Mag7 stocks: Estimated chance of crash by end of year
A little bit of Monte Carlo modeling for all
We had Claude Code whip us up a portfolio of equal dollar weighted Mag7 stocks:
Prices used (approx. Nov 3–4 2026 close):
┌────────┬─────────┬────────┬────────────────┬────────┐
│ Symbol │ Price │ Shares │ Position Value │ Weight │
├────────┼─────────┼────────┼────────────────┼────────┤
│ AAPL │ $331.90 │ 121 │ $40,159.90 │ 14.34% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ AMZN │ $250.69 │ 160 │ $40,110.40 │ 14.32% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ MSFT │ $520.99 │ 77 │ $40,116.23 │ 14.32% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ NVDA │ $206.69 │ 194 │ $40,097.86 │ 14.31% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ GOOG │ $278.27 │ 144 │ $40,070.88 │ 14.30% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ TSLA │ $398.50 │ 100 │ $39,850.00 │ 14.22% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ META │ $621.03 │ 64 │ $39,745.92 │ 14.19% │
├────────┼─────────┼────────┼────────────────┼────────┤
│ Total │ │ │ $280,151.19 │ 100% │
└────────┴─────────┴────────┴────────────────┴────────┘Close enough.
Then we do a backtest of this portfolio with our monte carlo portfolio modeler MCarloRisk. About 40 trading days till end of year (ooh, that’s scary!)
Since our app considers historical correlation among all assets in the portfolio, it assumes that these correlations will stay similar. And there’s a lot of correlation among these seven, noted in our prior article:
https://medium.com/@nttp/the-mag7-is-really-40-mag1-6-others-7401b5f58b88
A little report-out of metrics. We like that 5th percentile comes in at 4.4%. The top and bottom 99% and 1% is at the end of the respective “tentatively accept” ranges indicates that the model may be over-forecasting volatility over the last backtest range, but we are okay with that over-cautiousness for this crude of a model.
Now a forecast: 5% chance of this portfolio dropping about 12.5% at end of year.
And another forecast:
Get NTTP’s stories in your inbox
Join Medium for free to get updates from this writer.
1% chance of this portfolio dropping about… let’s call it 18% by end of year.
Now let’s add some Black Swan days!
What do you want them to be? You can put in your own!
We can set them up in the Tune window controls
So we set the up as 1% occurrence, 10% +-1% for this example. We don’t know if this has any historical precedent, we are just putting in some placeholder numbers to start. Kind of a lot of Black Swans, really. Don’t forget to flip the Black Swan switch on.
From this we can see a 5% chance of dropping over 19% at end of year.
And a 1% chance of this portfolio of the Mag7 dropping over 27%. See below.
And remarkably, the bright green line dips only slightly below neutral even when these large swans are modeled asymmetrically (e.g. only down side swans, not White Swans I guess we could call them, on the bullish side). The bulk of the non-black swan days stabilizes the forecast for the median.
What should you put in for Black Swan numbers?
Maybe do a little research on the worst stock market crashes in the past and see? We haven’t figured that out yet vis-a-vis our way of entering Black Swan numbers. Claude Code is giving us some enticing ideas, but it’s too early to report out. Stay tuned!
This work was done using the portfolio section of our MCarloRisk3D app for macOS and can be done also by the MCarloRisk iPhone app for back-of-the-envelope monte carlo model building.
