QuantML V11 · Stock predictions powered by machine learning

Stock Predictions. Your Decisions.

QuantML® helps people make better-informed investment decisions.Our machine-learning technology analyses S&P 500 stocks and predicts whether selected stocks are more likely to rise or fall.Our goal is to make those predictions clear and accessible to everyday investors.The investment decision always remains yours.

QuantML V11 is currently in development. Planned launch: Q1 2027.

Explore QuantML Become a Founding Member

How QuantML Works

QuantML V11 will analyse eligible stocks across the S&P 500 and publish predictions only when its models' qualifying criteria are met. Not every analysed stock will receive a prediction. A published prediction will indicate whether the models expect the selected stock to rise or fall over its stated evaluation period. You make the final decision.

QUANTML / HOW IT WORKS

From information to probabilistic insightANIMATED EXPLANATION · SELECT TO EXPLORE

QUANTML APPapp.quantml.ai · user-facing views

app.quantml.ai / Home

01 · DAILY PREDICTIONSSummary ResultsV11 in developmentFictional design preview

app.quantml.ai / Results

02 · RESULTS SUMMARYEvery result has a route.

91settled days+108.85%arithmetic sum

Historical · simulated

app.quantml.ai / My Portfolio

03 · WORKFLOW CONNECTIONConnect your workflowYou decide; QuantML never places orders

DIFFERENT VIEWS
SHARED DISCIPLINE

01 / GATHER INFORMATION

Different observations become structured features.

02 / TRAIN TEN VIEWS

Tap any anonymous model view to trace its contribution.

03 / TEST AGAINST HISTORY

TRAINTEST

Evaluate on periods beyond those used for training.

04 / COMBINE THE VIEWS

The ensemble brings the individual views together.

05 / GENERATE A SIGNAL

A signal is an insight, not a certainty.

06 / DAILY PREDICTIONS

The planned QuantML V11 App will present selected stock predictions. The tickers here are fictional examples, not published predictions.

07 / FINGERPRINT THE BATCH

The exact batch bytes become one fingerprint. A later change produces a different value.

08 / COMPARE THE BYTES

Recalculate the fingerprint and compare it with the original recorded value.

09 / BASE BLOCKCHAIN RECORD

A public record can establish the fingerprint and its block time. It does not establish prediction accuracy.

10 / PREDICTION RESULTS

Review individual historical prediction outcomes after their evaluation period. These are separate from any blockchain evidence of publication and do not represent portfolio returns.

11 / HISTORICAL PERFORMANCE

Completed simulated / hypothetical record. Arithmetic sum, gross of fees; not live capital or a forecast.

Illustrative ten-model ensemble. No live training, testing or predictions are displayed.

Why QuantML is Different

Built to make machine-learning stock predictions easier to understand and review

QUANTML / 06

From principles to decisionsINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Capital preservation precedes performance optimisation; exposure limits apply before a signal is ranked.

Candidate probability P(up) 78%Requested exposure 42%Illustrative position cap 20%

bounded = min(request, cap)= min(42, 20) = 20% · 22 pts held backrank = order by P(up), after the gateP 78% → #2 of 4; size is still 20%

Requested42%before the gate

Bounded20%probability not used

Unallocated80%of toy sleeve

Order createdNoneillustration only

Plainly. The gate runs before ranking, so probability can change where a candidate sits in the queue but never how far past the cap it can go. At P 78% the candidate ranks #2, yet its size is min(42, 20) = 20%. Toy rule and numbers, not QuantML’s deployed limits; a cap is a sizing bound, not protection from loss.

Educational experiments with synthetic, local numbers. Not advice, not live signals or predictions, not QuantML performance or deployed model settings.

HOW QUANTML WORKS

The process, at a glance

STAGE 01 / 04

Broad information context

A broad view of changing market conditions provides context. No single observation determines an outcome.

Wide market view

Conceptual illustration · No live predictions or model details shown · Not investment advice.

