Midcap Smallcap: breadth is not diversification

Pulled the factsheet on one of our higher-octane cards this weekend. A 40-stock momentum basket built from the smallest end of the market. Five-year backtest CAGR of 48.98 percent against the Nifty's 10.68. The kind of number that makes you stop scrolling.

So I scrolled one column further, to the drawdown. Minus 26.17 percent.

Here is what made me sit up. The concentrated 5-stock card on the same shelf fell less, minus 24.72. The 20-stock midcap card fell 24.87. The 40-stock basket dug the deepest hole of the three. You add 35 names and the floor drops, not rises.

That runs against everything we are taught about diversification. More holdings, less risk. Except it only works for one kind of risk.

Spreading across 40 names protects you when one of them blows up on a bad result. It does nothing when the whole mid-and-small universe sells off in the same two weeks. January and February 2025, the basket fell 10.9 then 14.1. All forty names fell together, because all forty were the same bet. Mid and small caps in a risk-off do not take turns.

The 50-stock Allcap card, drawn from the whole market instead of just the small end, actually fell less, 22.70. More names, shallower hole, because the names were genuinely different bets across the cap curve.

So the lesson I keep coming back to. Diversification is a count of bets, not a count of names. Forty tickers that rise and fall in the same week are one bet wearing forty jerseys. This card earns its place as a satellite you size small and hold through a 26 percent fall, not as the core you mistake it for because it holds forty lines.

Same Sharpe as the 5-stock card, 1.61. The breadth was a feeling, not a hedge.

The comparison on today's file

The opening uses the figures published on 30 June. On the current lineup file the order is unchanged. Midcap Smallcap, forty stocks: 42.64 percent CAGR, Sharpe 1.63, maximum drawdown 26.17 percent. The five stock Focused Alpha 5 fell 24.18 percent, the twenty stock Midcap 24.87, and the fifty stock Allcap 22.72 (Model, back tested, as of 27 Apr 2026). The forty stock card still dug the deepest hole.

On the monthly model values, the card fell 10.9 percent in Jan 2025 and 14.1 percent in Feb 2025. Allcap fell 12.1 and 11.4 in the same two months (Model, back tested nav files). Two cards built on the same momentum engine, both hit hard, and the one drawn from the smaller end of the market took the larger second blow.

Counting bets, not names

The opening's line, diversification is a count of bets, has a formula behind it. If every stock in an equal weight portfolio has the same volatility and the same average correlation with the others, the number of truly independent bets the portfolio holds is the number of stocks divided by one plus the number of stocks minus one, times the correlation.

Run the numbers. Forty names with an average correlation of 0.5 come to about 1.95 independent bets. Five names with a correlation of 0.3 come to about 2.27. Fifty names at 0.3 come to about 3.18. And no number of names at 0.5 can ever get past 2. Correlation sets a ceiling that adding names cannot break.

Those correlations are assumptions for the example; nobody publishes them for these cards. But the direction is well established: stocks from one narrow band of the market, sized similarly and owned by similar investors, tend to move together more than stocks drawn from the whole market, and more still in a sell off. That is how forty names can behave like two bets in a bad fortnight.

Two kinds of risk, two kinds of protection

That is why the fifty stock Allcap fell less. It is not the ten extra names. It is that its fifty come from the whole market, large companies included, so in a small cap sell off part of the book is fishing in calmer water. And it is why the multi asset cards on the shelf, which add debt and gold, fall less again: those sleeves are genuinely different bets.

Reading breadth on any portfolio

Two questions do more than a count of holdings. Where do the names come from, one band of the market or several? And what else is in the book that does not move with them? A forty stock fund of small companies and a forty stock fund across the market are not the same diversification, even though the holdings count matches.

1. Count namesThe number on the factsheet. The least useful of the four.
2. Map the bandsLarge, mid, small, micro: how many of the names sit in each.
3. Find the other assetsDebt, gold, anything that does not move with the stocks.
4. Count betsNames that fall together count as one.

The free Learn module on modern portfolio theory shows why correlation, not the count, decides how much risk diversification removes, with a quiz. The Sharpe and volatility calculator measures the swing of any return series. Related: the Allcap spotlight and 5 against 20 against 50 stocks.