Methodology & Transparency
How we produce our numbers, what we compute, and our limits — no black box.
This page explains the method behind our products. Everything we publish is descriptive: it characterizes past and present data, does not give investment advice, and does not forecast prices. Below you can see what data each product is built from, with which formula, and on which assumptions.
Data collection, freshness and integrity
We pull data directly from primary, publicly available sources; the full provider list is on the Sources page. Each source has its own refresh cadence (from seconds to months). When a price is not current, it is marked with a staleness indicator in the interface. Incoming values pass a server-side sanity check: invalid (zero/negative/undefined) prices are rejected and the last valid value is kept, while unusual jumps are flagged. Gap monitoring runs on the historical series we accumulate ourselves.
Trend score (Signals — Beta)
The trend score is NOT a buy/sell recommendation. The engine first computes a regime-aware directional technical stance: ADX separates trending from ranging regimes, mean-reversion gains weight in ranges, and trend, momentum, mean-reversion and volume factors are combined on a [-1,1] scale. This raw score is then calibrated: it is mapped to the distribution of forward returns that similar scores historically produced over set horizons (5/10/20 days). So the product says 'historically, at this score, this is what happened' — it makes no claim about the future. Every published label is tracked in a scorecard ledger (realized outcome) and stamped with the engine version.
Themed baskets
Baskets track the return of themed asset groups. The engine is Turkish-lira based and offers three lenses: nominal TRY, US dollar, and inflation-adjusted real. Daily returns are chained into a compound return; membership is computed point-in-time (with that day's composition), not retroactively. Maximum drawdown and volatility are provided as risk measures. History is limited to roughly two years (730 days), which constrains long-term comparisons and may carry survivorship bias.
Risk analysis
The risk tool describes a portfolio's past behavior: volatility, maximum drawdown, Sharpe ratio, and beta versus the market. The CBRT/EVDS deposit rate is used as the risk-free rate and the BIST 100 as the market benchmark. The computation is a quantity-weighted simulation over historical price series; past performance is no guarantee of the future and results are sensitive to the chosen period.
Real return and inflation
Real return is nominal return stripped of inflation, computed via the Fisher relation. We use the TurkStat CPI year-over-year change for consumer inflation, and the announced nominal rate together with inflation for the real deposit rate. Inflation series follow the official release calendar and may be revised retroactively.
Converters and calculators
Converter and calculator tools (gold/FX converter, loan, FIRE, gram-gold simulator, dividend, and similar) apply deterministic arithmetic on live data or the date you choose. They show the input clearly, state their assumptions, and promise no future return; scenario options are strictly 'what-if' in nature.
Limits and caveats
Historical depth is about two years for most series. Some assets have a single source; during a source outage that asset may appear stale. Hindsight and survivorship bias, currency/carry costs, and dividend effects can change results. None of our outputs account for taxes, FX spreads, or transaction costs on your behalf.
Disclaimer
All information and tools on this site are for informational purposes only and do not constitute investment advice. The content is descriptive; you are responsible for your own investment decisions. Investment advice is provided on a personalized basis by authorized institutions.