Fin Maverick
Foundations VocabularyAccounting & ReportingEconomics & MacroQuant Methods & ProgrammingBusiness & Company AnalysisCorporate Finance & ValuationBehavioural Finance
Banking & Market InfrastructureFixed Income & RatesDerivatives & Structured ProductsPublic EquitiesTransactions & DealsPortfolio ConstructionFunds & AMCs
Private Markets & AlternativesRisk, Treasury & ControlAI & Digital FinanceStochastic Calculus & PricingWealth & Personal FinanceIndian Markets & RegulationProfessional Practice
CalculatorComparison
Frameworks
Explore Bootcamps
Equity ResearchPortfolio ManagementMutual Fund MasteryFinancial LiteracyInvestment Banking Analyst
Private Equity AnalystHedge Funds AnalystBreaking Into VCBreaking Into QuantsAI For Finance
Financial Analyst ProgramRisk Management ProgramPrivate Wealth ManagementDebt Capital MarketsDerivatives Foundation
Explore Internships
Equity Research InternMutual Fund Intern
Portfolio Management InternFinancial Literacy Intern
Explore Micro Courses

Equity Research6

Writing an Investment ThesisBuilding a Discounted Cash FlowReading an Annual Report FastReading a Sector Before a CompanySpotting Quality of Earnings Red FlagsBuilding a Revenue Forecast From Drivers

Portfolio Management3

Rebalancing: When, Why and What It CostsStrategic and Tactical Asset AllocationMeasuring Risk in a Portfolio

Mutual Fund Mastery3

Comparing Funds Without Being FooledHow a NAV Is Struck and Which Day You GetReading a Fund Factsheet Properly

Derivatives Unlocked4

Hedging a Real ExposureThe Greeks, PracticallyFutures, the Basis and What Moves ItReading an Option Payoff

AI For Finance2

Retrieval and Grounding for FinanceDocument Extraction in Finance

Breaking Into Quants4

Backtesting a StrategyHypothesis TestingCleaning Financial DataRegression for Finance

Breaking Into VC3

Sizing a MarketReading a Term Sheet as a FounderHow a Venture Round Actually Works

Financial Analyst Program4

Common Size and Trend AnalysisReading a Cash Flow StatementRatio Analysis That Says SomethingBuilding a Working Capital Schedule

Risk Management Program2

Credit Exposure and How It Is ReducedValue at Risk and What It Hides

Investment Banking Analyst3

Precedent Transactions and Why They DifferReading a Term Sheet StructurallyBuilding a Comparable Companies Table

Private Wealth Management3

Tax Aware Portfolio DecisionsBuilding a Client Risk ProfileGoal Based Planning Arithmetic

Debt Capital Markets3

Analysing an Issuer's CreditDuration and What It Does Not Tell YouBond Pricing and Yield Mechanics

