal of Financial Markets
IntermediateIntradayLFTEmpiricalFree to readTestable with your data
A simple contrarian strategy lost heavily in August 2007 as crowded quant books deleveraged at once. Required reading on crowding risk.
Data: Daily bars · Intraday bars
Narasimhan Jegadeesh · 1990 · The Journal of Finance
BeginnerLFTEmpiricalTestable with your data
One-month return reversal: last month's losers beat last month's winners next month. The short-term reversal effect.
Data: Daily bars
William K. Bertram · 2010 · Physica A: Statistical Mechanics and its Applications
IntermediateIntradayLFTTheoryFree to readTestable with your data
Closed-form optimal entry and exit thresholds for trading an OU spread with transaction costs.
Jorge Guijarro-Ordonez, Markus Pelger, Greg Zanotti · 2026 · Management Science
AdvancedLFTEmpiricalFree to readTestable with your data
Builds residual portfolios, extracts mean-reversion signals with a CNN-transformer and optimises allocation end to end.
Data: Daily bars
Narasimhan Jegadeesh, Sheridan Titman · 1993 · The Journal of Finance
BeginnerLFTEmpiricalTestable with your data
The original cross-sectional momentum paper: stocks that won over the past 3-12 months keep beating the losers.
Data: Daily bars
Eugene F. Fama, Kenneth R. French · 1993 · Journal of Financial Economics
BeginnerLFTEmpiricalTestable with your data
Market, size (SMB) and value (HML) factors. The baseline any equity signal must beat or neutralise.
Data: Daily bars
Tobias J. Moskowitz, Yao Hua Ooi, Lasse Heje Pedersen · 2012 · Journal of Financial Economics
BeginnerLFTEmpiricalFree to readTestable with your data
An asset's own past 12-month return predicts its next return across dozens of futures markets. The academic basis of trend following.
Data: Daily bars
Clifford S. Asness, Tobias J. Moskowitz, Lasse Heje Pedersen · 2013 · The Journal of Finance
IntermediateLFTEmpiricalFree to readTestable with your data
Value and momentum premia show up in eight markets and asset classes and are negatively correlated with each other, so combine them.
Data: Daily bars
Kent Daniel, Tobias J. Moskowitz · 2016 · Journal of Financial Economics
IntermediateLFTEmpiricalFree to readTestable with your data
Momentum crashes in panic rebounds after market falls. A dynamically scaled momentum strategy roughly doubles the Sharpe ratio.
Data: Daily bars
R. David McLean, Jeffrey Pontiff · 2016 · The Journal of Finance
BeginnerLFTEmpiricalFree to readTestable with your data
Anomaly returns are lower out of sample and lower still after publication. Expect alpha found in papers to decay.
Data: Daily bars
Campbell R. Harvey, Yan Liu, Heqing Zhu · 2016 · Review of Financial Studies
IntermediateLFTMethodFree to readTestable with your data
With hundreds of factors already tested, a new one needs a t-statistic above about 3, not 2. Multiple-testing hygiene for strategy research.
Shihao Gu, Bryan Kelly, Dacheng Xiu · 2020 · The Review of Financial Studies
IntermediateLFTEmpiricalFree to readTestable with your data
Compares linear, tree and neural models for predicting stock returns. Trees and neural nets win, driven mostly by momentum, liquidity and volatility signals.
Data: Daily bars
Robert Novy-Marx, Mihail Velikov · 2016 · Review of Financial Studies
IntermediateLFTEmpiricalTestable with your data
Most high-turnover anomalies do not survive transaction costs. Simple cost-mitigation rules, like a buy/hold spread, rescue some of them.
Data: Daily bars
David H. Bailey, Marcos López de Prado · 2014 · The Journal of Portfolio Management
IntermediateAny horizonMethodFree to readTestable with your data
Adjusts a backtest's Sharpe ratio for the number of variants tried, skew and fat tails. Tells you whether your best variant is just luck.
David H. Bailey, Jonathan M. Borwein, Marcos López de Prado et al. · 2017 · The Journal of Computational Finance
IntermediateAny horizonMethodFree to readTestable with your data
Combinatorially symmetric cross-validation estimates the probability that your in-sample winner underperforms out of sample.
David H. Bailey, Jonathan M. Borwein, Marcos López de Prado et al. · 2014 · Notices of the American Mathematical Society
BeginnerAny horizonMethodFree to readTestable with your data
A readable warning: try enough configurations and a high in-sample Sharpe is nearly guaranteed and means nothing.