Market Data

Market data is structured into model-ready features using quantitative techniques. The system evaluates price behaviour, statistical patterns, and other market characteristics to identify potential opportunities.

Prediction Generation

V11 models will analyse eligible S&P 500 stocks and publish directional predictions only for those meeting their qualifying criteria. A confidence score represents the model's assessment of prediction strength. It is not a verified probability of success or a guarantee.

Risk-First Allocation

The research framework includes risk context alongside predictions. Portfolio or strategy results depend on execution, allocation and cost assumptions and are distinct from individual stock movement.

Transparent Reporting

The platform presents:

●published predictions and recorded outcomes, including incorrect directions
●portfolio weights
●benchmark comparison
●drawdown and volatility metrics

All results are displayed through the QuantML website and QuantML App.

V10 / V10.1 · Completed reporting period 1 Dec 2025 – 14 Jul 2026 · 91 settled days, every one on the record

+108.85%

$10,000 → $20,885

Arithmetic return sum

Chapter 4 of 4

Full period

SPY +10.52%

Simulated / hypothetical · $10,000 notional · gross of fees · not live capital · past performance does not guarantee future results.

$10k$15k$20kSPY (S&P 500) +10.52%

Arithmetic return sum+108.85%Avg +1.20% per settled day

Compounded return+159.01%$10,000 → $25,901.05

Arithmetic alpha vs SPY+98.33%Arithmetic return minus SPY

SPY · S&P 500 buy & hold+10.52%$10,000 → $11,051.94

Sharpe ratio2.80All trading days, risk-adjusted

Max drawdown−30.45%Peak to trough

Win rate59.3%54 of 91 settled days

Beta · correlation to SPY0.293 · 0.058Moderate — partial market link; Low — alpha-driven

Inspect Results →

Simulated / hypothetical · $10,000 notional · gross of fees · not live capital · past performance does not guarantee future results. Long/short · 1–3 names a day, equal weight · 7-day hold. Reproducibility: every figure derives from the QuantML signal ledger and can be reproduced from the daily signal outputs.

QUANTML / 02

Open the record. Follow the evidence.INTERACTIVE EXPLANATION · SELECT TO EXPLORE

Recorded checkpoint 2026-07-14 · 91/91

QuantML · arithmeticSPY · buy and hold

QuantML · arithmetic$20,885.10+108.85%

SPY · same checkpoint$11,051.94+10.52%

Basket at checkpoint+0.27%recorded seven-day holding return

Move the checkpoint to reveal history from the same completed record. Arithmetic adds the recorded basket returns; compounded reads the published equity curve. They measure different things.

Source: the same frozen public AuditSnapshot used above · v10_1_public_2026-07-14. Simulated / hypothetical, gross of fees, not live capital. Dates and values come from that record; no future performance or substitute ledger is generated.

IN FOCUSCompleted record

Context for this stage

The original completed, simulated performance record remains above this illustration.

QUANTML / 03

Inside the three failure modesINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Turbulence choppyShort window starts at step 18

Whole period+10.0%same in every setting

This 10-step window+4.6%depends on timing

Every 10-step window−6.1% to +8.7%range of outcomes

Deepest dip−9.6%largest step 6.6%

Takeaway. The destination is unchanged, but short windows now range from losses to gains. Over a short horizon, the path decides the outcome more than the average does.

Interactive explanations built from synthetic, fixed sample shapes. No live values, no QuantML performance, no claim that any risk is eliminated.

IN FOCUSMarket Volatility

Context for this stage

S&P 500 averages 10% yearly, but daily fluctuations make short-term investing risky. Traditional quant strategies no longer deliver consistent alpha.

Why Risk Matters

Markets are unpredictable. Long-term capital growth depends on disciplined exposure control.

This is structured risk management — not predictive certainty.

