Missing inputs reduce confidence
The model tracks inputs used versus inputs possible. If coverage is thin, the score is confidence-adjusted toward neutral instead of pretending the missing data is bullish or bearish.
Public model notes
How the TECHi Forward model turns data into a research score.
TECHi Intelligence is a transparent research layer for quote pages. It combines market, fundamental, analyst, technical, sentiment, leadership, and risk inputs into a confidence-adjusted 0-100 score. The score is directional context, not a guarantee.
The model will keep improving. Every meaningful change gets a version number, a short plain-English note, and a public record so readers can see how TECHi Intelligence is changing over time.
Current model v1.0 · Last reviewed May 28, 2026
Versioned model
Current public release: v1.0.
This release documents the first public TECHi Forward methodology: factor weights, confidence adjustment, stance bands, missing-data rules, and reader limitations. Future changes will be logged before they are presented as normal site behavior.
Formula
The current composite weights.
Composite = 12% technical tape + 12% momentum + 12% track record validation + 14% quality + 11% future value + 10% valuation + 10% analyst/revision + 6% sentiment + 4% leadership + 2% social/developer traction + 7% risk brake. Missing fields pull the final score toward neutral based on model confidence.
Technical tape
12%
Trend, realized volatility, range behavior, and price structure.
Momentum
12%
Recent return profile and persistence across short and medium windows.
Track record validation
12%
How matured historical horizon scores have tracked realized returns.
Quality
14%
Profitability, margins, cash-flow quality, balance-sheet durability, and operating consistency.
Future value
11%
Growth runway, revenue durability, estimate direction, and long-horizon setup.
Valuation
10%
Multiples, yield-style measures, and valuation pressure versus fundamentals.
Analyst and revision
10%
Target range, recommendation mix, revisions, and consensus dispersion.
Sentiment
6%
News and market attention where provider coverage is available.
Leadership
4%
Leadership and governance context when supported by issuer or curated data.
Social and developer traction
2%
Available community, product, and developer attention signals.
Risk brake
7%
Drawdown, volatility, debt, data gaps, and other penalties that reduce conviction.
Output bands
What the score labels mean.
74-100
Constructive setup
61-73
Positive but selective
47-60
Balanced watch
34-46
Cautious setup
0-33
High-risk setup
These labels describe a research setup. A constructive score can still be wrong, and a high-risk score can still precede a rally. Readers should inspect the factors, confidence level, and source data before relying on any output.
Guardrails
What keeps the model from looking more certain than it is.
Backtests are labeled by maturity
Rolling validation only appears after enough stored closes and factor snapshots mature. When accuracy tracking is pending, the UI says so.
Narrative does not create the score
The written explanation can make a score easier to read, but it does not create the score. The composite comes from the deterministic factor model and the resolved data layer.
No trade instruction
TECHi Intelligence is a research screen. It is not a broker recommendation, price target, or personalized investment instruction.
Reader limitation
TECHi Forward scores are informational research tools. They are not personalized advice, trade signals, guaranteed predictions, or a substitute for a broker quote, original filing, exchange feed, or licensed professional review.