A good stock can be a bad addition. Here is how to measure what a new position would do to your correlation, concentration and crash exposure before you buy it, with real numbers.
A stock fits when adding it makes the whole portfolio better for the risk you take. That is a different question from “is this a good company?” The same stock can be a sensible addition to one portfolio and a doubled bet in another.
Four things change when you add a position: how the portfolio moves with its largest risks (correlation), how much depends on a few names (concentration), how far it could fall in a crash (crash exposure), and how much return you get per unit of risk (Sharpe ratio). You can measure each one before you buy.
Check its correlation with your largest holdings and with the fund that best describes your portfolio. A correlation near 1 means it mostly adds more of the same risk; near 0 or below means it brings something different. Exent’s data on monthly returns over the five years to Sep 30, 2026:
| Stock | Correlation with SPY | Correlation with QQQ | S&P 500 fund volatility with 5% added |
|---|---|---|---|
| S&P 500 fund alone (SPY) | 1.00 | 0.92 | 15.7% |
| AMD | 0.55 | 0.71 | 17.1% |
| Micron (MU) | 0.51 | 0.59 | 16.9% |
| NVIDIA (NVDA) | 0.69 | 0.79 | 16.7% |
| Caterpillar (CAT) | 0.65 | 0.55 | 16.2% |
| Microsoft (MSFT) | 0.63 | 0.63 | 15.8% |
| Walmart (WMT) | 0.43 | 0.32 | 15.4% |
| Coca-Cola (KO) | 0.31 | 0.07 | 15.2% |
| Johnson & Johnson (JNJ) | 0.19 | −0.01 | 15.1% |
| Exxon Mobil (XOM) | 0.13 | −0.13 | 15.1% |
The last column is the volatility of a portfolio that is 95% S&P 500 fund and 5% the stock, over the same five years. Notice that correlation alone does not decide it: AMD is less correlated with the S&P 500 than Microsoft, yet raised volatility far more, because AMD itself is much more volatile. Fit depends on both how a stock moves with you and how much it moves.
Also check the stock against your own holdings, not just the index. If you already hold a Nasdaq-100 fund, the QQQ column is the relevant one: AMD at 0.71 adds the same engine, while Exxon at −0.13 moved the other way. Our guide to portfolio correlation explains how to read these numbers.
Work out the new weights before you buy, and group holdings that move together. A new stock that is 90% correlated with an existing position is, for risk purposes, an increase in that position.
This is where most portfolios already are stretched. Across 144 anonymized portfolios uploaded to Exent (as of Oct 2026), the median largest holding was 25.9% of the portfolio and the median top five were 77.2%. Adding a sixth idea that clusters with the top five does not diversify; it concentrates further. See single-stock concentration limits and position sizing.
Replay the new stock through past bear markets and see what it adds to the portfolio’s loss. A quick proxy is a crash replay: each holding’s weight times how far it fell in a given crash, summed. (Crash liability is a related measure: each holding’s share of the portfolio’s loss in its worst 5% of months.) Here is an illustrative portfolio of 50% S&P 500 fund, 30% Nasdaq-100 fund and 20% NVIDIA, using each holding’s real peak-to-trough fall in the 2022 bear market from the same data:
| Illustrative portfolio | New stock’s 2022 fall | Replayed portfolio loss, 2022 |
|---|---|---|
| Current: 50% SPY, 30% QQQ, 20% NVDA | — | 35.4% |
| Add 5% AMD (others scaled to 95%) | −61.6% | 36.7% |
| Add 5% Exxon Mobil instead | −20.5% | 34.7% |
| Add 5% Johnson & Johnson instead | −13.9% | 34.4% |
Illustrative only. The replay adds up each holding’s own worst fall, and holdings did not all bottom on the same day, so it is a rough worst-case proxy, not a forecast. The point is the direction: AMD made an already tech-heavy portfolio more fragile in the scenario it was most exposed to, while the other two softened it. Run several crashes, not just one, because each hits different holdings; our stress test guide shows how.
A full portfolio analysis runs these checks across all your holdings at once: correlation clusters, concentration, crash replays and Sharpe, Sortino and Calmar ratios. Stock pages such as NVIDIA show correlations and crash drawdowns for individual tickers.
Compare its correlation of monthly returns with your largest holdings and with the index fund closest to your portfolio. Above about 0.7 it mostly adds more of the same risk; near zero or negative it brings something different.
Not automatically. A high-return stock that moves with everything you own can raise volatility and crash losses more than it raises return. Check correlation, volatility and crash history along with return.
It is the change in the whole portfolio’s risk and return when you add the position. In Exent’s data, a 5% AMD position raised an S&P 500 fund’s volatility from 15.7% to 17.1%, while 5% in Johnson & Johnson lowered it to 15.1%.
Size depends on your risk tolerance and what the position adds. Many investors cap any single idea at a level whose worst historical fall they could live with, and treat highly correlated stocks as one position.
No. Fit tells you what a stock would do to your risk. You still need a view on the business and its price. Fit is the step most investors skip, not a replacement for research.
Upload your holdings and Exent shows your correlation clusters, concentration and crash replays, so you can see what a new position would add.
Educational content, not investment advice. Examples are illustrative; past performance does not guarantee future results.