QUANTML / 05

Discipline under changing conditionsINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Outer ring · uncappedInner ring · cappedCash (excess)Cap length

Asset A weight 40%Shock to Asset A −25%Illustrative cap per position 20%

Uncapped change−10.00 ptscash 0.0%

Capped change−5.00 ptscash 20.0%

Asset A · move −25.0%

Uncapped weight 40.0% → contribution −10.00 pts. Capped weight 20.0% → contribution −5.00 pts. The cap moves 20.0% of it to cash.

Inspect every weight and loss contribution

Weights and contributions (weight × move)
Holding	Move	Uncapped	Capped
A	−25.0%	40.0%−10.00 pts	20.0%−5.00 pts
B	0.0%	18.0%0.00 pts	18.0%0.00 pts
C	0.0%	15.0%0.00 pts	15.0%0.00 pts
D	0.0%	15.0%0.00 pts	15.0%0.00 pts
E	0.0%	12.0%0.00 pts	12.0%0.00 pts
Cash	0.0%	0.0%0.00 pts	20.0%0.00 pts
Total		100.0%−10.00 pts	100.0%−5.00 pts

Plainly. A position’s contribution is its weight times its move. Capping Asset A at 20% shrinks its direct loss contribution. Real holdings rarely move independently; try a co-move setting. Cash is held flat in this example and can miss gains. An illustrative cap is a sizing limit, not guaranteed protection.

Educational experiments built from synthetic, fixed inputs. Not advice, not live signals, not QuantML performance, and no claim that any risk is eliminated.

IN FOCUSMarkets are unpredictable

Context for this stage

Markets are unpredictable. Long-term capital growth depends on disciplined exposure control.

QUANTML / 07

Open the evidence behind each controlINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Evidence inspector · stage 1 · example only

Vote record: 7 up, 3 down, majority UP, vote share 7/10.

Up votes7

Down votes3

MajorityUPvote share 7/10

Disagreements3models against the majority

Plainly. Combining several models makes disagreement visible instead of hiding it. Vote share counts models; it is not a calibrated probability or confidence, and a tie is reported as a tie. Ten anonymous toy voters, not QuantML’s actual ensemble composition.

Educational examples with synthetic, local values. No real records, alerts, approvals, certifications or deployed QuantML settings are shown or created.

IN FOCUSML Ensembles

Context for this stage

Multiple specialised models trained per ticker and sector cluster, combining diverse predictors for robust signal generation.

About QuantML

QuantML® helps people make better-informed investment decisions. Our machine-learning technology analyses S&P 500 stocks and predicts whether selected stocks are more likely to rise or fall. Our goal is to make those predictions clear and accessible to everyday investors. The investment decision always remains yours.

Institutional-Grade by Design

QuantML integrates flexibly across institutional workflows, acting as a signal and risk layer.

Asset Managers & Funds

Signal generation for discretionary or systematic strategies. Independent risk-aware validation layer alongside existing models.

Signal integration
Risk validation layer
Managed strategy sleeve

Brokers & Trading Platforms

White-labelled QuantML signals for retail or professional clients. Premium subscription products without building a signal platform in-house.

White-label ready
API integration
Revenue diversification

Institutional Investors

Transparent governance, audit readiness, and compliance-friendly reporting for sophisticated allocators.

Full audit trails
Partner oversight

QuantML is designed for users who prioritise risk discipline and transparency over speculative trading.

QUANTML / 08

Your workflow. One intelligence layer.INTERACTIVE EXPLANATION · SELECT TO EXPLORE

Follow one fictional research signal. The animation stops where your process begins.

QUANTML PROVIDESResearch with uncertainty

YOUR INSTITUTION DECIDES

FROM QUANTMLFICTIONAL

Instrument A illustrative

DirectionPositive bias · short horizon

UncertaintyAdded at risk view

Risk contextAdded at risk view

SAMPLEA fictional research signal arrives with its direction and horizon.

YOUR DECISION

Fictional, local illustration. No live signal, order routing, execution or confirmed integration. Your institution controls every decision after the handoff. Not investment advice.

Market Opportunity

Explore three market figures and how they relate to the potential reach of QuantML.

