← Back to TradinZ

Methodology & Disclosures

This page describes how every computed number on TradinZ is produced, as required for investment recommendations under EU Market Abuse Regulation (Art. 20, Delegated Regulation 2016/958). Draft pending legal review — [Legal Entity] details must be completed before commercial launch.

Last updated: September 2026

1. Who produces this content

All model outputs on TradinZ are produced automatically, by software, operated by Dinz Studio ([Legal Entity], [address], Croatia — contact [email]). No analyst reviews individual outputs before display. Outputs regenerate whenever the underlying data refreshes; each model section carries its generation date. TradinZ is not an investment adviser and produces no personal recommendations; every output is identical for every user (see the full disclaimer).

Conflict of interest disclosure: the operator of TradinZ may personally hold long positions in securities covered by the site, including securities shown in example screens. Positions are not disclosed per-security at this time; no output is influenced by, or produced with knowledge of, those holdings — the models are mechanical. [Complete or amend per counsel before launch.]

2. Data sources

Facts (filed financials, prices) are shown as facts; everything derived from estimates or models is an estimate and is labeled as such where shown. Market data may be delayed.

3. How each output is computed

Analyst price target ("fair price")

The average 12-month price target of analysts covering the stock, as aggregated by stockanalysis.com. This is a reproduction of third-party opinion, not a TradinZ valuation. Analyst targets historically skew optimistic.

Bear / Base / Bull 5-year scenario models

Three mechanical what-if paths built from: this-fiscal-year consensus revenue; next-FY consensus growth faded yearly by a fixed schedule (100% → 55% → 35% → 25% → 20% of the starting rate) and scaled per scenario (bear ×0.55, base ×1.0, bull ×1.25, floor 3%); trailing-twelve-month net margin (bear compresses it 20%, bull expands 15%, ramped linearly across the horizon); share count held flat; and a price-to-earnings multiple taken from the company's own historical valuation band (25th percentile / median / 75th percentile of roughly five years of monthly observations). Where no usable history exists, an assumed ±30% band around the current multiple is used and flagged. Each year's model price = that year's earnings per share × the scenario multiple. "Worth today" discounts the year-5 price at a 10%/yr required return. These are hypothetical illustrations with every assumption displayed, not forecasts or price targets by TradinZ.

What's priced in (reverse DCF)

Solves a standard 10-year discounted-cash-flow model backwards: the constant free-cash-flow growth rate that makes the model's value equal today's market price, using a 10% discount rate and a 15× terminal multiple. The result is compared, for scale, against next-FY consensus revenue growth and the company's SEC-filed historical revenue growth. Caveat shown in-app: the implied rate is on free cash flow while the comparison lines are revenue growth — related but not identical metrics.

Revision momentum

A -100…+100 composite of three observable analyst behaviors: 30-day breadth of next-FY EPS estimate revisions (up vs down, ±45 points, available only when estimate-revision counts are published), 45-day named-analyst price-target raises vs cuts (±30), and 90-day upgrades vs downgrades (±25, deduplicated across sources). It summarizes what analysts are doing, not what anyone should do.

SEC-verified Piotroski F-score and Altman Z-score

Computed directly from XBRL facts in the company's annual 10-K filings: Piotroski's nine published accounting tests (each shown with its pass/fail and inputs; tests with missing inputs are skipped, not failed — the score reads X out of testable) and Altman's original Z formula using the filed balance sheet with live market capitalization for the equity term. Annual-report basis, so scores can differ from trailing-twelve-month variants published elsewhere. Z-scores are unreliable for financial companies and inflated for asset-light businesses; noted in-app.

10-K redlines

A sentence-level text comparison of the Risk Factors section (Item 1A) between the two most recent 10-K filings, with digits normalized so routine year/amount updates don't register as changes. Added and removed sentences are shown verbatim with links to both filings. This is a text diff of the company's own disclosures; TradinZ adds no interpretation.

Scenario builder

Interactive versions of the models above where you choose the assumptions. Builder outputs are your model, not a TradinZ forecast.

Portfolio X-Ray and signal labels

Descriptive aggregations of the same public, impersonal signals shown on every ticker page (consensus, analyst actions, 13F filings, revisions, quality scores), value-weighted across holdings you have connected or entered. Labels such as "Broadly positive signals" describe the balance of those third-party signals; they are not ratings, advice, or recommendations, and nothing considers your personal circumstances.

4. Reliability, limits, and errors

All outputs are mechanical estimates from the stated methodologies and public data. They may be wrong — through source errors, stale data, currency mismatches on foreign listings, model limitations, or bugs. Sanity gates suppress obviously broken values (for example, implausible multiples), but suppression is not verification. Past performance does not indicate future results. Verify important figures against primary sources before relying on them; to report an error, contact [email].

5. AI-generated content

Some summaries of public commentary are generated by artificial intelligence and carry an "AI-generated" label where shown. They may contain errors or misattributions; the linked sources are authoritative.