Bears Slowly Coming Back to Life?

---------------------------------------------------------

Update no.1 on November 09th 2025

---------------------------------------------------------

Just a quick update. Things didn’t take a bullish turn this week. If you’ve been following our dashboard signals, you’ve probably noticed that none of them have really improved. Four days since my last update, we might now be approaching the moment when a hedge signal appears.

In just a few days, our option model — which was recently at a very high level — has dropped sharply. It now sits only a few digits above its threshold.

The way it crosses that threshold line often indicates how deep the drop will go. With the current slope, unless things reverse soon, this would likely mean a significant reset for that indicator.

Our implied correlation is also extremely close to triggering. Not only has its signal inverted, but it’s now just about to reach its internal threshold. As you may recall, this internal threshold is actually quite complex, involving multiple variables — and the line shown on our dashboard is only an approximation. That said, we’re very close to meeting the flipping conditions, as shown by its current proximity to that line.

If this occurs on Monday along with the option model, our Risk Index would rise to 4 — a level that, based on backtesting, historically implies about a 75% probability that things will get worse before they get better.

We’re also seeing a notable reduction in the distance between bullish and bearish bets in our NYSE and Nasdaq derivative volume indicators.

This suggests that conviction in those markets is shifting. Sentiment has turned much more bearish than just a few days ago, and given how the options market is currently pressuring equities, this could easily translate into a significant change in the market’s trend.

This indicator is not yet in the red, which was expected — it often lags slightly when a correction begins (though not always, as we saw in summer 2024).

That’s the current state of our Risk Index and its components. As you know, this indicator is much more sensitive than our hedge model and will flip during both small pullbacks and strong corrections. It’s designed to capture the reality that sometimes even mild declines at the index level correspond to much larger moves in risk assets.

This time, however, the Risk Index is rising together with one of the hedge model’s main components. At this pace, we could very well see the hedge trigger early next week — unless we get a rebound, which would be significant news.

Indeed, our market-breadth-related indicator has been climbing steadily toward its threshold over the last two weeks, not pausing even on days when the market printed green candles.

This means that an increasing number of tickers and sectors are underperforming — a setup that often leads to a snowball effect. As I’ve mentioned before, this indicator alone acts as a reliable hedge, showing roughly the same hit rate as the hedge algorithm itself. It never misses a strong correction. See for yourself:

Put simply, the other nine components of our hedge strategy are there to help us anticipate downturns faster — and perhaps even more importantly, to detect the end of a correction before the breadth-based indicator does, since that one tends to lag significantly on the exit. Having redundancy also strengthens conviction. For example, in February of this year, we reached a point where three internal hedge components had already raised a flag.

Path Forward

As mentioned above, most of our metrics degraded significantly last week, and we’ve reached a point where things look increasingly bearish. However, there are still a few rays of sunlight suggesting that the long-awaited full-blown correction could be delayed — and that we might instead be seeing the setup for a rebound, or a correction that remains just a blip on the index while already causing damage in riskier assets.

First, the Skew remains low by historical standards. It hasn’t risen during this correction, nor cooled down significantly, but it still sits at a relatively low level compared to recent months — roughly 45 points below where it stood before the tariff correction and about 15 points below the range where it was stuck for most of August, when the market began that month with several red candles.

VIX also didn’t invert and actually began to decline on Friday afternoon across all durations.

The move down may appear modest, but the difference between this and how sharply it was rising earlier on Friday is significant. It was clearly on track for an inversion, yet it has now started to cool down across all timeframes.

Historically, it’s also quite rare for our market-breadth-related indicator to fall from a yearly low (bullish) all the way to the hedge threshold in a single move, except during black swan events — which, at the moment, we’re not experiencing.

And finally, on Friday, we recorded our first Downtrend Exhaustion (DE) signal since April (formerly known as our Buy-the-Dip signals). It came from the DE1 model — the same one that flashed a strong red signal on April 7th, the day we saw the bottom of the tariff correction. At this point, it’s showing a blue signal of small amplitude.

As a quick reminder, DE1 produces three levels of signal strength, from weakest to strongest: blue, orange, and red. Red appears only every two to three years during major corrections. The others depend on the correction’s strength. Given the current realized volatility, which has remained well-contained so far, a blue signal is exactly what we’d expect — consistent with what we’ve seen in several smaller pullbacks:

We have three different Downtrend Exhaustion signals, each looking at different areas of the market. The DE1 signal identifies moments when a correction or multi-leg downtrend begins losing momentum, using a combination of price structure (across multiple timeframes), trading volume, drawdown depth, and realized volatility.