Private Equity Analyst2

Fund Waterfalls and CarryThe LBO in Structure

Hedge Funds Analyst2

Short Selling MechanicsLong Short Mechanics
Courses
Explore Career Roadmaps
Investment Banking AnalystEquity Research AnalystVC AnalystPrivate Equity AnalystHedge Funds Analyst
Quant AnalystAI For FinanceFinancial Analyst ProgramPrivate Wealth ManagementDebt Capital Markets
Risk Management ProgramDerivatives FoundationPortfolio ManagementMutual Fund Mastery
PartnershipsShowdown
Log inSign up
Quant Analyst · CoreTrack
1Quantitative Methods, Financial Data & Programming
iProbability
Probability in FinanceRandom VariableProbability DistributionsThe Normal DistributionNormal Distribution ProbabilityThe Lognormal DistributionRandomness vs Uncertainty
iiStatistics and Inference
Population and SampleMean, Median and ModePrecision and AccuracyVariable TypesVariance, Standard Deviation and…Dispersion MeasuresStatistical BiasEffect SizeHypothesis TestingThe Sampling DistributionSkewnessKurtosisCovarianceConfidence IntervalArithmetic Mean vs Geometric MeanStatistical Significance vs Economic…Confidence Interval vs Prediction IntervalHow to Summarise a…
iiiCorrelation and Regression
RegressionCorrelation and CausationOrdinary Least SquaresInteraction TermsRegression CoefficientsRegression vs ClassificationHow to Build a…Spurious CorrelationRegression, Correlation and FitResidualsMulticollinearityAutocorrelation and Partial Autocorrelation
ivTime Series
Time Series in FinanceSimple, Weighted and Exponential…Moving Average CalculatorPrice, Return and Level SeriesHow to Prepare Time-Series…LagFrequencySeasonalityTimestampsTrendStationarity and the Unit RootHeteroskedasticityLeadRolling WindowsDifferencing
vSimulation and Numerical Methods
SimulationMonte Carlo SimulationHow to Run a…Numerical MethodsIterationResampling and the BootstrapPseudorandom Numbers and the SeedConvergence and ToleranceNumerical Stability
viOptimisation
OptimisationLocal and Global OptimaConstraintsConvex OptimisationThe SolverLinear ProgrammingThe Objective FunctionConstraint ViolationThe Feasible SetLagrange MultipliersQuadratic Programming
viiModelling Practice
Linear, Logistic, Ridge and…Training, Validation and Test…The ModelModel ErrorDependent and Independent VariablesThe ROC Curve and AUCWhat a Model HoldsMSE, RMSE, MAE and MAPEPrecision and RecallCross Validation and RegularisationOverfitting and UnderfittingReturn Series MeasuresSimple, Compound and Log Return
viiiBacktesting and Research Integrity
BacktestingBacktest vs Live PerformanceHow to Document a…How to Prevent Backtest…Out-of-Sample TestingWalk-Forward AnalysisMultiple TestingP-HackingData Snooping
ixData Quality and Structure
Data QualityThe DatasetSelection and Survivorship BiasVersioned DatasetsData Structures in FinanceData CleaningMissing Data and Null ValuesStructured Data vs Unstructured DataMissing Data vs ZeroData Validation vs Data CleaningOutliersDuplicate Records
xProgramming for Finance
Data PipelinesAPIs for Financial DataAPI vs CSV FileDatabases in FinancePython for FinanceJoinsSQL for FinanceThe Analysis Workflow
xiQuantitative Research
Research DesignThe Data Generating ProcessReproducibilityPeer Review in Analytical WorkThe Research HypothesisRobustness and Sensitivity
2Stochastic Calculus & Derivative Pricing Theory
iProbability Foundations
The Probability SpaceRandom VectorsSigma-AlgebraExpectationSample Space and EventsDensity and Distribution FunctionsRisk-Neutral ProbabilityState Price Density vs…
iiStochastic Processes and Jumps
Properties of a Stochastic ProcessMartingaleBrownian Motion and Its PropertiesBrownian Motion vs Geometric…Stopping TimeThe Markov PropertyState VariablesTransition ProbabilityQuadratic VariationQuadratic Variation vs Ordinary…Submartingale and SupermartingaleMartingale RepresentationMarkov Process vs MartingaleOptional StoppingFiltrationJump ProcessesThe Poisson ProcessLevy ProcessesJump Diffusion
iiiIto Calculus
The Ito IntegralThe Ito Integral vs the Riemann IntegralInfinitesimals in Stochastic CalculusQuadratic CovariationIto's LemmaHow to Apply Ito's…The Infinitesimal GeneratorIto Calculus vs Ordinary Calculus
ivStochastic Differential Equations
Stochastic Differential EquationsStochastic Differential Equation vs…Drift and DiffusionStrong and Weak Solutions ComparedDiscretisationGeometric Brownian Motion
vPricing Theory and No-Arbitrage
No-ArbitrageGirsanov, Radon-Nikodym and Change…Physical and Risk-Neutral Measures…The Fundamental Theorems of…The Law of One PriceThe Pricing KernelDiscount Factors and Zero-Coupon PricesReplication vs HedgingComplete Market vs Incomplete MarketClearing Margin Architecture
viOption Pricing Theory
European and American OptionsMonte Carlo European OptionThe Black-Scholes PDEBlack Scholes and the GreeksThe Payoff FunctionThe Binomial ModelBinomial Option PricingDelta Hedging in TheoryBoundary, Initial and Terminal ConditionsThe Exercise BoundaryHow to Check Put-Call…
viiVolatility Models
Constant, Local and Stochastic…Vasicek Model vs CIR ModelThe Heston ModelThe SABR ModelThe Volatility ProcessImplied VolatilityVolatility Smile vs Skew vs Surface
viiiInterest Rate Models
Interest-Rate DerivativesMean ReversionThe Zero-Coupon BondThe Ornstein-Uhlenbeck ProcessThe Discount CurveZero RatesShort-Rate Model vs Market Model
ixNumerical Pricing
Closed Form and Numerical…Monte Carlo PricingEuler and Milstein Schemes ComparedTree MethodsFinite Difference MethodsNumerical Error and StabilityVariance Reduction
xCalibration and Model Risk
Model OverrideMarket Price and Model PriceCalibrationHow to Document a Pricing ModelThe Educational Illustration LabelMarket ConventionsModel Uncertainty and LimitationsBacktesting a Pricing ModelIdentifiabilityCalibrated ParametersThe Calibration Loss Function

Iteration: Approaching an Answer Step by Step

Iteration means repeating one rule, with each pass starting from where the last pass finished. Two kinds matter. A fixed count repeats a stated number of times: the Nakshatra ladder runs twelve months and stops. A search repeats until a condition is met: bisection halves a bracket around the rate that reaches Rs 120/-, and ten steps shrink that bracket to one thousand and twenty fourth of its starting width.

Every rate, every reading and every bracket width below is the output of one arithmetic rule, and the rule is short enough to print in a line: begin at Rs 100/-, multiply by one plus the monthly rate, take Rs 2/- off, and do that twelve times. A calculator and some patience rebuild all twenty two steps and land on the same digits. Nothing in the sequence has to be taken on trust.

