NAV Candle: A Reproducible Candle-Based Visualization Framework for Mutual Fund NAV Series

Mutual funds generally publish one net asset value (NAV) observation per scheme and trading day. Consequently, conventional intraday OHLC candlestick charts cannot be constructed from NAV data without introducing values that are not observed in the underlying series. At the same time, a conventional line chart represents the level of NAV but provides limited visual encoding of period-by-period direction, magnitude, volatility, and drawdown structure. This paper introduces NAV Candle, a reproducible candle-based visualization framework designed specifically for once-daily mutual fund NAV data. The framework defines a body-only daily representation in which the opening value is the previous observed NAV and the closing value is the current NAV, while weekly and monthly representations use observed NAV values to construct genuine period-level OHLC values. In addition, the framework defines a fixed taxonomy of five descriptive patterns—up streak, down streak, extreme move, V-reversal, and whipsaw—and specifies deterministic rules for their detection. An empirical study is conducted using historical Indian mutual fund NAV records obtained from AMFI via MF API across 992 qualifying Direct Plan Growth schemes, strictly enforcing an observation cutoff of 31 December 2025. Within-fund return-order Monte Carlo permutation tests (B = 10,000) on the pilot scheme and benchmark panel demonstrate statistically significant departures from temporal exchangeability for streaks and whipsaws, while extreme moves conform to surrogate volatility distributions. Non-parametric Kruskal–Wallis tests confirm significant cross-category variation across 27 SEBI categories (p < 10^-5 across all patterns), and multi-resolution repeated-measures Friedman tests (N = 525) confirm significant attenuation of pattern rates at weekly and monthly aggregations (p < 10^-6). Event-study forward return evaluations across the 602-scheme development cohort under prespecified fund-level event eligibility reveal systematic post-down-streak mean reversion (+147.6 bps at h = 20, p = 4.1 x 10^-32) and post-whipsaw negative drift (-83.4 bps at h = 20, p = 1.4 x 10^-24). In prospective out-of-sample evaluation across 510 replication schemes (2020–2025), pattern frequency rank orderings demonstrate remarkable stability (Spearman rho >= 0.67 for streaks, whipsaws, and extreme moves). The framework provides a faithful visual encoding of historical mutual fund dynamics without fabricating unobserved intraday prices, serving as a descriptive visualization and analytical tool rather than a predictive trading or investment-advisory system.