The reason we designed multiple DE signals is that each is intentionally restrictive — we prefer a model that misses a signal rather than one that triggers without cause. In this context, redundancy becomes essential.

When we talk about weaker signals, like a blue one, they don’t always capture the exact bottom of a correction but consistently flag moments when the market is near a potential bounce. A good example is the summer 2024 correction. Just before the B-wave bounce, we had a weak blue signal (value around 10). The market then rallied for about five days before starting its C-leg. On Friday, August 2nd, we had another blue signal — this time of much stronger amplitude (value around 60). Everything that day pointed to the end of the correction, and it likely would have been if not for the black swan event that followed on Sunday night — the yen carry trade unwinding, which triggered one of the largest VIX spikes of the decade.

Even so, the next day the market opened sharply lower but immediately rebounded, printing a candle that ultimately marked the bottom of that correction — accompanied by a strong orange signal from DE1.

You can learn more about our Downtrend Exhaustion signals here

Conclusion

In summary, the market has deteriorated considerably since my last update only a few days ago. We’ve now reached a point where both our Risk Index and our Hedge Signal are very close to their respective thresholds and could trigger at any time. These thresholds are designed as the optimal dividing lines — separating periods where markets tend to recover from those where things typically worsen. It’s not unusual to see prices reverse right at these levels. In other words, being positioned right at that door means we must be ready for either scenario.

At this point, the appearance of a blue signal in our Downtrend Exhaustion indicator — even if it’s a weak one — is a positive sign, suggesting that we might soon see at least a short-term bounce, which could delay the hedge signal. If that happens, we would simply continue to ride the market until it reverses and the hedge signal confirms.

Over the next few days, our main focus will be on the components of the Risk Index, the Market Breadth indicator, and the Downtrend Exhaustion Dashboard. What these signals do in the short term will determine our next steps. If we do get a hedge signal, we’ll follow it without hesitation — unless it coincides with a strong orange DE1 signal, which is unlikely but not impossible.

If this scenario unfolds, it would mark the conclusion of a very successful trade for the hedge signal, one where we would have captured roughly 80% of the move with leveraged exposure (UPRO and QQQ) and the remainder through non-leveraged positions.

We’ll keep you informed by updating this post soon. We’re at a crossroads — the path forward is uncertain, but clarity should come quickly.

---------------------------------------------------------

End of Update no.1

---------------------------------------------------------

The market didn’t have a good day today. After last week’s strong gap-up driven by impressive earnings from the major AI players, it was inevitable that earnings-related volatility would eventually swing in the opposite direction.

Today’s weakness seems to be the result of several overlapping factors: a wave of AI-related layoffs over the past few days, disappointing employment data from Indeed (since we still have no official data due to the government shutdown), concerns about tomorrow’s court hearing in the tariff case, and Palantir’s earnings yesterday, which reignited fears of overvaluation among the major AI names.

Price action wasn’t catastrophic at the index level — the S&P 500 (SPY) closed down a reasonable -1.19% — but the Nasdaq (QQQ) saw a steeper -2.07% decline, reflecting where most of the overvaluation anxiety is concentrated. A look at the S&P 500 heatmap confirms that tech was the main culprit, explaining why today’s gap between QQQ and SPY performance far exceeded their normal beta relationship. While tech-led risk assets sold off, more defensive sectors of the S&P 500 held up relatively well.

Despite the pullback, both major indices remain close to their all-time highs. Investors positioned in them are still wealthier today than they were two weeks ago — or at the end of almost any other month this year.

However, outside the major indices, things look more concerning. For investors with higher exposure to risk assets, today may have felt like the cherry on top of a very bad sundae rather than just a minor hiccup. Market breadth has been slowly but steadily deteriorating for the past three weeks.

On Friday, October 10th — after the market’s sharp -3% “tariff déjà vu” sell-off — I mentioned that I didn’t expect an immediate correction, even though many were calling for one and valuations could have justified it. The main reason was that our indicators were still sitting in overly bullish territory — too high for a straight “elevator down” move. Going down quickly and deeply is not the market’s normal regime — it’s a failure mode, and it requires an alignment of multiple factors. One key metric, our Market Breadth Indicator, had just printed a yearly low, meaning that this was when the largest number of companies were performing well simultaneously.

Sharp market crashes are driven by panic — and panic tends to occur when two conditions align:

1. Investors suddenly fear the collapse of the world or its economy (examples: COVID in 2020, the Fed’s rapid rate hikes in 2022).