The Nakshatra ladder has already appeared. The ladder, an invented arithmetic rule, runs like this: start at Rs 100/-, multiply by one plus a monthly rate, subtract a standing chargeA fixed amount taken off every month whatever else happened, which here is the Rs 2/- the rule subtracts before the next month begins. of Rs 2/-, and repeat that twelve times. Two things about the ladder are already settled. At a rate of exactly 2.00 per cent a month the ladder sits flat at Rs 100.0000/- for all twelve months, and that case has a closed formAn answer that can be written down as a finished expression and worked out in a single pass, with no repeating involved. Here it is one division.: the charge over the starting figure, Rs 2/- over Rs 100/-, and no arithmetic beyond that division. At a target of Rs 120/- there is no such expression at all. The rule cannot be rearranged to hand over the rate, and no amount of algebra changes that.

Which leaves the only option that remains: try a rate, look at what the ladder reads, and try again somewhere better. Trying has a shape. A rule repeated so that each pass builds on the last can be organised to close in rather than wander, and it can be told when to stop. None of that needs anything but arithmetic and a willingness to do it several times over.

Before any of that, the everyday version. Ten shops in one mall have rents that go up by ten per cent every year. A shop paying Rs 20,000/- a month this year pays Rs 22,000/- next year, then Rs 24,200/-, then Rs 26,620/-. After eleven of those raises it pays Rs 57,062.33/-, and that figure does not depend on the first year except through the ten years in between. Each year's answer is the next year's starting point. Every shopkeeper in that mall runs an iteration in their head without ever needing a word for it.

What is iteration, and how is it different from doing something twice?

Iteration is repeating one rule where each pass starts from where the last pass finished. The second half of that sentence is the whole of it. Doing the same thing several times over is repetition, and repetition hands back several answers to the same question. The passes of an iteration are joined end to end rather than run side by side, so iteration hands back one answer that has moved. The joining is exactly why an iteration can travel somewhere. A repetition can only stand where it started and say the same thing again.

A rent rise put through both arrangements makes the difference concrete. A ten per cent rise applied to Rs 20,000/- four separate times gives Rs 22,000/- four times over. Four copies of one fact. Applied four times in a chain, each rise landing on whatever the last one produced, it gives Rs 22,000/-, then Rs 24,200/-, then Rs 26,620/-, then Rs 29,282/-. Same rule, same number of applications, and the second arrangement has reached somewhere the first cannot get to however many times it is run.

The joining charges something as well as buying something, and the charge is worth naming now because these notes on method keep walking into it. Because every pass carries the last one forward, whatever went wrong on pass three is still sitting inside pass four. A repetition that goes wrong once is wrong once and no more. An iteration that goes wrong once may carry that wrongness all the way to the end, and whether the damage fades or grows as it travels is a separate question with its own notes.

THE SAME RULE, RUN TWO WAYS, FOUR TIMES EACH The rule is a rise of ten per cent. Only the starting point of each pass differs between the two rows. REPETITION: every pass starts from the same place PASS ONE PASS TWO PASS THREE PASS FOUR Rs 20,000/- Rs 20,000/- Rs 20,000/- Rs 20,000/- Rs 22,000/- Rs 22,000/- Rs 22,000/- Rs 22,000/- Four answers to one question, and no pass knows that any other pass happened. ITERATION: every pass starts where the last one finished START AFTER ONE AFTER TWO AFTER THREE AFTER FOUR Rs 20,000/- Rs 22,000/- Rs 24,200/- Rs 26,620/- Rs 29,282/- One answer that has travelled Rs 9,282/- from where it began, which four separate passes could never do. The rule is identical in both rows. Only the starting point of each pass changed, and that is the entire difference.
Four independent passes of a ten per cent rise return Rs 22,000/- four times, while four joined passes reach Rs 29,282/-, so the connection between passes and not the rule is what lets an iteration travel.
Try it out

A clerk applies the same ten per cent rise to a rent four times and writes down four figures. What would make that an iteration rather than a repetition?

Breaking Into Quants Bootcamp — Fin Maverick

What are the two kinds, and which single question tells them apart?

One question separates them, and it can be asked before writing a single line of arithmetic. Is the number of passes known in advance?

If the answer is yes, the arrangement is a fixed count. Twelve months is the question being asked, so the Nakshatra ladder run over twelve months is a fixed count. Twelve passes is what the rule gets, and it would still be twelve if the readings came out at Rs 40/- or at Rs 400/-. The count was never in the arithmetic's hands. Nothing the readings produce along the way can lengthen it or cut it short.

If the answer is no, the arrangement is a search. Bisection on the Rs 120/- target is a search. The run repeats until the two ends it is working between are close enough together, and how many passes that takes depends entirely on how close was demanded. Ten steps satisfies one demand and twenty two satisfies another, and neither number was known before the run began.

A fixed count states its count before it starts, so a fixed count always finishes. A search finishes only if the condition it is waiting for can be reached at all. The gap between the two is not a small difference in temperament, but the difference between a run that has an end and a run that has a hope of one. Every serious failure described below grows out of it, and so does the reason a search always needs a cap on how many passes it may take before it gives up and says so.

