Research
AINFER Limited · Hong Kong · Business Registration No. 81518145
TradeFlags research is quantitative work on market structure: volatility regimes, breadth, sentiment, supply-chain concentration and forecast behaviour. Every published series on this site is the output of that work — the research programme and the data service are the same thing seen from two sides.
We publish the negative results alongside the positive ones, with the out-of-sample window, the transaction-cost assumption and the sample size stated in the report. A number without its validation status is not a research output.
Upcoming research reports
The queue below is drawn from work in progress across our research repositories. Dates are given
as reporting periods, not as commitments to a publication date.
01
NEXT
Exit timing by regime
The current workhorse strategy’s optimal EXIT is regime-dependent: volatility-normalisation exits are right in V-shaped recoveries and too early in slow recoveries. The next report isolates the exit rule from the entry rule and tests each against the same three out-of-sample windows.
02
PLANNED
Cross-asset capitulation confirmation
The edge lives in buying capitulation, so the trigger should fire on SYSTEMIC stress rather than a local wobble: VHSI, VIX, MOVE and credit spreads spiking together. Tested as an entry filter on the same windows, with the direction fixed before the grid is run.
03
PLANNED
Basket quality filter
The current basket buys the most beaten-down names, which sometimes means buying structurally impaired companies in credit events. A solvency and quality screen over the same candidate set would test whether the edge survives once distress is separated from deterioration.
04
IN PROGRESS
Supply-chain risk scoring
The graph-theoretic risk model — bottleneck scores, shock propagation, concentration by sector and country — is written but not yet scored into the live dashboard. This report carries the scoring method and its validation against the published scenario set.
05
PLANNED
Multi-ecosystem supply-chain graph
Extending the model beyond one ecosystem into a merged graph covering the major semiconductor and systems-integration chains, so that cross-company correlation and shared-bottleneck exposure can be measured across them rather than within each in isolation.
06
MONITORING
Live-regime watching brief
A standing brief on how the deployed live books are behaving against their out-of-sample reference, including the regime conditions under which the historical edge has been observed to weaken. Published as a monitoring note rather than a performance claim.
Newsletter
One email per report, at publication. No other mail.
Subscribe to the research newsletter
Every research report is emailed to subscribers as it is published. Subscription is confirmed
by email — you will receive a confirmation link that must be clicked before the list is activated.
We store your email address and the date of your subscription in order to send you the newsletter.
You can unsubscribe from any email through the link it contains, at any time.
See the Privacy Policy.
Where research is published
Every report reaches readers through all three channels. The newsletter is the complete record;
the other two carry the same findings for audiences that do not read email.
Email newsletter
LIVE
The full report, including the method, the validation status and the negative results,
delivered to confirmed subscribers on publication.
Subscribe on this page — confirmation by email.
Substack
COMING SOON
The same findings written up for a general investment audience, with the tables and charts
that do not fit in a data package.
Publication not yet open — the link will appear here when it launches.
WeChat 公众号
COMING SOON
Chinese-language summaries of each report for mainland institutional readers, published on our
corporate account shortly after the English edition.
Corporate account in preparation — the handle will appear here when it is released.
Not investment advice. TradeFlags outputs are computational audit data only. No dealing, advising or asset-management
activity is conducted. Research reports describe the behaviour of models and data; they are not recommendations.