2. There are no remaining “pockets of strength” — when every stock seems to be falling at once.

When that happens, everyone loses regardless of what’s in their portfolio, which fuels panic selling. Historically, that’s a sine qua non condition for deep corrections.

For instance, in 2022, while the market officially began its crash in January, our Market Breadth Indicator had already flipped to hedge mode by late November 2021. Similarly, during the tariff-related correction earlier this year, that signal reached its danger threshold at the end of February, when the S&P 500 was only -3.1% off its highs. In both cases, it foreshadowed the real damage.

In short, history shows that major “elevator-down” corrections typically only begin once that indicator hits dangerous territory. While we were far from that level on October 10th, the situation has changed — market breadth has now degraded significantly.

This deterioration is mirrored in our Implied Correlation Signal, which reflects a similar concept in the options market. This indicator measures the divergence between index volatility (S&P 500) and the volatility of its individual components. Normally, the two move together — but when fear rises in individual stocks, they begin to diverge. That’s exactly what we’re seeing now.

So, we need to acknowledge this de-risking phase. The implied correlation is very close to flipping red, which would raise our Risk Index to level 2. However, our market breadth indicator still sits comfortably above its hedge threshold. This isn’t a contradiction — implied correlation feeds into our Risk Index (which is designed to detect smaller risk shifts that affect high-beta assets outside the main indices), while market breadth is part of our Hedge System, which only triggers in the case of major corrections.

Outside of these warning signs, most of our other indicators remain relatively positive.

Our Option Model remains elevated and didn’t drop much today, even with the tech sell-off.

The spread between bullish and bearish options remains wide — still clearly favoring the bulls.

This tells us that sentiment in the options market remains far from panic and, in fact, still fairly optimistic.

Even though the Skew Index ticked up slightly today — signaling a modest rise in fear — it remains low by the standards of the last four years, having recently bounced off a local low of 139 late last week.

Wrap-Up

So, what should we make of all this? We’ve been quieter lately because we’re at one of those crossroads where certainty meets uncertainty.

Our certainty is clear: the market is overbought, valuations are stretched, and we are well overdue for a correction. That’s precisely why, at the end of the summer, we shifted our leveraged exposure back to a non-leveraged position. We know that, at some point not too far ahead, we’ll have the opportunity to re-enter the market with leverage — at a discount relative to today’s prices. That’s the part we’re confident about, and it’s what we’ve been consistently repeating lately.

The uncertainty, however, lies in timing. Every backtest we’ve ever run confirms that it’s nearly impossible to consistently call the exact top of a bullish leg. In a strong uptrend like this one, momentum can persist far longer than logic would suggest. What we do know is that, eventually, our hedge system will flash a clear warning — signaling that the elevator might be heading down. This may not be the top, but when it comes, we’ll act decisively — exiting positions, sending SMS alerts, and publishing an update. We’re just not there yet.

Today’s setup isn’t as bullish as it was back on October 10th, when our full suite of metrics gave us confidence that a deep correction wasn’t imminent — despite that day’s sharp sell-off. Now, our tone is more cautious, especially toward riskier assets.

If I had to speculate, I’m not convinced we’re heading straight into a full-blown correction just yet. Market breadth rarely falls from yearly highs directly to hedge-triggering levels in one move. It usually comes in waves — back and forth — that gradually bring it down. For nearly a month now, risk assets have already been in correction mode. The next move might briefly favor these beaten-down names. As valuations among large caps become increasingly unattractive (if not outright repulsive), investors might start rotating into smaller, discounted plays.

Another potential short-term catalyst could come from the Supreme Court, which begins hearings tomorrow on the so-called “Liberation Day” tariff case. While this likely won’t end the saga, even a temporary reprieve could be viewed positively for smaller companies.

Still, all this remains speculation. The reality is that, despite today’s large red candle on the Nasdaq — and to a lesser extent on the S&P 500 — the options market and most of our other indicators aren’t yet positioned for a major breakdown.

When that changes, you’ll know — we’ll keep you informed. Until then, we’re comfortable maintaining our current moderate-risk positioning, including on Bitcoin, which we believe is finally undergoing the correction it needed. With Bitcoin now down nearly 20%, this move looks much more in line with what we typically see at this stage of a bull cycle.

A drop toward the 95K level would still fit within the range of a normal pullback — consistent with late-stage bull market behavior. We actually think we’re now closer to a major bottom than to the beginning of something larger. We’ll share a dedicated update soon, once the latest on-chain data becomes available, as today’s readings have not yet been fully released.