ONE QUESTION SEPARATES THEM, AND THREE CONSEQUENCES FOLLOW Ask it before writing any arithmetic: is the number of passes known in advance? THE TEST A FIXED COUNT A SEARCH Passes known before it starts? Yes. The count is the question being asked. No. It depends on what the readings turn out to be. What makes it stop? The count runs out. A stated condition is met. Can it fail to finish? Never. Yes, if the condition it waits for is out of reach. The case here The ladder over twelve months, at any rate at all. Bisection hunting the rate that reaches Rs 120/-. A fixed count always finishes because its count was stated. A search finishes only if its condition is reachable.
A fixed count always finishes because the count was stated before it began, while a search finishes only if the condition it waits for can be reached at all.
Try it out

A question asks what a price becomes after eighteen months of a stated monthly change. Which kind of repeating is that, and how can it be told before doing any arithmetic at all?

How does bisection actually work, step by step?

Bisection needs one thing before it starts: a pair of rates the answer is confidently taken to sit between. Given that, the whole method is three lines, and they repeat until something stops them.

The whole of bisection. Nothing has been left out of these three lines.
LineWhat happens
OneTake the value exactly halfway between the two ends.
TwoRun the rule there and read off what it gives.
ThreeIf that reading is below the target, the answer lies in the upper half; if it is above the target, the answer lies in the lower half. Throw the other half away and go back to line one.

The only operation anywhere in those three lines is a comparison: is this reading below the target or above it. No slope, no rate of change, no clever guess at where the answer might be hiding, and no assumption about the shape of the rule beyond one. The plainness is not a limitation dressed up as a virtue. Nothing in bisection is sophisticated enough to be wrong about, and a method with nothing to be wrong about cannot fail in the ways cleverer methods fail.

The one condition line three needs is worth saying out loud. The rule's reading has to move in a single direction as the rate rises, and this one does: the ladder reads Rs 87.3175/- when the monthly rate is 1.00 per cent, then Rs 100.0000/- at 2.00, Rs 114.1920/- at 3.00, Rs 130.0516/- at 4.00, and Rs 167.4798/- once the rate reaches 6.00. A rule that behaves that way is monotoneMoving in one direction throughout and never turning back, so a higher input always gives a higher reading., and on a monotone rule a reading above the target really does mean that every rate above that middle value reads above the target too. Throwing half away is therefore safe rather than merely convenient.

One more property is worth noticing while the method is this simple. Bisection is deterministicProducing the same output every time from the same starting point, with nothing random anywhere in it.. Give it the same two ends and the same rule and it produces the same middle values in the same order, on any machine, on any afternoon. No drawn number enters bisection anywhere, so there is no starting value for a stream of drawn numbers to name. Seeding such a stream is covered under computer randomness, and bisection does not need it.

ONE STEP, THREE OPERATIONS, AND NOTHING ELSE IN IT Drawn on the tenth step of the search below, where the two ends are 3.375000 and 3.382812 per cent. ONE. TAKE THE MIDDLE Halfway between 3.375000 and 3.382812 per cent. 3.378906 per cent a month TWO. READ THE RULE Run the ladder twelve months at that rate. Rs 119.9961/- against a target of Rs 120/- THREE. DISCARD A HALF The reading is below the target, so the answer is higher than this middle. the lower half goes and the upper half survives A comparison is the only judgement in the step. No slope, no rate of change and no guess enters anywhere. Then go back to line one with the half that survived, and do the same three things again.
A bisection step is a middle value, a reading and a comparison, and no slope, rate of change or guess about the answer enters the method anywhere.
Try it out

A middle value of 4.000000 per cent gives a reading of Rs 130.0516/- against a target of Rs 120/-. Which half survives, and which would have survived had the reading been Rs 110/-?

AI For Finance Bootcamp — Fin Maverick

What does the whole search look like on the Rs 120/- target?

Start with a bracket of 2.00 to 6.00 per cent a month. The pair is a legitimate starting point for the Rs 120/- target for exactly one reason: the ladder reads Rs 100.0000/- at the lower end and Rs 167.4798/- at the upper, so the target sits between the two readings. A pair of ends earns the right to be called a bracket by having the target fall between what they read, never by looking wide enough to be safe. That distinction costs one minute at the start and it is the check most searches are missing.

Here are the first ten steps in full. Each row shows the two ends going in, the middle value tested, what the ladder reads there, and which half survived the comparison.

Bisection on the Nakshatra ladder, target Rs 120/-, starting bracket 2.00 to 6.00 per cent. Every reading recomputed from the rule.
StepThe two ends going inMiddle testedThe ladder readsWhich half survived
12.000000 to 6.0000004.000000Rs 130.0516/-above, so the lower
22.000000 to 4.0000003.000000Rs 114.1920/-below, so the upper
33.000000 to 4.0000003.500000Rs 121.9029/-above, so the lower
43.000000 to 3.5000003.250000Rs 117.9941/-below, so the upper
53.250000 to 3.5000003.375000Rs 119.9350/-below, so the upper
63.375000 to 3.5000003.437500Rs 120.9156/-above, so the lower
73.375000 to 3.4375003.406250Rs 120.4245/-above, so the lower
83.375000 to 3.4062503.390625Rs 120.1795/-above, so the lower
93.375000 to 3.3906253.382812Rs 120.0572/-above, so the lower
103.375000 to 3.3828123.378906Rs 119.9961/-below, so the upper

Every step halves the width, and it does so without consulting the readings at all. Four percentage pointsThe plain gap between two per cent figures, so 3.38 per cent less 3.37 per cent is one hundredth of a percentage point. becomes two, then one, then half of one, and so on down. After those ten steps the two ends run from 3.378906 to 3.382812 per cent. The width is 0.00390625 of a percentage point, one thousand and twenty fourth of the width it began with. Ten comparisons bought a thousandfold narrowing, and not one of them required knowing anything about the rule except whether a reading came in high or low.

