MSN is running a piece this week declaring that the stock market is repeating a pattern seen only once since 1871, with history supposedly telling us what comes next. You can read it here. The year 1871 does a lot of work in that headline. It sounds ancient and authoritative, like the data itself is leaning across the desk to whisper in your ear.

Headline writers have been circling this same watering hole for weeks. I've seen pieces claim the Shiller CAPE ratio is near dot-com levels, claim that some valuation signal has fired only a handful of times in over a century and a half, and cite polling supposedly showing most Americans expect a crash. I can't vouch for any of those numbers, and neither, I suspect, can the people writing the headlines. Each one arrives dressed slightly differently, and each is selling the same underlying product: the idea that history has a script, and somebody has found it.

I'd gently point out a few things. A pattern that has appeared once or twice in 155 years is not a pattern in any statistical sense that should move your money. It's an anecdote with a long memory. If you flip a coin twice and it lands heads both times, you don't have a law of coin behavior. You have two flips. When an analyst says "the only other time this happened was 1929," what they're really saying is "I have one data point, and it was a famous one." One instance tells you nothing about the distribution. It only tells you which example was dramatic enough to remember.

The other problem is survivorship of the comparison itself. There were dozens of conditions in 1871, in 1929, in 1999, and today. If you let a writer pick any two of them to match, you can connect almost any market moment to almost any historical bogeyman. Match on valuations and you get 2000. Match on interest rates and you get 1987. Match on investor sentiment and you get whatever year fits the narrative this week. The matching rule does the predicting, not the history.

And notice what these headlines never contain: a date. "What comes next" is doing nothing in that sentence. No timeframe, no magnitude, no confidence level. If the market drops 30 percent sometime in the next five years, the piece gets to claim it was right. If it rallies 40 percent first and then drops 30 percent, it still gets to claim it was right. A forecast that can never be wrong is really just a horoscope with a chart attached.

None of this means valuations are necessarily comfortable, and I'd never tell you they are. In my own view, sober preparation is reasonable regardless: know your allocation, keep some cash or bonds if you'll need money soon, and don't let any single position grow into a bet. That advice was sound when valuations were boring, and it's sound now.

But the pattern-matching genre isn't analysis. It's content. It exists because fear is the most reliable click in finance, and 1871 sounds so much better than "nobody really knows." The sober version of the story is that historical studies have generally suggested high starting valuations tend to line up with lower returns over the following decade, while saying close to nothing useful about the next twelve months. Slow words. Bad headline.

So ask yourself a plain question. If this signal is real, what exactly would you do differently today that you shouldn't already be doing? For most people the answer is nothing, and that's the point. Keep buying, keep costs low, and let the once-in-155-years headlines do what they do, which is mostly generate other headlines.

The usual disclaimer: I'm sharing perspective built on forty years of my own mistakes, not professional financial advice.