How articles are produced
on this site.
FX Strategy Analyzer is a one-person project. Articles are written with the help of AI tools and reviewed before publication. Here's the workflow, the sourcing rules, and what to do if you find an error.
1. Who runs this site
FX Strategy Analyzer is built and maintained by a solo developer with hands-on experience in algorithmic FX trading and MetaTrader 5 Expert Advisor development. The same person writes the tools (EA Analyzer Pro, StratForge MT5, Strategy Lab, AI Signal Trainer, and the calculators) and the articles on the blog.
Articles are not bylined under a personal name. The "I" in articles refers to the developer running the site, and the "we" sometimes used in framework articles is a stylistic choice — there is no editorial team, no investment-research desk, and no third-party contributors. What you see here is one trader's work on tools and writing, in public.
2. Editorial mission
The blog exists to support the tools. Articles cover the questions that come up while building or using EA Analyzer Pro, StratForge MT5, and the calculators — how to read a backtest report properly, how to evaluate an EA before risking real capital, how to think about market regimes, how a macro event like an NFP or CPI release transmits into intraday FX. The intended reader is intermediate-to-advanced: someone who can already run an EA in MT5 and is looking for the next layer of rigor.
3. Production workflow
Every article on this site goes through three stages before it is published.
4. AI tools and disclosure
The three stages above are run with the help of large language models — specifically, Anthropic's Claude family of models. The AI is part of the workflow at every stage; it is not the final voice and it does not have final say on what is published. Final accountability for what appears on this site rests with the developer running it.
Every article carries a footer line noting that AI tools were used in production and that the article was reviewed before publication. There are two reasons that disclosure is there:
- Transparency. Readers have a right to know how the content they consume is produced. AI involvement is a material fact about the process.
- Accountability. Disclosing AI use makes it the developer's responsibility — not the model's — to ensure the published output meets these standards.
5. Sourcing and citations
Sourcing rules vary by article type. The underlying principle is consistent: don't dress an unsourced claim as cited research.
For event-driven articles
Specific numerical claims — payroll prints, CPI readings, central-bank decisions, currency pair levels at a given moment, statements from named officials — are drawn from primary sources. That includes statistical releases from official agencies (BLS, BEA, the Federal Reserve, ECB, BoE, BoJ), exchange or broker quote feeds for price levels, and the official transcripts or video records of central-bank press conferences.
For framework articles
Articles that explain methods (walk-forward optimization, EA evaluation criteria, regime detection) draw on a mix of primary literature, vendor documentation (such as the official MT5 Strategy Tester documentation), and recognized practitioner sources. Where a method originates with a specific author — Pardo's work on walk-forward optimization, Kaufman's Efficiency Ratio — that author is named.
What this site doesn't do
- Invent attributions. If a claim can't be tied to a specific source, it's phrased as a general observation ("a common practice", "many systematic traders") rather than presented as cited research.
- Fabricate quotes. Quoted statements come from records that can be pointed to.
- Invent URLs or studies. Cited papers and releases exist.
6. Illustrative numerical examples
Several framework articles include specific numerical thresholds, ratios, percentile readings, and worked calculations — for example "modeling quality ≥ 90%", "ATR percentile 38th", "5-day / 60-day ADR ratio of 1.25". Readers should understand how to interpret these.
Numerical examples used in framework articles are illustrative unless they are explicitly cited to a source. They demonstrate how a metric is computed or what a typical range might look like — they are not measurements of current market conditions or industry-mandated standards.
The threshold values recommended in framework articles (for example "Profit Factor between 1.4 and 2.5", or "OOS profit factor at least 70% of in-sample") reflect commonly-used heuristics in the systematic-trading community and the developer's own experience with EA validation. They are starting points for your own evaluation framework, not industry standards.
7. Financial framing
FX Strategy Analyzer publishes educational content about systematic FX trading. It does not provide personalized financial advice, investment recommendations, signals, or trade calls. References to specific currency pairs, price levels, central-bank decisions, or strategy parameters are part of analysis being explained — they are not guidance for individual trading decisions.
8. Review and fact-checking
Beyond the three-stage production workflow, every article is checked against a specific set of items before it goes live:
- Numerical claims. Specific percentages, dollar figures, and counts must either come from a cited source or be clearly framed as illustrative.
- Quoted statements. Quotes attributed to a named individual must be tied to a specific public statement, press conference, official document, or published interview.
- External links. Links must point to live, relevant pages. Placeholder or fabricated URLs are not used.
- Hedging. Causal and predictive claims about markets are hedged appropriately. Absolute language ("guaranteed", "always", "definitely") is not used in the context of trading outcomes.
- Method vs. recommendation. Articles distinguish between method (how something is calculated) and recommendation (what a trader might do with it), so the reader can disagree with the recommendation without losing the underlying method.
An article that fails any of these checks is held back for revision. Standards are not softened under deadline pressure.
9. Corrections policy
Material corrections — corrected numbers, corrected attributions, corrected dates — are made promptly. When a substantive correction is made, a dated correction note is added at the bottom of the article describing what was changed.
Non-material edits (typos, broken links, formatting issues) may be made silently. Updates that reflect new information after publication (for example, a follow-up event) are added as dated update notes rather than silent edits to the original analysis.
To report a factual error, send an email to fxstrategyanalyzer@gmail.com with a link to the specific article and a clear description of the issue. Identifying the exact sentence and, where possible, the correct source helps it get fixed faster.
10. Contact
For factual corrections, suggested topics, tool feedback, or questions about this editorial process, email fxstrategyanalyzer@gmail.com. Every message gets read, even when an individual reply isn't possible.
Commercial inquiries (advertising, partnerships, sponsorship) use the same channel — please identify the inquiry as commercial in the subject line. Commercial inquiries don't change these standards.