Keep going and the narrowing keeps its pace. By the twenty second step the middle being tested is 3.379155 per cent, where the ladder reads Rs 119.999995/-, and the two ends run from 3.379155159 to 3.379156113 per cent. The width is then roughly one part in 41,94,304 of where the search started. Twenty two doublings of precision look like that written out.

A search that hands back a number has proved nothing at all. The returned answer goes back through the original rule first. Carried to nine decimal places of a per cent the answer is 3.379155475, and the ladder run at that rate reads Rs 120.000000/- to six decimal places and past them. Carried to six decimal places instead, 3.379155, the ladder reads Rs 119.999993/-, short of the target by about seven millionths of a rupee. Both figures are perfectly fine to print. Only one of them is fine to put back into the rule and call an exact hit, and knowing which is which is the whole of what the precision of a returned answer means.

There is a second check and it costs even less. Point the same procedure at the case whose answer is already known. Bisection on the Rs 100/- target, from a starting bracket of 1.00 to 4.00 per cent, returns 2.000000 per cent. The closed form for the flat case says the rate is the charge over the starting figure, Rs 2/- over Rs 100/-, or 2.00 per cent exactly. The two agree to fifteen decimal places. A method that gets the known case right has earned the right to be pointed at the case nobody can check, and a method that gets it wrong has said so before costing anything.

TEN STEPS, EACH DRAWN AT TWICE THE MAGNIFICATION OF THE ONE ABOVE IT Green is the half that survives. Grey is the half thrown away. The faint lines carry the survivor down and stretch it to full width. STEP 14.0000002.0000006.000000Rs 130.0516/-above, keep the lower half STEP 23.0000002.0000004.000000Rs 114.1920/-below, keep the upper half STEP 33.5000003.0000004.000000Rs 121.9029/-above, keep the lower half STEP 43.2500003.0000003.500000Rs 117.9941/-below, keep the upper half STEP 53.3750003.2500003.500000Rs 119.9350/-below, keep the upper half STEP 63.4375003.3750003.500000Rs 120.9156/-above, keep the lower half STEP 73.4062503.3750003.437500Rs 120.4245/-above, keep the lower half STEP 83.3906253.3750003.406250Rs 120.1795/-above, keep the lower half STEP 93.3828123.3750003.390625Rs 120.0572/-above, keep the lower half STEP 103.3789063.3750003.382812Rs 119.9961/-below, keep the upper half kept, because the answer has to be in here let go, because the answer cannot be in there Ten halvings take four percentage points down to 0.00390625 of one, which is one thousand and twenty fourth of the start.
Each step throws away the half that cannot hold the answer, so a starting width of four percentage points becomes one thousand and twenty fourth of itself after ten steps.
Try it out

After ten steps the two ends run from 3.378906 to 3.382812 per cent. To how many decimal places of a per cent may the answer honestly be reported now, and to how many may it not?

Try it out

Ten bisection steps are about to be taken from a bracket of 2.00 to 6.00 per cent. How wide will the two ends be once they are done?

Play with it

Walk the search one step at a time and watch the two ends close around a fixed line.

One control moves: how many bisection steps have been taken, from the first to the twenty second. So that the two ends can be seen vanishing into it, the upper panel keeps the whole starting range in view. The middle panel magnifies whatever they currently are, so they never disappear, and the lower panel draws the ladder's reading against the target of Rs 120/-. The opening setting is ten steps taken. The step just completed tested 3.378906 per cent, the ladder read Rs 119.9961/-, and the two ends now run from 3.378906 to 3.382812 per cent, the same figures the table above prints as plain text.

Jump straight to a step, or walk one at a time:
The middle tested at this step
3.378906 per cent
What the ladder reads there
Rs 119.996102/-
Against the target
Rs 0.003898/- below
The two ends now
3.378906 to 3.382812 per cent
Their width
0.00390625 points
As a share of the start
one part in 1,024
Loading the panel.

Educational illustration. The starting bracket of 2.00 to 6.00 per cent is legitimate for one reason: the ladder reads below the target at one end and above it at the other, and nothing else qualifies two numbers to be called a bracket. A rate that changed month by month would be a different question, so the rate is held constant across all twelve months. The true answer drawn as a fixed line is 3.379155475 per cent, computed by carrying the same halving far past the twenty second step.

Try it out

A simulated estimate and a pair of bisection ends both indicate roughly where an answer is. Which of the two makes the stronger claim, and which word marks the difference?