Varia

While the market has been relatively calm lately, that’s given us some time to focus on a project we’ve been quietly working on for a while — bringing a real AI inside the WealthUmbrella Dashboard. In the not-too-distant future, I’ll write a full update about it, but considering that I’m a professor who actually does research in AI, it was about time WU had this:

(it's just our internal code name).

Why bring AI into WU?

1. Because investors are already using AI, but there’s also a slightly scary side to it

   I recently read that more than 25% of younger investors now rely primarily on AI for their investment decisions. The problem is that when you ask most public AIs about a company, they tend to regurgitate incomplete facts, often based on outdated data from their training period — and they rarely offer real insight. Even worse, they sometimes hallucinate with absolute confidence, which makes it hard to tell what’s true and what’s made up. Just try the new Yahoo Finance AI Analyst and you’ll see what I mean — it’s often a random list of loosely related facts. Perplexity’s finance module is better, but still far from what I knew we could build. I trained my first neural network back in 2003, and I knew we could do something far more robust.

   
   

   So, we’ve been secretly developing a multi-model AI trained on institutional-grade data — including our own proprietary datasets — and supervised by a “committee” that prevents it from fabricating random facts. The goal: provide deep, reliable insights with the kind of analytical quality you’ve come to expect.

   
   

2. Because our members fall into two very distinct groups.

   Some of you — especially our long-term and tech-savvy members (thank you for being with us all these years) — can’t get enough data. The more we add, the more you dig in. I’ve even seen some of you build genius personal strategies using our metrics.

   
   

   On the other side, we have members who find the amount of data overwhelming — especially newer investors still learning how to interpret our indicators. And I get that. The learning curve can be steep. Earlier this year, I mentioned that we were planning better onboarding videos to help with that, but now we believe AI might actually be the better solution.

   
   

   What we’re building is a conversational AI that you’ll be able to poke (no chatBot) directly within the platform to get a personalized interpretation of our data — one that reflects how we would read it.

   
   

   Our first step is to integrate this into the Stock Health Dashboard. That’s partly why we recently added the “Valuation in Context” module — to give the AI not only the current data but also the full historical context it needs to interpret it accurately.

   
   

   For example, with a simple click, you might soon be able to ask for our AI’s interpretation of Fiserv’s technical trend — and it would respond with something like this:

   

   (That’s actually a test we ran last week to see how quickly our AI committee could assess Fiserv’s sharp downtrend. )

Or ask what we think of Apple current valuation:

Step 1 will be AI integration into our Stock Health section, since that’s an easier environment — most of the metrics are already well understood by the system. It’s also a pivotal stage in the investment process: evaluating whether a company is worth investing in.

Step 2 will connect Pand.AI to our S&P 500 market signals, allowing it to generate a proper market interpretation. My ultimate goal is that you’ll be able to ask for — and receive — a complete market update whenever you want, even if I haven’t written one that day.

(Did you really think a Robotics and AI professor wouldn’t automate himself? Just kidding — it’s not replacing me, just adding to what I have to say.)

We expect you’ll start seeing Pand.AI appear in the Dashboard around the end of November. Once it’s ready, we’ll of course post a detailed announcement.

Varia 2

I recently passed some tests that confirmed what I had anticipated — my health is almost fully back to 100% (50 EV for our cardiologist member that told me I would be back to 55% in no time). The improvement has been significant enough that I feel completely normal again and can even run at a decent pace.

This closes a rather difficult chapter of my life — one I would have preferred not to experience, but which ended up being philosophically insightful.

I’m mentioning it here for a few reasons:

1. Many of you have written to ask how I’m doing.

   I’ve received a lot of kind emails from members checking in, but since all this happened, my inbox has gotten a bit out of control. So for those who’ve asked — here’s the update.

2. To say thank you again.

   When I first shared what happened to me on July 28th, I received so many thoughtful messages and personal stories. Those words genuinely lifted me up at a time when I wasn’t feeling my best. People who know me would say I’m usually in a good mood — that period was an exception, and your encouragement helped more than you might think.

3. And finally, to close the loop.

   When I first shared what had happened, it was because my doctors warned me that I might struggle during treatment for heart failure, and I was worried I’d have to step back from some of my projects. Fortunately, that phase turned out to be only temporary — and I actually managed to keep working almost as much as usual.

So, now that I’m fully back to normal, I don’t expect my health to interfere with work anymore — and I’m just grateful to be back at full speed.