IndiaPulse research methodology

U-Score fundamental model

U-Score is a 0-100 research ranking. Five pillars contribute 100 base points. A Piotroski F-Score bonus can add up to five points, while liquidity and adverse momentum adjustments can reduce the total. The final score is clamped to 0-100. Banks use a sector-appropriate valuation track because a leveraged balance sheet is part of their operating model.

Verdict bands are model labels: Deep Value at 75 or above, Undervalued at 60-74, Fair Value at 46-59, Watchlist at 32-45, and Overvalued below 32. They describe the model output, not a buy, hold, or sell recommendation.

Technical analysis and chart signals

Technical calculations run on stored daily OHLCV candles. The standard set includes EMA 20, EMA 50, EMA 200, RSI(14), MACD(12,26,9), ATR, volume versus its 20-session average, and a true 252-session high-low range. A setup is not promoted when the required history is missing; shorter windows are not silently substituted for EMA 200 or a 52-week range.

Candlestick and chart-pattern candidates are rule-based detections. IndiaPulse combines pattern evidence with liquidity, trend, volume, fundamental U-Score, and management trust filters so a visual pattern is never treated as sufficient evidence on its own.

Volatility Contraction Pattern screen

The VCP screen runs after the daily candle refresh across every active NSE stock in the IndiaPulse database. A stock needs at least 260 valid daily OHLCV sessions and a price of at least Rs 50. The trend template requires price above the 50, 150, and 200-day simple moving averages; 50-DMA above 150-DMA above 200-DMA; rising 50, 150, and 200-day averages; price within 25% of its 52-week high and at least 30% above its 52-week low; and price above its 10-DMA.

Weekly confirmation uses completed exchange weeks only. The latest completed weekly close must be above its 10-week average, which must be above its 50-week average. Relative strength is ranked from 20% three-month, 30% six-month, and 50% twelve-month returns across the full history-eligible universe. Candidates require an RS percentile of at least 70 and a rising stock-to-Nifty-50 relative-strength line.

Contraction is a deterministic approximation, not visual recognition. The nested 40, 20, and 10-session high-low ranges must progressively tighten, the final 10-session range must be no more than 10%, 14-day ATR must compress, and 10-day and 50-day average volume must show dry-up. The first 40-session range is constrained to 8-35% to reject both flat noise and excessively loose structures.

A triggered breakout needs a close above the prior 50-session high, at least 1.4 times average 50-day volume, and a close in the top 35% of that candle. It remains Triggered for at most three sessions and while no more than 5% above the pivot; later detections are Extended. A close 2% below the pivot or below the final-contraction low marks the structure Invalidated. Large, mid, other, and micro or SME stocks must average at least Rs 5 crore, Rs 2 crore, Rs 1 crore, and Rs 0.5 crore of daily traded value respectively.

The Nifty 50 above its 50-DMA is a separate market gate. Valid structures remain visible when that gate is off but are marked market-blocked. Scheduled results within seven calendar days are flagged rather than silently removed. The displayed structural low is an invalidation reference, not a promised stop price; gaps and illiquidity can cause materially worse execution.

Signal Ledger and outcome measurement

The weekday morning publication freezes candidates at 8:45 AM IST using the latest completed stored data. Technical entries come from the Nifty 500 bullish pattern scan, require a setup score of at least 60, a model reward-risk ratio of at least 1.5, minimum liquidity and volume evidence, and valid target and invalidation levels. The published list is capped at eight names and two names per sector to reduce concentration. Long-term entries are capped at six and come from the existing quality-value and compounder Playbook filters.

The reference price is the source-data snapshot used for measurement; it is not an execution claim. Technical outcomes are written once after 1, 5, and 20 complete trading sessions. Long-term candidates are measured after 20, 60, and 252 sessions. Returns use the horizon close, while maximum favourable and adverse movement use stored daily highs and lows. Excess return uses the Nifty 500 close over the same dates.

A target or invalidation touch is inferred from daily high-low data. When both levels are touched in one candle and intraday order cannot be known, the result is labelled Ambiguous. Published rows and completed outcomes are append-only: reruns can fill missing measurements but cannot rewrite or delete prior entries.

The Performance Laboratory groups completed outcomes by algorithm version, setup, publication-day market regime, sector, and market cap. It reports sample size, positive-return rate, median return, median excess return, median MFE and MAE, and an equal-weight publication-cohort drawdown diagnostic. Market regime uses only benchmark closes available on or before publication: close versus a trailing 50-session average plus trailing 20-session return. Samples below 10 are marked insufficient, 10-29 early, and 30 or more established; none of those labels proves future edge. The cohort path is not an executable portfolio and can contain overlapping fixed-horizon observations.