Risk Management Program Bootcamp — Fin Maverick

What can a bracket promise that an estimate cannot?

At every one of those twenty two steps the run could have stopped and said something exact. The answer is between these two numbers. Not probably, not within a margin, not nineteen times out of twenty. Between these two numbers, and that is the end of the sentence.

One other way of getting a number out of a rule that cannot be solved stands beside bisection: draw a great many random cases, work each one through and report what came back. An estimate built that way arrives with a standard errorThe typical distance between an estimate built from a limited number of draws and the quantity it is estimating. Earlier notes build it in full. attached, and what a standard error buys is a statement of the form the answer is probably within so much of this. Probably. A pair of bisection ends is a different kind of sentence altogether: it says the answer is definitely inside this range, and on most problems that difference is worth a great deal more than speed.

Watch what the answer never does across the whole run. The answer never leaves. After step one it is inside 2.000000 to 4.000000 per cent. After step ten it is inside 3.378906 to 3.382812. After step twenty two it is inside 3.379155159 to 3.379156113. Every one of those statements was exact when it was made, none of them was ever revised, and each one is simply a tighter version of the one before rather than a correction of it.

The reason it never leaves is the comparison at line three, and it is worth seeing why rather than taking it on trust. On a monotone rule every rate in the discarded half reads on the wrong side of the target, so the half being thrown away is the half that cannot contain the answer. Nothing is being estimated when that half goes. The half is being ruled out. A method that only ever discards what it has already checked keeps the answer between its two ends from the first step to the last, and that, rather than any question of speed, is the whole of what bisection promises.

Choosing a tolerance depends on what the answer is for. How narrow is narrow enough is settled under convergence and tolerance, and so is the comparison between the speed of this narrowing and the speed of a simulated estimate. The promise itself has one shape: certainty about a range, bought with comparisons, one factor of two at a time.

THE ANSWER NEVER LEAVES, AND THE BAND CLOSES AROUND IT The dashed line is the answer, 3.379155475 per cent. The band is the pair of ends after each step. Distance from the answer, in percentage points, on a scale where each labelled step is a hundredfold. 1 1 0.01 0.01 0.0001 0.0001 0.000001 0.000001 0 5 10 15 20 22 steps taken the answer, 3.379155475 per cent, and it never leaves the upper end, always above the answer the lower end, always below the answer
The answer sits inside the pair of ends at every one of the twenty two steps, so each step states with certainty rather than with confidence that it lies between these two numbers.
Reading an Option Payoff — free micro-course from Fin Maverick

When should the repeating stop, and which rule is wrong?

By definition nobody told a search how many passes to run, so a search has to be told when to stop. Three stopping rules are in common use. Two of them are defensible and one is wrong.

The first is to stop when the two ends are close enough together. The gap between the ends is a real statement about where the answer is rather than a hint about it, so on bisection stopping on the width is the honest rule. When it is small, what is known is correspondingly narrow.

The second is to stop when the reading is close enough to the target. The gap between the current reading and the target has a name, the residualThe distance between where a reading currently stands and the target it is chasing, measured in whatever units the reading uses., and this rule watches that instead of watching the rate. The residual is the right thing to watch whenever what matters is the consequence of the answer rather than the answer itself. On the ladder that means the rupees rather than the rate.

The third is to stop when the step gets small, and that is the wrong one, for a reason that does not show up on this problem at all: a small step means the method is moving slowly, not that it has arrived. Those two sound like the same thing and are not, and telling them apart is the difference between a search that can be trusted on a new problem and one that cannot.

On the ladder the two happen to coincide, and the coincidence is precisely what makes the fault hard to see. Nudge the rate by a millionth of a whole unit and the ladder's reading at the end of its twelve months shifts by Rs 0.001564/-, so by the time the steps are that small the readings really have gone still, and a search stopping on step size lands somewhere sensible. The agreement is a fact about how gently this particular rule responds, not a fact about the stopping rule. Draw a rule ten times as steep through the same point and the same step of one in a million moves the reading by Rs 0.015642/-. A hundred times as steep and it is Rs 0.156421/-. More than fifteen paise of error hides behind a step that looked reassuringly tiny. The method halves whatever problem it is pointed at, so the step size is a fact about the method. The residual is a fact about the problem. Only the residual needed to be made small.

THE SAME SMALL STEP, ON TWO SLOPES, AT ONE SHARED SCALE Both panels run from Rs 119.982/- at the bottom to Rs 120.018/- at the top. The step is one in a million in the rate. THE LADDER, WHICH IS GENTLE HERE A RULE TEN TIMES AS STEEP the target, Rs 120/- the target, Rs 120/- one step of one in a million in the rate the identical step, side by side Rs 0.001564/- Rs 0.015642/- The step is identical in both panels. What it costs in rupees is not, and the step size cannot say which panel applies. The steeper line leaves the top and bottom of its panel because this shared scale cannot hold it.
One step of the same size moves the ladder's reading by Rs 0.001564/- and a rule ten times as steep by Rs 0.015642/-, so the size of a step says nothing about how far the reading still sits from the target.
Try it out

A search stops when its step falls below one in a million and reports its answer. On the Nakshatra ladder that is fine. What would have to be true of a different rule for the same code to fail, and how would the failure show up?