Campbell R. Harvey, Yan Liu · 2015 · The Journal of Portfolio Management
BeginnerAny horizonMethodFree to readTestable with your data
Haircut Sharpe ratios for multiple testing. A practical rule for discounting backtests.
Halbert White · 2000 · Econometrica
AdvancedAny horizonMethodTestable with your data
A bootstrap test of whether the best of many strategies really beats a benchmark once you account for the search.
Andrew W. Lo · 2002 · Financial Analysts Journal
BeginnerAny horizonMethodTestable with your data
Standard errors for Sharpe ratios, and why annualising by √12 is wrong when returns are autocorrelated.
Zihao Zhang, Stefan Zohren, Stephen Roberts · 2019 · IEEE Transactions on Signal Processing
AdvancedHFTEmpiricalFree to readTestable with your data
CNN plus inception module plus LSTM on the top 10 book levels to predict mid-price direction. The most-cited deep LOB model and the usual baseline.
Data: L2 depth (snapshots)
Adamantios Ntakaris, Martin Magris, Juho Kanniainen et al. · 2018 · Journal of Forecasting
IntermediateHFTDatasetFree to readTestable with your data
FI-2010: a public 10-level LOB dataset for five Nasdaq Nordic stocks. The standard (if small) benchmark for LOB models.
Data: L2 depth (snapshots)
Alec N. Kercheval, Yuan Zhang · 2015 · Quantitative Finance
IntermediateHFTEmpiricalTestable with your data
Hand-crafted LOB features (spreads, imbalances, arrival intensities) fed to SVMs. The feature list is still worth borrowing.
Data: L2 depth (snapshots)
Matteo Prata, Giuseppe Masi, Leonardo Berti et al. · 2024 · Artificial Intelligence Review
IntermediateHFTEmpiricalFree to readTestable with your data
Re-runs 15 published deep LOB models on common data: most fail to generalise beyond FI-2010. A reality check before trusting DeepLOB-style results.
Data: L2 depth (snapshots)
David Byrd, Maria Hybinette, Tucker Hybinette Balch · 2020 · ACM SIGSIM Conference on Principles of Advanced Discrete Simulation
IntermediateHFTMethodTestable with your data
Open-source agent-based exchange simulator with nanosecond messaging. Useful for testing strategies where your own impact matters.
Data: L3/MBO (order-level)
Fischer Black, Myron Scholes · 1973 · Journal of Political Economy
BeginnerAny horizonTheoryTestable with your data
Price an option by dynamically hedging it. Know its assumptions so you know exactly where real markets break them.
Steven L. Heston · 1993 · Review of Financial Studies
AdvancedLFTTheoryTestable with your data
The workhorse stochastic-volatility model with semi-closed-form prices. Explains skew through spot-vol correlation.
Jim Gatheral, Antoine Jacquier · 2014 · Quantitative Finance
AdvancedIntradayLFTMethodFree to readTestable with your data
How to fit SVI smiles across expiries without calendar or butterfly arbitrage. Directly usable for an index-options vol surface.
Data: L1 quotes
Jim Gatheral, Thibault Jaisson, Mathieu Rosenbaum · 2018 · Quantitative Finance
AdvancedIntradayLFTEmpiricalFree to readTestable with your data
Log-volatility behaves like fractional Brownian motion with a Hurst exponent near 0.1. Rough vol forecasts well and fits the short-dated skew.
Data: Intraday bars
Fulvio Corsi · 2009 · Journal of Financial Econometrics
IntermediateIntradayLFTMethodFree to readTestable with your data
HAR-RV: regress tomorrow's realized vol on daily, weekly and monthly realized vol. A hard-to-beat forecast in a few lines of code.
Data: Intraday bars
Lan Zhang, Per A. Mykland, Yacine Aït-Sahalia · 2005 · Journal of the American Statistical Association
AdvancedHFTIntradayMethodTestable with your data
Microstructure noise biases tick-level realized volatility. The two-scales estimator corrects it.
Data: Trades
Tim Bollerslev · 1986 · Journal of Econometrics
BeginnerLFTMethodTestable with your data
GARCH(1,1): volatility clusters and mean-reverts. The baseline volatility model.
Data: Daily bars
Peter Carr, Liuren Wu · 2009 · Review of Financial Studies
IntermediateLFTEmpiricalFree to readTestable with your data
Synthesises variance swap rates from option prices. Implied variance exceeds realized on average: the premium option sellers harvest.
Data: Daily bars
Dmitriy Muravyev · 2016 · The Journal of Finance
AdvancedIntradayEmpiricalFree to readTestable with your data
Option market makers' inventory risk is priced: customer order imbalance predicts option returns.