QUANTML / 09

Explore the market landscapeINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Focus regionEuropeillustrative placement

Routes emphasised6/10capital layer

Traced routeAmericas → Europecapital · emphasised

The Capital layer is on: trunk routes link centres in different regions. Routes touching Europe are emphasised. The $120T+ figure does not size or drive anything on this globe.

Mode-button figures are QuantML marketing figures; they are not independently verified or recalculated here. Continents are approximate. Node positions near financial centres are for orientation only — not clients, offices, users or a live footprint. No live data, feed or backend.

IN FOCUSGlobal Asset Management

Context for this stage

A $120T+ addressable industry seeking improved allocation and risk insight.

OPEN ENGINEERING · QUANTML V11

Watch V11 take shape

A governed AI operating system: bounded build capacity, ten specialist roles, two AIs that must agree, and human approval before anything from the pipeline merges or ships.

Live service feed

Delivery and worker activity do not establish improvement over V10. Comparative performance evidence is not published in this panel.

V11 delivery progress

Canonical V11 roadmap · Published 10/31/2026, 8:02:56 AM · updates when the roadmap changes

Capabilities deliveredRecorded status, not release or performance proof.

Evidence recorded

64/ 213

Blocked capabilities

Stage 117 / 23Data integrity

Stage 27 / 20Feature engineering

Stage 356 / 66Model optimisation

Stage 415 / 15Backtest trust

Stage 523 / 60Pilot retrain

Stage 60 / 5Full-universe retrain

Programme delivery fabric

Governed work moves from backlog → execution → independent review → human approval → release.

V11 DELIVERY FEEDV11 roadmapCanonical V11 capability roadmap

Ready to start

Roadmap items cleared for autonomous dispatch now

Still to deliver

78of 213

V11 capabilities not yet delivered

Also queued: approved backlog 0 · bug fixes 0Roadmap evaluated 12:53:20 AM. The arrow into ORC moves only while this is fresh and work is ready.

Slots occupied

0of 4

Under review

2independent

BUILD SLOTS

SLOT 1 · OPEN

SLOT 2 · OPEN

SLOT 3 · OPEN

SLOT 4 · OPEN

WHERE THE WORK LANDS

GitHubCode · pull requests · reviews

HEALTHY

Base44Operations UI · API · deployment

CONNECTED

Git worktreesIsolated implementation checkouts

0 IN USE

Moving green: verified live work
Moving amber: independent review
Still: idle, unverified or unconfirmed
Ivory: human approval gate
Providers are interchangeable per role: Claude, ChatGPT.

HUMAN AUTHORITYAI agents propose. Human decides.

Claude and ChatGPT must agree before a human is asked. Nothing merges from the pipeline, and nothing reaches production, without human approval.

DUAL-AI AGREEMENT
ClaudeBUILDSChatGPTACCEPTS
Both AIs must agree first
Whichever provider builds, the other must accept the exact final code. A rejection sends the work back for fixes. Only then is a human asked.
HUMAN GATEApprove merge
Human - Release Manager
EVERY PIPELINE PULL REQUEST
AUTOMATEDGoverned merge
Executed automatically, only after approval.
HUMAN GATERelease to production
Human - Head of Product
EXPLICIT APPROVAL · NO EXCEPTIONS
SHIPPEDLive
In production for QuantML users.

Service health

Simplified service view · heartbeats observed on this page over the last five minutes · Updated 11/3/2026, 12:56:07 AM

WatcherHealthy
SentinelHealthy
ORC sessionNeeds attention
Telemetry PublisherHealthy
Command ApplierHealthy

Transparency First

Every signal, every decision — fully auditable

Risk-Led Framework

Exposure controls are applied before allocation

Repeatable Results

Systematic, not discretionary

Institutional Benchmarking

Performance vs SPY shown transparently

Next Generation

QuantML V11 is currently in development

New market opportunities and daily signals are temporarily unavailable while we complete and validate the next generation of the QuantML platform. Historical V10/V10.1 results remain available in Results summary through 14 July 2026.