Two stopping rules are defensible and one is wrong. See where the repeating stops.

How is it known that the repeating will finish at all?

For a fixed count the question barely exists. A fixed count finishes for one reason only: the count was stated. The only way it fails to finish is if somebody forgot to state one.

For bisection the answer is almost as firm, and it comes from the halving rather than from the rule being searched. The width after any number of steps is four percentage points divided by two raised to that number of steps, and that is a complete description with no readings in it. Any named width is reached in a countable number of steps: one in a million of a whole unit of the rate takes sixteen steps and one in a hundred million takes twenty two. The halving never consults a reading before halving, so no shape of rule and no unlucky run can stretch those counts.

One case still remains: the condition itself is out of reach, and here the cap earns its keep. The only alternative to stopping is running forever, so a search written without a stated maximum number of passes is a defect rather than a matter of style. Ask for a residual below one ten thousandth of a paisa on a rule whose arithmetic cannot carry that many digits and the condition will never be met, however patient the machine is. With a cap, the run ends and says it did not converge, and a failure announced is information. Without a cap the run simply never comes back, and silence is not.

The commonest reason a condition turns out to be unreachable is a bracket that never contained the answer in the first place. Hand the same procedure a starting pair of 4.00 to 6.00 per cent for the Rs 120/- target. The ladder reads Rs 130.0516/- at one end and Rs 167.4798/- at the other, and the target sits below both of them. Every middle it tests reads above the target, so the upper half is discarded every single time and the pair marches steadily down onto its own lower end: 5.000000 per cent reads Rs 147.7514/-, then 4.500000 reads Rs 138.6601/-, then 4.250000 reads Rs 134.2970/-, and it goes on shrinking towards 4.000000 per cent forever. A search stopping on the residual will never get there and the cap will fire, and the firing is the warning working. A search stopping on the width will stop cleanly and report 4.000000 per cent, whose reading is Rs 130.0516/-, a full Rs 10.0516/- from what was asked for, with nothing in the run to say so. Only putting the answer back through the rule catches that one.

THE WIDTH FALLS IN A STRAIGHT LINE, AND STRAIGHT IS THE WHOLE POINT Width of the pair of ends, in percentage points, on a scale where each labelled line is a tenth of the one above it. 1 0.1 0.01 0.001 0.0001 0.00001 0.000001 0 4 8 10 12 16 20 22 steps taken 10 steps: 0.00390625 16 steps: 0.0000610 22 steps: 0.0000010 A straight fall on this scale means every step buys the same factor, and here that factor is exactly two.
The width falls by exactly one factor of two per step whatever the readings do, so sixteen steps reach a millionth of a whole unit of the rate and twenty two reach a hundred millionth.
Try it out

A bisection is handed a pair of ends whose readings are both above the target. What will it do, and which check would have caught it before a single step ran?

The failure: a stopping rule that was correct on the only problem it was tested on

An analyst writes a search that stops when the step gets small. Stopping on the step reads as the natural way to say the method has settled down, it is one line of code, and on the Nakshatra ladder it works. The ladder's reading shifts by only Rs 0.001564/- when the rate moves by a millionth of a whole unit, so here a small step really does mean an arrived answer. The code is written, it is tested against the Rs 100/- case and the Rs 120/- case, it passes both, and it goes into use.

Then the same code is pointed at a rule that responds more sharply. The shrinking never depended on the rule, so the steps shrink at exactly the same rate. The readings are still a long way from the target. The only thing the search was watching has gone small, so it stops anyway and reports an answer with no warning attached to it. The fault is invisible until the problem changes, and that is precisely what makes it expensive: the code was tested, it passed, and it was genuinely correct for the case it was tested on. The cost lands as an answer that is confidently wrong on the one occasion when nobody thinks to recheck it, because it has been right every time before.

The fix is two lines and neither is clever. Stop on the width of the pair of ends or stop on the residual, and never on the size of the step. The step belongs to the method; the other two belong to the problem. And cap the number of passes, so a search that cannot meet its condition ends by saying so out loud rather than by quietly handing back whatever it happened to be holding.

What does an analyst check before trusting a returned answer?

Iteration is where a great deal of everyday quantitative work actually happens, and almost none of it looks dramatic. Someone in a lending team backs out the rate implied by a schedule of payments. Someone in a valuation team backs out the growth figure that makes a set of assumptions land on a stated total. A household working out what monthly saving reaches a school fee in six years is doing the identical thing on a smaller sheet: guess, run it forward, look at the gap, guess again. In every one of those, a number comes back out of a repeating procedure, and the person receiving it has to decide whether to build anything on it.

Three checks decide that, and together they take about a minute.

Check one, and it costs a single line: put the returned answer back into the original rule and see whether it produces the target. Take 3.379155475 per cent, run twelve months of the ladder, and read Rs 120.000000/-. Almost every fault a search can suffer ends with a number that does not satisfy the thing it was supposed to satisfy, so this one check catches almost everything a search can get wrong. A search that stopped too early fails it. A search that ran on a bracket that never held the answer fails it. A search whose stopping rule was watching the wrong quantity fails it. A returned number that has not been put back through the rule is a claim, not a result.

