129 questions · full solutions
Statistics
Estimators, regression intuition, time-series traps, and the standard-error scaling that decides whether a backtest means anything.
- 01Estimating the Fleet From Four Serial NumbersMedium
- 02Variance of a DieMedium · Private
- 03Sixty HeadsMedium · Private
- 04Carrying the Normal TableEasy
- 05Halving the ErrorEasy
- 06What the Slope Really IsMedium · Private
- 07Diversification ArithmeticMedium · Private
- 08π from DartsEasy
- 09MLE for an Arrival RateMedium · Private
- 10The Biased MaximumHard · Private
- 11n or n − 1?Medium · Private
- 12Method of Moments in One LineEasy
- 13Consistent, Unbiased, or Both?Hard · Private
- 14What a p-Value Is NotMedium · Private
- 15Ninety-Five Percent of What?Medium · Private
- 16How Many Orders for the A/B Test?Medium · Private
- 17One Hundred Junk SignalsMedium · Private
- 18t or z?Easy
- 19Pair Your ExperimentsMedium · Private
- 20Covariance by HandMedium · Private
- 21Long Two, Short ThreeEasy
- 22Flows Predict Returns. Or Do They?Easy
- 23The Mean Above the BarHard · Private
- 24Best of Two NormalsHard · Private
- 25Two Uniforms Past 1.5Medium · Private
- 26The R² Beauty ContestMedium · Private
- 27The Ghost RegressorHard · Private
- 28Unequal Noise, Unequal DamageMedium · Private
- 29Why Backtest t-Stats LieHard · Private
- 30Why Betas Get Pulled Toward OneMedium · Private
- 31Two Random Walks Walk Into a RegressionHard · Private
- 32AR(1) Vital SignsMedium · Private
- 33Permanent or Transitory?Hard · Private
- 34Quiet Days, Loud DaysMedium · Private
- 35Error Bars on a Track RecordHard · Private
- 36The Price of RobustnessMedium · Private
- 37Updating a Hit RateMedium · Private
- 38Sizing the Sample for a Target WidthEasy
- 39Power of a Fixed-Size ExperimentMedium · Private
- 40Covariance of Functions of a NormalHard · Private
- 41Correlation of X and X SquaredMedium · Private
- 42The Algo That Wins Everywhere but LosesMedium · Private
- 43The k-th Smallest of n UniformsMedium · Private
- 44The Sophomore Slump, QuantifiedMedium · Private
- 45Information Ratio and the Fundamental LawMedium · Private
- 46Correlation Doesn't Care About UnitsEasy
- 47Error Bars on a Fill RateEasy
- 48Three Sigma Without a DistributionEasy
- 49MLE Pinned to the BoundaryMedium · Private
- 50Who Makes It Into the Resample?Medium · Private
- 51The Smallest p-Value Money Can BuyMedium · Private
- 52The Posterior Splits the DifferenceMedium · Private
- 53Correlated by ConstructionMedium · Private
- 54Counting a Rival's OrdersMedium · Private
- 55The Slope That Noise AteMedium · Private
- 56What a 52% Hit Rate BuysMedium · Private
- 57One Number to KeepMedium · Private
- 58Two Sharpe-One Strategies, One BookMedium · Private
- 59Fitting the Tail Thickness by MomentsMedium · Private
- 60Everyone Can't Hedge EveryoneMedium · Private
- 61The Fills You Never SawHard · Private
- 62The Floor Under Every EstimatorHard · Private
- 63Conditioning Your Way to 13x Less VarianceHard · Private
- 64A Whole Interval of MLEsHard · Private
- 65Why L1 Loves the MedianHard · Private
- 66Peeking Once Costs YouHard · Private
- 67What Fraction of Your Discoveries Are Junk?Hard · Private
- 68Significant and Probably FalseHard · Private
- 69Same Data, Two p-ValuesHard · Private
- 70How Long to Certify a 0.05 CorrelationHard · Private
- 71Thirty-One Years to Prove ItHard · Private
- 72Error Bars for an Option SellerHard · Private
- 73Good Managers Lose OftenHard · Private
- 74Root-T Breaks When Returns TrendHard · Private
- 75The Beta That Grows With the ClockHard · Private
- 76The R-Squared You Get for FreeHard · Private
- 77Weight the Quiet DaysHard · Private
- 78Regress Out the Market FirstHard · Private
- 79The Correlation Regime That Wasn'tHard · Private
- 80Ticks Sharpen Vol, Never DriftHard · Private
- 81Correlations Come in TrianglesHard · Private
- 82Two Events That Can't Fully CorrelateHard · Private
- 83The Hedge That Shorts the HedgeHard · Private
- 84Hedged Within Regimes, Correlated Across ThemHard · Private
- 85One Factor to Rule the MatrixHard · Private
- 86Fat Tails From Nothing but Vol SwitchingHard · Private
- 87Error Bars on a Vol MarkHard · Private
- 88How Wobbly Is an Empirical VaR?Hard · Private
- 89Six Breaches and a VerdictHard · Private
- 90Correlated Until It MattersHard · Private
- 91Shrinking Ten Traders at OnceHard · Private
- 92Your Best Day Means NothingHard · Private
- 93Bootstrapping Away the AutocorrelationHard · Private
- 94The Free Lunch in Mirrored DrawsHard · Private
- 95Factors From Thin AirHard · Private
- 96Five Bad DaysEasy
- 97Most Months Are Below AverageMedium · Private
- 98Forty-Five Percent, Still Not ProvenMedium · Private
- 99Two Extra KnobsMedium · Private
- 100Is the Router Playing Favorites?Medium · Private
- 101The Next Two RFQsMedium · Private
- 102How Fast Does the Edge Rot?Medium · Private
- 103One Print to Call the RegimeHard · Private
- 104Twin Tripwires at 19 to 1Hard · Private
- 105The 1.96 That Wasn'tHard · Private
- 106Selected Into AnticorrelationHard · Private
- 107Don't Control for the MessengerHard · Private
- 108Loud Noise on the Big DaysHard · Private
- 109Twenty Days Out, Two Different WorldsHard · Private
- 110Whose Correlation Is It?Hard · Private
- 111How Hard to Lean on the Peer GroupHard · Private
- 112The Optimizer Grades Its Own HomeworkHard · Private
- 113Aim Where the Tail LivesHard · Private
- 114Subtract What You Already KnowHard · Private
- 115The Information in a Day's RangeHard · Private
- 116The Second-Worst Day Is Not Your VaRHard · Private
- 117When the Prior Stops MatteringHard · Private
- 118Diversification Raised My VaRHard · Private
- 119The Size That Maximizes the TakeHard · Private
- 120Skew Arrives With the RegimeHard · Private
- 121Patching the Error Bar, Lag by LagHard · Private
- 122Three New Signals, Worse Than NoiseHard · Private
- 123Vol That Ignores the Fat FingerHard · Private
- 124Proving the Hedge DecayedHard · Private
- 125The Correction That Flips the VerdictHard · Private
- 126The Overlap IllusionHard · Private
- 127Does the Second Bet Help?Medium · Private
- 128Add a Third EngineHard · Private
- 129The Best of Twenty BacktestsHard · Private