Corporate red flags and personal calendar

The Corporate Red-Flag Radar applies deterministic text rules to recent exchange or issuer documents and threshold rules to stored financial snapshots. It looks for auditor exits or qualifications, promoter-share encumbrance, credit-rating deterioration, fresh dilution, insider sales, regulatory action, receivables stress, and cash-conversion stress. Current promoter pledge is flagged from 5%; receivable days from 120; and a non-financial company's operating cash-flow deficit only when it is at least 5% of revenue. Increase and deterioration labels require an earlier stored snapshot.

Broad keyword matching only selects documents for inspection. Type-specific exclusions remove routine regulatory boilerplate, proceeds-monitoring reports, and subsidiary rights investments. Every displayed flag retains its source link. Reprocessing first retires previous derived matches, so corrected rules do not leave stale labels active. These are review prompts, not allegations or trade directions, and users must read the complete filing and later clarifications.

The personal calendar combines explicitly dated filing events with result-calendar and reported-result records for stocks in all saved portfolios and the watchlist. A filing-derived event is admitted only when its date is tied to a concrete phrase such as an AGM being held on a date, a dividend record or payment date, or an issue opening or closing date; generic filing dates are rejected. Reminder lead time and event types are user-owned settings. Reminders are in-app highlights, not guaranteed push, email, exchange, or broker notifications.

Trade plans and paper-portfolio measurement

Paper plans are user-owned simulations and never reach a broker. Position size uses the less favourable edge of the entry range, the selected account-risk percentage, and the distance to invalidation. Quantity is capped by both that risk budget and the stated account capital. The source snapshot, planned levels, quantity, and planned risk are retained after activation so later outcomes do not rewrite the original setup.

Completed-candle activation considers only sessions after plan creation. A range touch fills a long at the entry-range high and a short at the entry-range low; manual fills and exits use server time and cannot be backdated. Outcome evaluation begins with the following completed session. Adverse gaps through invalidation use the opening price, while favourable target gaps remain capped at the target. A candle containing both target and invalidation without a known sequence is labelled Ambiguous and receives no fabricated exit return.

Target hit, invalidation, manual exit, and maximum-session time exit are final states. MFE and MAE use completed daily highs and lows through exit. Live quotes may update the current open mark but do not alter those completed-candle excursion measures. Simulations exclude brokerage, taxes, partial fills, market impact, most slippage, and practical liquidity limits.

Nifty 500 momentum and market breadth

The momentum workspace uses an equal-weight Nifty 500 lens. Relative-strength ranks compare cross-sectional 1-month, 3-month, 6-month, and 1-year returns. Breadth measures advances and declines, participation above the 20, 50, and 200-day moving averages, and new 52-week closing highs and lows. Sector and industry rotation groups require enough eligible constituents before a signal is shown.

Trend stages use price, the 50-day average, a 30-week proxy, and the direction of those averages. Relative strength and breadth are descriptive measures of participation; they are not forecasts.

Quarterly-result classification

Quarterly results receive a 0-100 evidence score from year-on-year revenue and profit growth, operating-margin direction, quarter-on-quarter revenue momentum, and operating leverage. Scores of 75 or above are labelled Excellent, 55-74 Good, 30-54 Average, and below 30 Bad. Missing revenue or profit data caps the classification so incomplete evidence cannot earn the highest label.

Result quality is then shown beside the post-result market reaction. A strong result followed by a falling price, or a weak result followed by a rising price, is a research prompt rather than a conclusion.

Sources, management evidence, and trust

Price history and index context are sourced from market-data archives. Financial metrics, results, order announcements, and management claims are drawn from public company or exchange material and upstream data providers. Stock pages expose source labels and timestamps where available. When a cited claim matters, users should open the original filing and verify it directly.

Management trust analysis separates claims from outcomes. It records the source document, period, stated direction or target, and later evidence of delivery, delay, contradiction, or failure. Automated extraction can be wrong, so the source link remains the authoritative record.

Freshness and quality controls

Quotes, fundamentals, scores, tags, filings, and result records have independent refresh cycles. IndiaPulse displays their timestamps separately and flags stale or low-coverage rows. A current quote does not make old fundamentals current, and a recent score does not prove that every underlying source has updated.

Corporate actions, symbol changes, restatements, exceptional items, illiquidity, and parsing failures can distort ratios and charts. Missing evidence is retained as missing rather than guessed whenever the model supports that distinction.

Use and limitations

Models compress complex businesses into comparable fields and will miss context. Rankings can change after data refreshes or methodology revisions. Backtests can contain survivorship, look-ahead, data-quality, liquidity, cost, and slippage bias. No score, pattern, target, stop, or portfolio weight guarantees a return or limits a loss.

IndiaPulse is not registered with SEBI as an Investment Adviser or Research Analyst. Use the platform as a starting point for independent research and consult an appropriately registered professional for personalized advice.