Check two is about where the search began rather than where it ended. Confirm that the two starting ends really did sit on opposite sides of the target, by reading the rule at both of them. At 2.00 per cent the ladder gives Rs 100.0000/-, at 6.00 per cent it gives Rs 167.4798/-, and the target of Rs 120/- sits between the two. The straddle check would have caught the 4.00 to 6.00 pair before a single step ran, and it costs two runs of the rule.

Check three is about the method rather than the run. The same procedure is pointed at a case whose answer is already known, to see whether it comes back with the right thing. Bisection on the Rs 100/- target returns 2.000000 per cent, and the closed form says exactly 2.00 per cent. The known case is a small ceremony, and it has caught reversed comparisons, wrong targets and off by one loops in more places than anybody enjoys admitting.

WHAT COMES BACK, AND THE THREE THINGS DONE TO IT BEFORE IT IS BELIEVED The slip on the left is everything the search reports. On its own it is a claim rather than a result. WHAT THE SEARCH RETURNED the rate it settled on 3.379155475 per cent steps taken 22 the pair of ends it stopped on 3.379155159 to 3.379156113 per cent the stopping rule it used width of the pair of ends Nothing on this slip proves the number is right. CHECK ONE, ONE LINE Put it back into the rule. reads Rs 120.000000/- passes, and this one catches almost everything CHECK TWO, TWO RUNS Did the starting ends straddle it? Rs 100.0000/- and Rs 167.4798/- passes, the target lies between them CHECK THREE, ONE RERUN Same method on the known case. 2.000000 against 2.00 exactly passes, so the method itself is sound Check one alone would have caught an early stop, a bad starting pair and a reversed comparison.
Putting the returned answer back into the original rule costs one line and catches almost every fault a search can produce, including the two the other checks are aimed at.
Try it out

Name the single check on a returned answer that costs one line and catches almost everything a search can get wrong.

What sits alongside iteration, and where each of those subjects is settled. What a numerical method is and which kinds of them exist is covered under numerical methods. How close is close enough, and how the narrowing speed of a search compares with the speed of an estimate built from drawn numbers, both belong to the notes on convergence and tolerance. Whether repeated arithmetic makes a small error grow belongs to the notes on numerical stability. A bracket width is a plain fact about what bisection does, and which width to be satisfied with is settled with the tolerance.

Where do these numbers come from, and how would they be checked?

Each number above is the output of a rule stated in full rather than a value read off somewhere else, so every one of them is checkable without going elsewhere. The subject is arithmetic, and arithmetic carries its own proof: what four divided by two raised to the tenth comes to is settled by dividing.

The Nakshatra ladder was written to cooperate. Rs 2/- over Rs 100/- is a tidy division, so its flat case lands on exactly 2.00 per cent, and real problems are rarely that obliging. On an untidy problem the shape of the method survives unchanged: a comparison, a halving, and a pair of ends the answer never leaves.

The table below lists what was printed above and what would have to be done to reproduce each figure.

Every figure in this guide, the rule that produces it, and the arithmetic needed to redo it.
What is printedThe rule that produces itWhat is needed to redo it
Rs 100.0000/-, Rs 114.1920/-, Rs 130.0516/- and Rs 167.4798/-The ladder run twelve times at 2.00, 3.00, 4.00 and 6.00 per cent a monthTwelve multiplications and twelve subtractions for each
The twenty two middle values, from 4.000000 down to 3.379155 per centHalve the pair of ends and keep the half whose readings straddle Rs 120/-An addition and a division by two, twenty two times over
The widths, from four percentage points down to 0.0000010 of oneFour divided by two raised to the number of steps takenOne division, with no reading involved at all
The answer, 3.379155475 per cent a monthThe same halving carried well past the twenty second stepNothing beyond the two rows above it
Rs 0.001564/- of reading for a step of one in a millionThe change in the ladder's reading either side of the answer, worked at full precisionTwo runs of the ladder and one subtraction
Rs 22,000/-, Rs 24,200/-, Rs 26,620/-, Rs 29,282/- and Rs 57,062.33/-Rs 20,000/- raised by ten per cent, each rise landing on the last answerOne multiplication per year

The Nakshatra ladder is invented.
Educational material. Not advice on any investment, tax, budget or market position.

← PreviousNext →
Fin Maverick Micro CoursesExplore Micro Courses
Fin Maverick BootcampsExplore Bootcamps
Fin Maverick

Finance education that ends in a job, not a certificate that gathers dust. Built for young India.

LEARN
CalculatorsFrameworksComparisonsCareersShowdown
RESOURCES
All CoursesMicro CoursesBootcampsInternships
COMPANY
AboutJob openingPartnership
LEGAL
Privacy PolicyTerms & ConditionsContent LicenseReturn & Refund Policy
© 2026 FIN MAVERICK / BUILT FOR INDIA.DO FINANCE, DO NOT JUST READ ABOUT IT.