You Don't Have to Call the Top

The question every investor asks

Ask a room of investors what they want from a market analyst and the answer is almost always the same: tell me when to get out.

It is an understandable request, because most people have only one direction to worry about. Selling short — positioning to profit when prices fall — is a professional's tool. It belongs to active traders with the appetite and the capital for it, and to institutions hedging positions they already hold. The ordinary investor does not do it, and mostly should not. Someone putting money into an index fund for retirement buys, and holds, and that is the entire strategy.

For that investor, the danger is not the ordinary drop. A market that falls five or ten percent and recovers within months does no lasting damage to a thirty-year plan; by the time it is over it barely registers. What does the damage is the steep, drawn-out bear market — the decline of forty or fifty percent that takes years to climb back from, and takes a piece of the investor's remaining time along with it. That is the event actually worth defending against.

So the entire profession bends toward that one question. Analysts spend their careers trying to call the top, because that is what their audience is asking for. Berg's position, arrived at over a career that began in 1978, is that this is the wrong job to be attempting.

Why tops are so hard to call

The difficulty is structural, not a matter of effort or skill.

Market bottoms tend to be sharp. Prices fall, panic peaks, and the turn happens in a window of days — often a single day, sometimes with a violent reversal that is visible in the data almost immediately.

Tops behave differently. In Berg's description they are rolling affairs rather than sharp points. One index peaks in one month, another peaks the next, a third has already peaked and nobody noticed. There is rarely a single moment when the whole market turns over — there is a slow, uneven rotation that only resolves into a clear picture months later, once the damage makes it obvious.

That is why so many top calls arrive either far too early — the analyst who has been warning of a crash for three years and is eventually right, having missed an enormous advance — or far too late, after the decline is already well underway.

Berg is blunt about the scoreboard, including his own profession's:

“I've never found a technician who has been successful in calling market tops consistently.”

He is equally direct about the other half of the ledger, and this is the half that matters most. Bottoms, he argues, can be called consistently — and he says he has found exactly one analyst who does it from data, himself, on the basis that his work has modeled every major market low since 1957. His own summary of a career spent testing both ends:

“I realized that it's almost impossible to call a precise top. But we're very, very good at calling precise bottoms — or near-precise bottoms — within days of a market low.”

The asymmetry is the whole thesis. The two ends of a market cycle are not equally knowable, so a sensible method should not treat them as if they were. Give up on the end you cannot measure. Build everything on the end you can.

The rule that replaces the prediction

If you cannot reliably identify the top, what do you do instead? Berg's answer is to stop trying, and to react rather than predict:

“It's not necessary to anticipate a bear market. Wait till the market declines 4% or 5% or 6% or 8% or 9%, and get out then… You don't have to call the exact top.”

This is a smaller claim than it first appears, and that is its strength. It does not require knowing the future. It requires only a decision made in advance about how much of a decline you are willing to sit through before you step aside — and then actually doing it when the moment comes.

The cost is explicit and it is paid every single time: you always give up the first stretch of every decline, and on the occasions when the drop turns out to be an ordinary wobble, you have stepped aside for nothing.

What you buy for that cost is the other tail. You are out of the way of the declines that destroy a decade of compounding, because every one of those begins as an ordinary drop of a few percent and simply keeps going. And, crucially, you never have to be correct about which is which at the moment it starts.

Getting out is only half a strategy

Here is where most market-timing advice quietly falls apart, and it is the part worth reading twice.

An exit rule on its own is not a strategy — it is half of one, and the dangerous half. Anyone can devise a reason to sell. The hard question arrives the morning after: when do you buy back? Sit in cash waiting for the news to improve and you will wait through the entire recovery, because the news is at its worst at the bottom and only improves long after prices have. That is how investors who correctly avoid a crash still end up worse off than the ones who did nothing — they miss the bottom, and missing the bottom means missing the bull market that follows it.

So the exit rule is only as good as the re-entry that follows it. Step aside without a tested way back in and you have not reduced your risk; you have swapped the risk of a decline for the risk of missing the recovery, which for a long-term investor is just as expensive and far easier to rationalize.

This is precisely where Berg's work is pointed. His models are not built to tell you the market is falling — anyone can see that. They are built to identify the low, and to do it while it is still happening: readings that have historically clustered at or immediately after major bottoms, drawn from every major market low since 1957. In practice the signals arrive within days of the low, sometimes the day after. When one came eight days late in April 2026, he called that late.

Why the two halves need each other

Put the two together and the logic closes. You cannot call the top, so you do not try: you let the market take its first few percent from you and step aside on a rule. You can call the bottom, so that is where the precision goes: you return on a data-driven buy signal rather than on a feeling that things look better.

Precision at the bottom is what buys you permission to be imprecise at the top. Without it, the whole approach collapses into guesswork — and the honest version of that sentence is that an investor without reliable bottom signals is genuinely better off buying and holding than stepping aside and hoping.

That is the entire design of the MB Edge model, and the asymmetry is visible in how much machinery sits on each side. The sell side is deliberately plain: the model holds the S&P 500 and steps aside on a trend-following, pattern-based rule once a decline is underway. Note what that exit therefore does not claim — it is a signal to step aside, not a declaration that the top is in.

Then comes the part that makes the exit safe: it buys back on the first model signal that a low is in. For individual investors the model takes that first signal rather than waiting for confirmation, deliberately — for someone holding the index for retirement, being a few days early to a recovery costs far less than being months late to one.

Why the same person can be confident and cautious at once

This asymmetry explains something that otherwise looks like a contradiction in Berg's public commentary. He is often specific and confident about lows — willing to name a day, and to act on it — and noticeably more hedged about highs, offering conditions and warning signs rather than a date.

That is not inconsistency. It is a direct consequence of believing that one end of the cycle leaves clear evidence in the data and the other does not. An analyst who is equally confident at both ends is telling you something about their marketing, not about the market.

What it means for an ordinary investor

The practical translation is almost disappointingly simple.

Stop waiting to be told the top is in. Nobody is going to tell you, and the people who sound most certain are the ones to trust least. Decide instead — while you are calm, before anything is falling — how far you will let a decline run before you step out of the way.

Then solve the harder problem, the one almost nobody plans for: know in advance what will bring you back. Not a feeling that the worst is over, not a headline, not a round number — a tested signal. If you have no answer to that question, you are better off staying invested through the whole thing than stepping aside with no way home.

Both halves are hard to obey in the moment, which is exactly why a mechanical version exists. The MB Edge model makes the same decision the same way every time — out on a simple drawdown rule, back in on signals drawn from decades of modeling — without asking how anyone feels about it on the day. One line of code gets you out. The research is all on the way back in, which is the half worth paying for.

Watch the full conversation

This article draws on Milton Berg's August 2026 interview with Jack Farley on Monetary Matters — a wide-ranging hour and a half that also covers gold, bonds, and the Federal Reserve. The passage on calling tops begins around the twenty-three-minute mark.

Perguntas frequentes