Home Verified Methods Reports Papers Methodology About
Last close
Market data updates when available
Overview Research Areas Timeline Publications Contact
QM
Independent quantitative research — published with code

QuantMedia

Written and operated by Cemil Ertürk · market microstructure, portfolio construction and execution costs
5 research notes 2 daily proprietary indices Runnable code · 28 tests VPIN · HRP · PSR Not peer reviewed
Contact
Overview
Research

Overview

QuantMedia is an independent quantitative research site written and operated by Cemil Ertürk. Its subject is whether published quantitative finance methods work as described once they are implemented: each method here ships as runnable Python with a synthetic input whose correct answer is known in advance, tests that guard the fixes, and a dated verification report of what the code got wrong before it was corrected.

Two methods, VPIN and Hierarchical Risk Parity, have runnable implementations with example data and 28 tests. The Probabilistic Sharpe Ratio is covered by a worked calculation, checked step by step, and a free calculator. The remaining research notes are analytical and are not presented as validated backtests.

A daily pipeline scores a fixed 180-name US equity universe after each close and publishes two indices derived from that scan, Signal Breadth and Sector Confluence, together with their freshness state. Nothing on this site is investment advice.

Research Areas

📊
Order Flow Toxicity
A runnable VPIN estimator covering equal-volume bucketing, bulk volume classification and the tick rule, with 15 tests and a verification report. The example data is synthetic and states its own limits.
VPINPython15 tests
🧩
Portfolio Construction
A runnable Hierarchical Risk Parity allocator compared with minimum variance, shrinkage and equal weight on a synthetic panel, with 13 tests and a verification report.
HRPPython13 tests
📐
Backtest Statistics
The Probabilistic and Deflated Sharpe Ratios: a worked calculation checked step by step, a free calculator, and explainers on track-record length, skewness and kurtosis.
PSRDSRCalculator
💱
Execution Costs
Bid-ask spread dynamics and slippage modelled as spread, square-root impact and delay cost, with a worked example of what turnover does to a strategy net of costs.
SpreadSlippageImpact
📡
Daily Signal Pipeline
A post-close scan of a fixed 180-name US universe against 30 technical signals, publishing Signal Breadth and Sector Confluence with an explicit freshness state and a fail-safe that never labels stale data as live.
BreadthConfluencePipeline

Research Timeline

Aug 2026
Reproducible implementations and proprietary indices published
Runnable VPIN and Hierarchical Risk Parity packages released with example data, expected output and 28 tests. Two original daily metrics published with formulas, JSON and non-backfilled history: the Signal Breadth Index and Sector Confluence Index.
Apr–Jun 2026
Data pipeline outage — detected, disclosed and repaired
The scheduled update job failed silently from 14 April and the automated schedule was disabled after 60 days of inactivity. The cause was a continuous-integration misconfiguration, not a data-source failure. The pipeline was rebuilt with per-source failure isolation, last-known-good preservation and a published freshness state at /data/status.json, so a future stall is visible rather than silent. Recorded here under the corrections policy.
Apr 2026
Signal methodology v2.0
The current confluence engine took effect on 14 April 2026: 30 technical signals scored across a fixed 180-name liquid US universe, with a 22-of-30 agreement threshold. Parameters are published in signal_config.json and the build fails when site copy stops matching them.
Jan–Mar 2026
Research library published
Eleven research papers covering order-flow toxicity, spread decomposition, portfolio construction, performance evaluation under multiple testing, and execution costs. Each carries primary references and an explicit limitations section. Six were retired in September 2026 when the site narrowed to verified methods; five remain.

Research Notes & Publications

VPIN & Order Flow Toxicity
Research note · runnable Python implementation · 15 tests
Hierarchical Risk Parity
Research note · runnable Python implementation · 13 tests
Probabilistic Sharpe Ratio
Research note · free interactive calculator

Read the worked examples or browse all five research notes. Code availability and limitations are listed in the implementation index.

Contact & Collaboration

QuantMedia is open to research collaboration with academic institutions, independent quantitative researchers, and technology partners working on market microstructure, low-latency systems, or statistical finance.

All research outputs are published on this portal. For collaboration inquiries, reach out via the contact page. The lab does not accept unsolicited trading strategy proposals or investment-related inquiries.

QuantMedia is independently operated and unaffiliated with any fund, broker, exchange or data vendor. It currently earns no revenue; an advertising application is pending, and the editorial policy sets out what would and would not change if it were approved. All findings are provided for academic and heuristic purposes only, and do not constitute financial advice under any regulatory jurisdiction.