Section 5

Winners' characteristics

Two competing hypotheses tested head to head — did winners keep winning, or did losers bounce? Plus the size and volatility profile of the top performers, presented together because reading either one alone is misleading.

Survivorship bias applies to every figure on this page

Winners are drawn from the 2026 Nifty 500. Stocks that failed and were removed are absent — which means the "winners" here are, if anything, understating how bad the worst outcomes can be, and the momentum/reversion test is run on a survivor set.

Momentum vs mean reversion, head to head

Each row compares what happened to the prior year's best quartile versus its worst quartile. Baseline for both "stays" and "jumps" is 25% under randomness.

Transition Prior BEST → next-year return Prior WORST → next-year return Prior best stays in top quartile Prior worst jumps to top quartile
Momentum rate vs reversion rate, by transition
Momentum = prior best quartile stays in the top quartile. Reversion = prior worst quartile jumps to the top quartile. Dashed baseline is 25%.
The exception that breaks the rule: FY20 → FY21

The strongest mean-reversion signal in the entire sample. In the post-COVID recovery, prior-year LOSERS returned +213.5% against +98.0% for prior-year winners, and 45.1% of the prior worst quartile jumped into the top quartile. This is the classic post-crash rebound: the names most oversold in March 2020 bounced hardest.

So the finding is not a clean rule. Momentum was the better default (30.1% vs 24.7%, ahead in 6 of 9 transitions, and notably it was about avoiding losers going into FY20 where the prior worst quartile fell −47.7%). But after a market-wide crash, expect violent reversion instead. Both regimes occurred inside this single ten-year window.

Small caps dominate the winners — but read this next to the volatility

These two facts belong together. Shown apart, either one is misleading.

Market-cap band of the top 30
Tercliles of estimated FY-start market cap vs the 33/33/33 universe.
Average across the ten years: small 63.0% · mid 24.7% · large 12.3%. Small caps ~1.9× over-represented in 9 of 10 years.
Volatility of winners vs the rest
Annualised daily volatility during the year.
Mean annualised vol: winners 47.7% vs rest 36.9%. Winners were more volatile in 9 of 10 years — FY20 (COVID) is the sole exception, where less volatile names won (41.6% vs 46.7%).
This is an order-statistic artefact, not an exploitable size premium

Small caps carry higher return volatility. In any ranking of extreme outcomes, high-dispersion names are over-represented by construction — that is what dispersion means. The two panels above say the same thing from different angles: winners are the high-dispersion names, and small caps are where high dispersion lives.

The decisive check is the other tail: the bottom-20 lists carry the same small-cap skew. The same property that makes small-cap winners common makes small-cap disasters common. Once liquidity and impact costs are included, this is a statement about volatility, not a tradable size edge. The FY20 exception is instructive — in the crash year, low-volatility defensives won.

FYSmallMidLargeUniverse

Prior-year decile → next-year outcome

Pooled across the decade. The relationship is U-shaped, not linear — both extremes beat the middle.

Mean next-year return by prior-year decile
Decile 1 = worst prior year, decile 10 = best.
Read this pooled table carefully

The +25% to +44% means are dominated by a handful of very strong market years (FY21 and FY24 in particular — the mean stock returned +139.9% in FY21 alone). This is not an expectation for any single year, and pooling ten very different years into one table hides enormous regime variation. The shape (both tails beating the middle) is the finding; the levels are not.

Year-over-year correlation

Positive = momentum, negative = mean reversion. Note these are single-year cross-sectional correlations with n≈320–450 stocks, so they are far better identified than the macro correlations on the previous page.

TransitionPearsonSpearmannReads as

Sector concentration of winners

FYMost-represented sector in top 30NamesShareHHISectors represented
Winners cluster, but never into one sector

The top 30 concentrated into roughly 10–11 of 20 sectors, never all 20, with an HHI of 0.096–0.209 against a 0.05 benchmark for 20 equal sectors — i.e. two to four times the market's own sector concentration. Capital Goods was the single biggest cluster in 5 of 10 years (the domestic capex / PSU / defence cycle). Sector selection mattered, but a single-sector bet would still have missed most of the top 30.