APEX TQQQ Algo — Weekly Signal Report (May 31, 2026)

Current Allocation: 100% BIL (cash proxy). TQQQ position: 0%.

The APEX QQQ Trading Algo remains entirely in BIL. The risk-on gate is fully open and the machine learning model is in sole control of the allocation decision. The gate is a structural filter the algo uses to determine whether broad market conditions are healthy enough to even consider holding TQQQ. When it is closed, the algo sits in cash regardless of anything else. Right now it is open, meaning the market has cleared that bar — QQQ has been trading significantly above its long-term trend for eight consecutive weeks, well clear of the 200-day moving average that sat near $610 when the gate opened on April 24. The ML model is the reason money is sitting on the sideline.

It is worth addressing the intuition that a more gradual approach — allocating, say, 30% or 50% to TQQQ rather than sitting at 0% — might feel like a smarter middle ground. In a ten-year backtest, however, that intuition does not hold up. Incremental allocation versions of the strategy arrive at roughly the same cumulative gains as the binary all-or-nothing approach, but they carry significantly more risk along the way. The reason is structural: partial TQQQ exposure tends to be highest precisely during the periods when the model has the most doubt — which are also the periods most likely to precede drawdowns. A leveraged instrument like TQQQ punishes hesitation harshly. The binary approach, uncomfortable as it looks in weeks like this one, is what keeps the risk-adjusted profile intact over a full cycle.

The Near-Term / Long-Term Divergence

The algo runs two independent probability forecasts — a 3-month horizon and a 6-month horizon — blended together with heavy weight toward the nearer-term 3-month result. The most recent readings tell a striking story of disagreement between timeframes. The probability of an up move over the next three months came in at just 24.1% — emphatically bearish. The probability over the next six months, however, printed at 96.8% — about as bullish as the model can express. The blended joint probability works out to 0.350, well below what is needed to trigger a TQQQ buy. The near-term signal is dominating and saying: 'not yet'.

This divergence has been building over the past several weeks. The 6-month probability has been above 90% since mid-May, while the 3-month probability has been sliding — from 84% in late April down to 24% by month-end. The model sees a strong medium-term setup but believes the near-term path is bumpy.

What Is Influencing the Model?

The clearest near-term warning sign is coming from the options market. Investors are buying meaningfully more puts relative to calls compared to recent history — the put/call ratio is running more than 1.3 standard deviations above its 6-month average. Elevated hedging demand of this kind is historically associated with near-term caution, and it is the single most influential signal in the current read.

Trend momentum is strong but the rally has advanced into thinner air. QQQ was already 6.8% above its 200-day moving average when the gate opened in late April and has extended further since, with QQQ trading in the $714–$720 range by late May against that long-term average near $610. While a large positive spread confirms trend strength, at sufficiently stretched levels the model has historically associated this with elevated mean-reversion risk. The slope of the 200-day moving average itself — a measure of whether the underlying trend is accelerating or flattening is reinforcing this picture.

Smoothed momentum indicators across multiple timeframes are adding to the cautious near-term tone. The model assesses technical momentum across many different lookback periods to filter out single-period noise, and the current reading is leaning bearish in the near term. Capital flow, by contrast, is telling a more constructive story — and notably becomes the dominant influence in the 6-month model, which is consistent with the idea that money flow trends are slow-moving but persistent. That is a big part of why the 6-month outlook is so bullish even as the 3-month signal flashes caution.

Taken together, the picture is of a market that has run hard, where options participants are hedging more aggressively than usual, and with a medium-term foundation looks solid, but the model is not yet convinced the next three months will be a smooth ride.

Outlook

The algo will remain in BIL until the blended probability improves substantially. Given the 6-month probability is already near its ceiling, the critical variable is whether the near-term signal recovers — which the model will reassess at the next retrain. 

The macro backdrop this week supports the cautionary approach. Today brings ISM Manufacturing PMI, followed by ADP Employment Change and ISM Services PMI on Wednesday, and the week culminates with Friday's Nonfarm Payrolls report — a data-heavy stretch that will directly shape Fed expectations heading into the June 16–17 FOMC meeting. Layered on top of that is the looming risk of a resumption in the Iran conflict. Any re-escalation carries the potential for an abrupt spike in oil prices and a sharp risk-off move in equities. The market is also entering a traditionally risky pre-Q2 earnings period, where the 'sell in May and walk away' mentality looks ever more attractive.

Sitting in BIL while watching QQQ rip higher is psychologically one of the harder positions this strategy demands. The temptation to chase is real and understandable. But this is precisely the scenario the APEX algo was designed and backtested for — a market that looks compelling on the surface while the underlying signals quietly accumulate risk. History and a decade of backtesting have consistently shown that holding fast during these high-uncertainty, momentum-driven stretches produces better outcomes than increasing exposure to a 3x leveraged instrument at the wrong moment. The structural setup remains intact, the 6-month outlook is strongly bullish, and the algo will be ready to rotate aggressively back into TQQQ when the near-term picture clears.