Event-driven algorithms use time-stamped corporate events—earnings, dividends, and material news—to predict short-term price moves. This guide explains the signals, data plumbing, execution, and risk controls Indian traders need, with formulas, backtest traps, and real-world examples from NSE/BSE.
Why event-driven trading matters
Markets reprice quickly around new information. In India, Regulation 30 of SEBI LODR requires timely disclosure of material events (results, board decisions on dividends, etc.), creating predictable data bursts that algorithms can harvest—if your timestamps, expectations, and execution are right. (Securities and Exchange Board of India)
What qualifies as an “event” in India?
- Earnings: Board-approved quarterly/annual results and commentary.
- Dividends & corporate actions: Interim/final dividends, splits/bonuses, and the ex-date mechanics impacting cash and derivatives. For extraordinary dividends (>2% of price), F&O strikes/base prices are adjusted per exchange rules. (NSE Clearing)
- Material news: Anything disclosed under SEBI LODR Reg. 30—M&A, management changes, regulatory actions, capex, etc. (Securities and Exchange Board of India)
Market microstructure anchor: India’s pre-open session runs 9:00–9:15, with order collection and random closure 7–8 minutes; regular trading begins 9:15. If you trade gaps or auctions, your algos must align to this clock. (NSE India)
Core signals & simple formulas
1) Earnings surprise & post-earnings drift
- Surprise: EPS Surprise = (Actual EPS − Consensus EPS) / |Consensus EPS|
- Where no consensus is available, use proxy expectations (analyst revision momentum, option-implied skew, or historical seasonal models).
- Drift (PEAD): Prices often continue in the surprise’s direction for days to weeks; evidence is well-documented globally and still studied in the AI era. (CFA Institute Research and Policy Center, CFA Institute Daily Browse)
Abnormal return math:
AR_i,t = R_i,t − (α + β·R_m,t); CAR(t1,t2) = Σ_t AR_i,t (event window).
Classify winners/losers by standardized surprise (z-score) and go long winners / short losers post-announcement with liquidity and risk filters.
2) Dividend capture via futures basis
Indian equity futures reflect cost-of-carry minus expected dividends. A quick diagnostic:
Theoretical F ≈ S · (1 + r·T) − PV(dividends)
or in practice from market quotes: Implied Dividend ≈ S · (1 + r·T) − F
Mispricings appear when futures under-reflect upcoming dividends. BUT: if a dividend is extraordinary (>2%), option strikes/futures base are adjusted on ex-date—your models must switch regimes accordingly. (Zerodha, NSE Clearing)
3) News/NLP + timestamped disclosures
- Parse exchange dissemination (NSE/BSE announcements) and company press releases tagged under Reg. 30.
- Start simple: keyword & entity rules with polarity; progress to domain-tuned transformers.
- Control for pre-open vs continuous session: a 7:05 pm disclosure vs 8:55 am changes your execution path. (Securities and Exchange Board of India)
Data pipeline (practical)
- Calendars & disclosures
- Results & board meetings from exchange “Corporate Filings/Announcements.” (NSE India, BSE India)
- Reg. 30 text + Industry Standards Note (2025) to keep classification consistent across issuers. (NSE India Search Archives)
- Event normalization
- Standardize fields: symbol, isin, event_type, board_time, dissemination_time, ex_date, record_date, dividend_per_share, is_extraordinary.
- Price alignment
- Use tick or 1-min bars; mark auction prints separately; snap to exchange timestamps (not vendor lag).
- F&O adjustments
- Ingest NSE corporate action circulars; apply strike/base price transforms only when flagged extraordinary. Log the adjustment factor and reconcile P&L continuity. (NSE India Search Archives)
Strategy modules (ready-to-code blueprints)
A) Post-Earnings Drift (PED-10)
- Universe: NIFTY100 with L2 depth ≥ threshold; exclude F&O ban days.
- Signal: SUE_z (z-score of EPS surprise or proxy).
- Entry: At 9:20 next trading day if |SUE_z| ≥ 1.5 and gap < 3%.
- Holding: 10 sessions; early exit if CAR < −2% from entry.
- Risk: Volatility-scaled sizing; max 4 names/sector; slippage adders around 9:15–9:25.
- Rationale: Monetize behavioral under-reaction. (CFA Institute Daily Browse)
B) Dividend Basis Arbitrage (DIV-BAS)
- Universe: Liquid single-stock futures with declared dividends.
- Signal: ImpliedDividend − AnnouncedDividend > fee + buffer.
- Trade: Long cash, short future (or options overlay) T−3 to T+1 of ex-date; unwind post adjustment.
- Guardrail: If event tagged extraordinary, expect strike/base adjustment—switch to post-adjusted series. (NSE Clearing)
C) Reg. 30 News Momentum (R30-NLP)
- Feed: Live exchange announcements; classify positive/negative/material using a rules + ML hybrid.