Complete architectural redesign

QuantML V11 has been rebuilt from the ground up rather than being an incremental update to previous versions.

Controlled model development

Uses defined review and validation steps as models are developed and updated.

Execution intelligence

Focuses on closing the gap between theoretical AI performance and realised investment performance through enhanced execution, validation, and monitoring.

Institutional-grade transparency

Provides end-to-end auditability, execution traceability, and performance attribution to support governance and trust.

Foundation for future innovation

Establishes a scalable platform architecture capable of supporting new AI capabilities, automation, and continuous evolution.

View historical V10/V10.1 Results (through 14 July 2026)

Partner with QuantML

QuantML is opening discussions with select strategic partners with expertise in trading, risk management, and portfolio construction

QUANTML / 10

Connect expertise with capabilityINTERACTIVE EXPLANATION · SELECT TO EXPLORE

Selected3 of 5a count, not a score

Cards on agenda3example evidence

REGIME NOTES adds how signals and drawdowns behaved across different market regimes
RISK CHECKLIST adds institutional-level questions to put to the risk controls
CONSTRUCTION not on this example agenda
REVIEW CADENCE adds operational discipline for review and sign-off rhythms
HORIZON not on this example agenda

Example only. Each switch decides which evidence card assembles on the shared agenda. Nothing here rates, ranks or qualifies a partner.

Partner brings expertise, risk oversight and strategic input.

QuantML brings the platform, model and risk evidence, and a roadmap seat.

Boundary training code stays with QuantML; capital deployment stays external.

Illustrative example of a possible collaboration. No partner score, agreement, delivery, certification or terms are shown or implied; specific terms are co-defined.

IN FOCUSThe Right Partner

Context for this stage

Experience across market regimes, institutional risk oversight, portfolio construction and governance, with long-term alignment. Terms are flexible and co-defined.

Ready to Explore?

Important Information

QuantML provides general, non-personalised stock predictions and supporting research for information only. Predictions can be wrong and outcomes or returns are never guaranteed.

QuantML does not:

•Provide personalised investment advice
•Manage client funds
•Place live orders for members
•Operate as a broker, custodian, or discretionary portfolio manager

All outputs are general information made available equally to all users and are not tailored to individual financial circumstances.

Users remain solely responsible for all investment decisions and trade execution.

Investments can fall as well as rise, and you can lose some or all of the money invested. Historical simulated results and paper trading do not guarantee future results. Optional Alpaca paper trading uses simulated money and requires the user's separate authorisation.

For full details, please review our:

Risk Disclosure Terms of Use Privacy Policy

QUANTML / 12

Read the boundariesINTERACTIVE EXPLANATION · SELECT TO EXPLORE

QuantMLAnalytics & information

WITHIN THE INFORMATION LAYER

Market signals

QuantML market signals are provided for informational purposes only.

Choose an action. The connection opens for information and stops at the boundary for advice, funds or execution.

Read the complete termsRisk Disclosure Terms of Use Privacy Policy

The complete original disclosures remain above. These local illustrations explain their boundaries; they do not change the platform’s scope or your responsibilities.

IN FOCUSPlatform scope

Context for this stage

QuantML provides general, non-personalised stock predictions and supporting research for information only. Predictions can be wrong and outcomes or returns are never guaranteed.

Stock Predictions. Your Decisions.

How QuantML Works

Why QuantML is Different

The process, at a glance

Broad information context

Market Data

Prediction Generation

Risk-First Allocation

Transparent Reporting

+108.85%

The Problem

Why Risk Matters

Technology & Governance

About QuantML

Institutional-Grade by Design

Asset Managers & Funds

Brokers & Trading Platforms

Institutional Investors

Market Opportunity

V11 delivery progress

Programme delivery fabric

Service health

QuantML V11 is currently in development

Complete architectural redesign

Controlled model development

Execution intelligence

Institutional-grade transparency

Foundation for future innovation

Partner with QuantML

Ready to Explore?

Important Information