Data: Trades · L1 quotes
Sophie Xiaoyan Ni, Neil D. Pearson, Allen M. Poteshman · 2005 · Journal of Financial Economics
IntermediateIntradayEmpiricalFree to readTestable with your data
Optionable stocks close near strike prices on expiry days, driven by hedge rebalancing and manipulation. Relevant to expiry-day dynamics in Indian index options.
Data: Daily bars
Harry Markowitz · 1952 · The Journal of Finance
BeginnerLFTTheoryTestable with your data
Mean-variance optimisation: diversification measured by covariance. The starting point of every portfolio construction step.
J. L. Kelly · 1956 · Bell System Technical Journal
BeginnerAny horizonTheoryTestable with your data
Bet the fraction that maximises the growth rate of log wealth. The basis of position sizing; most practitioners use a fraction of Kelly.
Olivier Ledoit, Michael Wolf · 2004 · The Journal of Portfolio Management
IntermediateLFTMethodFree to readTestable with your data
Shrink the noisy sample covariance toward a structured target. A free improvement for any optimiser.
Data: Daily bars
Nicolae Gârleanu, Lasse Heje Pedersen · 2013 · The Journal of Finance
AdvancedIntradayLFTTheoryFree to readTestable with your data
Trade partway toward an 'aim' portfolio that weights each signal by how slowly it decays. When costs bite, fast alphas get less weight.
Marcos López de Prado · 2016 · The Journal of Portfolio Management
IntermediateLFTMethodFree to readTestable with your data
Hierarchical risk parity: cluster assets and allocate recursively, avoiding covariance-matrix inversion.
Data: Daily bars
Igor Makarov, Antoinette Schoar · 2020 · Journal of Financial Economics
IntermediateIntradayLFTEmpiricalFree to readTestable with your data
Large, persistent price gaps across crypto exchanges, especially across countries with capital controls. Arbitrage is limited by moving capital, not by spotting the gap.
Data: Trades
SEBI Department of Economic and Policy Analysis · 2026 · SEBI
BeginnerAny horizonReportFree to readIndiaTestable with your data
Individuals' net losses in equity derivatives fell 18% to ₹91,685 crore in FY26 and loss-makers fell to 87.7%, yet those still trading lost slightly more on average. Also measures the effect of SEBI's November 2024 F&O measures.
SEBI Department of Economic and Policy Analysis · 2026 · SEBI
BeginnerAny horizonReportFree to readIndiaTestable with your data
97% of individual F&O traders are mainly option buyers and about 90% of them lose money. Option sellers lose less often but far more when they do. A sample-based picture of the retail flow on the other side of index options.
SEBI Department of Economic and Policy Analysis · 2024 · SEBI
BeginnerIntradayReportFree to readIndiaTestable with your data
71% of individual intraday traders in the cash market lost money in FY23, rising to 80% for very frequent traders. Under-30s grew to 48% of intraday traders. Based on data from the top 10 brokers.
Nidhi Aggarwal, Susan Thomas · 2014 · IGIDR Working Paper 2014-023
IntermediateHFTEmpiricalFree to readIndiaTestable with your data
Uses the introduction of co-location at NSE as a natural experiment. Stocks with more algorithmic trading show lower liquidity costs, less order imbalance and fewer extreme intraday price moves.
Data: L3/MBO (order-level)
Nidhi Aggarwal, Susan Thomas · 2019 · Journal of Futures Markets
IntermediateHFTIntradayEmpiricalFree to readIndiaTestable with your data
On NSE's highly liquid single-stock futures, futures carry about 55% of the information share around news, rising to 61% on bad news when short-sale limits push informed traders to futures. The leverage-versus-liquidity trade-off decides which market leads.
Data: Trades · L1 quotes
Ekkehart Boehmer, R. L. Shankar · 2014 · NSE-NYU Stern Working Paper
IntermediateHFTEmpiricalFree to readIndiaTestable with your data
Order-level NSE data around the co-location launch: more algorithmic trading reduces commonality in order flow, returns, liquidity and volatility across stocks, which argues against the idea that low-latency trading raises systemic risk.
Data: L3/MBO (order-level)
Vipul · 2005 · Journal of Futures Markets
BeginnerIntradayEmpiricalIndiaTestable with your data
Underlying share prices dip slightly the day before expiry and rise significantly the day after, with abnormally high expiry-day volume, mainly due to arbitrage unwinds and short-selling restrictions.
Data: Daily bars