Skip to content
Forticia
Research
Quantitative FinanceEquities, FX, futures. Factors and backtests, every run replayable.Computational BiologySequence, folding, simulation. Versioned, reproducible labs.Cultural IntelligencePhilosophy, governance, ethics. How institutions decide and answer for it.AI InstrumentationPrivate models. Multi-agent orchestration. Guardrails on write.View all research
Quantitative Finance
  • Equities, FX, futures
  • Factors and backtests
  • Every run logged and replayable
InfrastructurePapersPolarisLink™About
Sign inRequest access
Request access

Research

Quantitative FinanceEquities, FX, futures. Factors and backtests, every run replayable.Computational BiologySequence, folding, simulation. Versioned, reproducible labs.Cultural IntelligencePhilosophy, governance, ethics. How institutions decide and answer for it.AI InstrumentationPrivate models. Multi-agent orchestration. Guardrails on write.

Platform

InfrastructurePapersPolarisLink™About
Request accessSign in
Forticia

A private institute for computational research. It publishes original research and runs a governed environment where every run is logged and replayable.

Sign inSystem status

Research

Quantitative FinanceComputational BiologyCultural IntelligenceAI Instrumentation

Infrastructure

WorkspacesHelper planeGoverned AIPapersStatus

PolarisLink™

SpecificationQuickstartClientsChangelogGitHub repository

Institute

AboutFAQRequest accessContactSign inForticia on GitHub

Forticia publishes research and simulations. Nothing on this site is investment advice.

Official channels: forticia.uk and github.com/Forticia

© 2026 ForticiaPrivacyTerms
  1. Home
  2. Research
  3. Quantitative Finance

Quantitative finance research at Forticia: Alpha is a claim.
We try to break it.

Equities, FX, futures, options and market microstructure. Every run is logged and replayable, every finding is written up with its limits, and every figure is tied to the run that produced it.

Enter the gauntlet Request access
  1. Hypothesis
  2. In-sample study
  3. Blind holdout
  4. Placebo
  5. Friction
  6. Verdict
The validation laboratory

Every idea walks
the same corridor.

Six stages, in a fixed order. The order matters more than any single test: nothing is allowed to look good before it has been allowed to fail.

  1. S1Hypothesis

    A reason before a result.

    Every study starts from a mechanism: a rule, constraint or incentive that forces prices to behave a certain way, and a statement of what would prove us wrong. Patterns without a reason are what any large dataset hands you for free.

    We ask
    What forces this to happen, and what would falsify it?
    Where ideas die
    A pattern found first and explained afterwards.
    Research notebookidle
    md
    Claim. A candidate series leads the target. Mechanism: none stated. Search for the best match.
    In [1]
    best = max(corr(c, target) for c in candidates[:36])
    Out [1]

    Hover a thumbnail to inspect it. The best match so far is outlined.

    Best candidate against the target 0.00
    Best of 0 tries, repeated on noise
    Chance noise alone gives a strong match 0%

    Waiting for the search to run.

    candidates triedstrongest apparent match
    Running maximum
    36
    FewMany
  2. S2In-sample study

    Freedom flatters the past.

    Estimation happens on a fixed window under one pre-stated specification. The more flexible a model is, the better it explains history, which is exactly why explaining history proves nothing.

    We ask
    Is the specification fixed before the data is touched?
    Where ideas die
    Parameters tuned until the past looks good.
    In-sample studyFitted curveStudy observationspolynomial · degree 1
    featureresponsecomplexityin-sample error

    A straight line. Misses the structure, and misses it the same way everywhere.

    degree 1
    RigidFlexible
  3. S3Blind holdout

    Opened once, behind an air gap.

    A partition the research environment cannot reach is held back. It is opened once, after the specification is frozen. If the idea is adjusted after the look, the look is spent and the result no longer counts.

    We ask
    Did the idea survive data it never saw?
    Where ideas die
    Edge that was only ever a description of the study window.
    Blind holdoutStudy windowBlind partitionsealed
    Study windowGapSealed
    timecumulative resultinputs in the specificationrisk-adjusted result
    16 inputs
    SimpleMany inputs
    The research environment cannot read this partition.
  4. S4Placebo

    Does the machinery find edge in nothing?

    The same pipeline runs on versions of the idea that should have no edge: signal inverted, timing shuffled, noise standing in. A real effect stands clear of every one of them. Drift, leakage and bugs do not.

    We ask
    Does the result stand clear of every control?
    Where ideas die
    A market trend mistaken for a signal.
    Placebo controlsThe ideaControl distributions0 permutations
    test statisticInvertedShuffledNoisenormal quantileshuffled control

    Running permutations.

  5. S5Friction

    Edge has to survive being collected.

    Spreads, commissions, impact, financing and execution delay are subtracted, not footnoted. A signal formed at one close is traded at the next open. What remains is the edge that scales to the size being asked of it.