- Entry: IOC/limit in pre-open for after-hours disclosures; or immediate in-session with impact score ≥ threshold.
- Exit: Intraday at VWAP bands; hard stop at 2× 1-min ATR.
- Compliance: Ensure sourcing from official dissemination and maintain an audit log. (Securities and Exchange Board of India)
Backtesting pitfalls (India-specific)
- Look-ahead on timestamps: Use dissemination time, not news article time.
- Corporate action continuity: Stitch pre/post-adjustment prices correctly for extraordinary dividends, splits/bonuses. (NSE India)
- Auction mechanics: Pre-open random close (7–8th minute) changes fill probabilities; simulate indicative equilibrium price logic. (NSE India)
- Carry math: Ensure futures fair value uses local rates and dividend PV; reconcile with observed basis. (Zerodha)
Execution & risk overlays
- When to trade
- Pre-open for after-hours earnings or Reg. 30 disclosures; 9:15–9:30 if in-session & spreads wide. (NSE India)
- Order styles
- Use limit-with-protection, post-only in auction, or TWAP/VWAP for larger names post-open.
- Position limits
- Respect broker/exchange risk checks and MWPL/F&O ban constraints. (Check daily exchange lists.)
- Event halts & bands
- Consider price bands and circuit filters; widen slippage plugs around results and ex-dates.
- Portfolio-level
- Cap sector correlation, run intraday VaR, and apply kill-switches on multiple concurrent adverse events.
Quick reference table
| Event | Primary data source | Signal idea | Typical horizon | Key pitfalls |
| Earnings (results) | Exchange filings, company PR | SUE z-score, guidance tone | 1–10 days | Timestamp, liquidity shocks |
| Dividends | Board outcome, ex-date circular | Implied vs announced dividend | T−3 to T+1 | Extraordinary dividend adjustments in F&O |
| Material news (Reg. 30) | Exchange disclosure feed | NLP polarity × materiality | Intraday–3 days | False positives, rumor noise |
Sources include SEBI LODR Reg. 30 and NSE corporate filings/circulars. (Securities and Exchange Board of India, NSE India, NSE India Search Archives)
Indian examples (illustrative)
- IT majors (e.g., TCS/Infosys): Strong or weak deal-win commentary alongside EPS often drives multi-day drift—pair an earnings surprise filter with low pre-open gap entries. (CFA Institute Daily Browse)
- High dividend PSU names: Watch for extraordinary dividend flags; plan strike/base adjustments to avoid P&L discontinuities. (NSE Clearing)
- Banks: Reg. 30 disclosures on NPA trends, management changes, or regulatory communications can trigger swift repricing—route through auction logic if after hours. (Securities and Exchange Board of India)
Implementation checklist
- Data contracts: Subscribe to official exchange announcement feeds; archive PDFs + JSON. (BSE India)
- Calendar store: Persist event_id, symbol, t_dissemination, ex_date, is_extraordinary.
- Signal stack:
- Earnings: SUE, revision momentum, tone (MD&A).
- Dividends: ImpliedDividend vs declared; switch regime on extraordinary. (Zerodha, NSE Clearing)
- News: Reg. 30 classifier with confidence scores. (Securities and Exchange Board of India)
- Backtest hygiene: Corporate action stitching, auction fills, slippage curves by time-of-day. (NSE India)
- Execution: Prefer auction-aware engines for after-hours events; otherwise TWAP 9:20–10:00 with circuit-aware guards.
- Risk: Per-name exposure caps, portfolio VaR, and event heat throttles (limit concurrent event bets).
FAQs
Q1. Do Indian exchanges adjust option strikes for all dividends?
No. Under NSE rules, only extraordinary dividends (above 2% of market value) trigger F&O strike/base adjustments. Ordinary dividends do not. (NSE Clearing)
Q2. Is post-earnings drift still alive in 2025 with faster AI news-reading?
Research debates the decay, but drift remains a documented anomaly; profitability depends on costs, speed, and filters. (CFA Institute Daily Browse)
Q3. Where should I source “official” news for algos?
From exchange dissemination (NSE/BSE corporate announcements) and company filings under Reg. 30—these are timestamped and auditable. (BSE India, Securities and Exchange Board of India)
Q4. How do I model futures around an ex-dividend date?
Use cost-of-carry with dividend PV; compare implied dividend from S and F against declared amounts to find basis trades. (Zerodha)
Key takeaways
- Event-driven algos in India hinge on clean disclosure timestamps, corporate action logic, and auction-aware execution.
- For dividends, know when derivatives adjust (only on extraordinary payouts). (NSE Clearing)
- Earnings-surprise drift is tradable with tight risk and liquidity filters, but edge depends on implementation quality and costs. (CFA Institute Daily Browse)