    We ask
    What is left after every cost, at the size we would trade?
    Where ideas die
    A gross edge that costs more to collect than it pays.
    Net edge after frictionNet positiveNet negativecapital × cost level
    edge, selected pointGrossSprFeeImpFinDlyNet

    Most of the edge survives here.

    SmallLarge
    LowHigh
  6. S6Verdict

    Certified, or kept in the graveyard.

    A candidate must clear every gate; one failure ends it. Failures are not deleted. Each is recorded with its mechanism and screened against every future hypothesis before a line of code is written.

    We ask
    Does it clear every gate, and is it independent of what already exists?
    Where ideas die
    Any single gate. There is no averaging across them.
    Certificationevaluating 1 of 10
    CheckEvidence requiredResult
    Data admissionTimestamps as known, next-open execution
    In-sample studyOne specification, fixed in advance
    Blind holdoutSingle look at the partition
    StabilityHolds outside the study window
    Placebo controlsStands clear of every control
    Parameter neighbourhoodNeighbouring settings agree
    Costs and capacityPositive net of cost, at size
    Drawdown budgetWithin the loss budget
    Sample sizeEnough independent events
    IndependenceNot a copy of existing exposure
    EvaluatingEach gate runs in order. One failure ends the run.
    Data
    Pinned snapshot
    Specification
    Frozen before data
    Partition
    Opened once
    Seed
    Recorded
    Failure registry
    • Duplicated an existing exposure
    • Died on costs
    • Only described the study window
Research desks

Five desks.
One contract.

Each desk keeps its universe, its calendar, its cost model and its window on the same run as the figure. The ground differs; the discipline does not.

Cross-section, ranked by signalTop groupOther groupsfive groups
signal group, low to highforward returnperiodscumulative rank correlation

The configuration a live book would have faced.

Corporate actions and borrow travel with the signal. Cross-section and event study in one contract.

What the desk studies

  • Cross-sectional factors and their interaction
  • Event studies around corporate actions
  • Relative value between related securities

How it usually breaks

A universe that quietly drops the names that failed, or a short side nobody could borrow.

What rigour means here

Words that
have to mean
something.

Rigour is not a tone. It is a short list of specific habits, each one aimed at a way empirical finance is known to fool itself.

  • A partition kept out of reach of the environment where ideas are developed. It is read after the specification is frozen. Reading it twice turns it into more study data.

    AskCould anyone have seen this partition before the idea was fixed?

How the work is organised

Many hands.
One record.

Researchers and agents work on separate machines at once. A frontier can be held by one of them at a time, every run lands in a single immutable history, and a failure found on one desk is known to all of them.

  • Exclusive claims. Two desks never spend a week testing the same idea.
  • One history. Each experiment is committed with its author, data and specification.
  • Shared scars. A falsified idea becomes a screen that every future hypothesis passes through.
Frontier boardclaims
Acting as
  • EquitiesCross-sectionDesk two
  • FuturesTerm structureOpen
  • OptionsSurface fitCompute node
  • MicrostructureExecution costOpen

Claim a frontier that is already held, then release one as falsified.

Latest papers

Quantitative Finance

All papers
  • 5 Oct 2026Gate stacks are not products: measuring the rubber-stamp risk of a research pipelineIn a simulated research pipeline, correlated gates protect up to 18 times less than their pass rates imply, and a one-draw placebo is a coin flip.19 min read
  • 5 Oct 2026When a failure record lies: the economics of shared negative results in a research swarmIn a simulated swarm, sharing falsified ideas saves a third of compute, but neighbourhood rules close off real effects unless records carry their test power.17 min read

RelatedIsolated research workspacesHelper plane

Further reading

The published work this discipline leans on. Not endorsements, only where to start.

  • The Probability of Backtest OverfittingBailey, Borwein, López de Prado, Zhu · Journal of Computational Finance
  • The Deflated Sharpe RatioBailey, López de Prado · Journal of Portfolio Management, 2014
  • Pseudo-Mathematics and Financial CharlatanismBailey, Borwein, López de Prado, Zhu · Notices of the AMS, 2014
  • … and the Cross-Section of Expected ReturnsHarvey, Liu, Zhu · Review of Financial Studies, 2016
  • BacktestingHarvey, Liu · Journal of Portfolio Management, 2015
  • Replicating AnomaliesHou, Xue, Zhang · Review of Financial Studies, 2020
  • Does Academic Research Destroy Stock Return Predictability?McLean, Pontiff · Journal of Finance, 2016
  • Optimal Execution of Portfolio TransactionsAlmgren, Chriss · Journal of Risk, 2000–01
  • Continuous Auctions and Insider TradingKyle · Econometrica, 1985
  • Empirical Market MicrostructureHasbrouck · Oxford University Press, 2007
  • The Volatility Surface: A Practitioner’s GuideGatheral · Wiley, 2006

Bring a claim.
We will try to break it.

Quantitative workspaces are provisioned per mandate, with lead approval on every write.

Request access