Quantarded Weekly Signals #008 — Week 7, 2026
Market note: U.S. markets are closed today for Presidents Day (February 16).
This week closed positive, but again with meaningful intraweek dispersion.
Early weakness was followed by a sharp mid-week recovery and a partial fade into Friday. The final print was constructive, yet the path highlighted how concentrated baskets behave under volatility.
For context, this week Quantarded processed 454 House trade disclosures filed during the week, 227,217 Reddit comments analyzed and 32,914 stock ticker mentions detected and classified. As always, what matters is not volume alone, but how agreement evolves as volume grows.
Reddit picks
As a reminder, a ticker is labeled BUY or SELL only when it clears a minimum imbalance threshold. High visibility alone is not sufficient; divided sentiment is explicitly penalized.
This week’s basket is anchored by one dominant BUY, with the rest of the names clearing the threshold with meaningfully lower conviction. In practice, that means the basket is top-heavy: it can work, but it is more path-dependent than weeks where agreement is broad.
$MU: BUY, 34% share
$MU is the clear anchor this week, with the highest confidence in the basket and the strongest net imbalance.
This is not just “high volume”, it is persistent directional weight. When a single name pulls this far ahead of the rest, it usually reflects consistent agreement rather than a short-lived comment cascade.
As contextual background only, Micron has remained in the news cycle around supply expansion and AI-driven memory demand, including its announced Singapore investment plan: Micron plans $24-billion memory chipmaking plant in Singapore.
$SNDK: BUY, 23% share
$SNDK clears the BUY threshold with moderate conviction: it is a valid signal, but structurally much weaker than $MU.
In the extended report, this looks more like concentrated enthusiasm than broad participation. These signals can work, but they tend to be more fragile if the week turns risk-off.
As contextual background only, SanDisk has been covered recently around guidance and storage demand tied to AI-related themes: SanDisk forecasts profit surge, secures supply deal as AI fuels storage demand
$COIN: SELL, 16% share
$COIN is the only SELL in the basket, with lower conviction than the main BUYs.
This is a threshold-clearing SELL, not a dominant one. The model is picking up a negative imbalance, but the signal is not “overwhelming consensus”. That matters, because crypto-linked names can flip quickly when the broader tape turns.
As contextual background only, Coinbase was in the headlines this week around results: Coinbase posts surprise loss on crypto trading slowdown.
$NBIS: BUY, 14% share
$NBIS enters as a low-conviction BUY: it clears the imbalance requirement, but sits firmly in the “fragile” part of the basket.
This is the kind of inclusion that adds breadth, but not stability. If $MU underperforms, names like this rarely compensate on their own.
As contextual background only, Nebius has been covered around capex and data-center expansion:
- AI cloud firm Nebius posts surge in capex on GPU, data center expenses
- Nebius to build 240MW data centre near Lille, France
$RIVN: BUY, 13% share
$RIVN completes the basket as another low-conviction BUY.
This is again a threshold signal. It can contribute diversification, but it does not carry the basket. In weeks like this, the performance profile is dominated by the top-weighted name(s), not the tail.
As contextual background only, Rivian had prominent coverage this week tied to forecasts and its R2 rollout narrative:
- Rivian sees 2026 delivery jump driven by rollout of smaller, more affordable R2 SUVs
- Rivian shares surge as upcoming affordable SUV powers EV delivery forecast
Note: several highly visible tickers appeared heavily discussed in the extended report but did not clear the imbalance threshold for inclusion. Visibility without conviction is not a signal.
House trades
House disclosures this week were concentrated around a single filer. All qualifying trades in the dataset were disclosed by Michael T. McCaul, spanning multiple tickers and both BUY and SELL directions:
- SELL $NTK (~$250k)
- BUY $FCN (~$100k)
- BUY $IFF (~$100k)
- SELL $DLTR (~$100k)
- SELL $UST1 (~$250k)
Key observations: no cross-filer clustering, no repeated accumulation in a single ticker within a short window, and activity appears diversified rather than consolidating.
Performance review
Last week’s results
Mid-week strength (particularly Wednesday) drove most of the gains. Friday’s pullback erased a significant portion of that performance,especially in the more volatile names.
Dispersion was clear: two semiconductor names delivered strong positive weeks, one mega-cap detracted materially, and the most volatile name swung aggressively but finished modestly positive.
This was not a smooth week. It was path-dependent and concentration-sensitive.
Portfolio tracking
Using the running history you provided, the portfolio now stands at:
- End of 26W6 return: +1.64%
- YTD (2026) return: +13.40%
- Cumulative return since inception (weighted): +24.94%
The trajectory remains positive, but volatility clustering is visible in the equity curve. Weeks with top-heavy construction continue to amplify dispersion.
Knowing the algorithm — introducing the benchmark (Nasdaq via $QQQ)
Up to now, Quantarded has mostly talked about itself in absolute terms: weekly returns, the running portfolio value, and the occasional note about fragility or dispersion.
Starting this issue, we are adding a benchmark line.
The reason is simple. A weekly signal system is only meaningful relative to what you could have earned by doing nothing clever. If the whole market is drifting up, almost anything can look smart for a few weeks. And if the tape is hostile, even good signals can get dragged around by the regime. A benchmark keeps us honest.
Nasdaq is the natural reference point for this project. The Reddit universe that Quantarded samples tends to over-index toward the same ecosystem of names: tech, semis, software, high beta growth, and whatever the current “AI-adjacent” narrative is. Even when the weekly basket includes non-tech tickers, its behavior often rhymes more with Nasdaq than with a broad-market index, because crowd attention still anchors the portfolio in that growth-and-volatility risk bucket. Nasdaq defines the Nasdaq-100 as the 100 largest non-financial companies listed on the Nasdaq exchange. Nasdaq-100 fact sheet
To make that benchmark investable, we use $QQQ as the proxy. The choice is intentionally boring. QQQ is a liquid ETF designed to track the Nasdaq-100, which makes it a reasonable “default baseline” for the risk bucket Quantarded tends to live in. Invesco overview
In the tracking sheet, the Nasdaq line is computed as a $10,000 notional compounded by the daily “NASDAQ Change” series, aligned to the same dates as the Quantarded curve. It is not trying to simulate real-world frictions. It is trying to answer the core benchmark question: did we beat the baseline over the same calendar? Investopedia on benchmarks
Here is where we stand as of 2026-02-16, starting from $10,000 on 2025-12-21:
- Quantarded portfolio: $12,494.16 (+24.94%)
- Nasdaq proxy (via $QQQ series): $9,754.78 (-2.45%)
That puts the spread at +27.39 percentage points.
This is a meaningful gap, but the benchmark is here to narrow the story, not inflate it. The disciplined interpretation is: so far, Quantarded has done much better than the passive Nasdaq-100 proxy during a window where the proxy itself is down.
What it does not prove is that this outperformance is stable, regime-proof, or guaranteed to persist. If the market flips into a sustained Nasdaq-led risk-on rally, the hurdle rate rises immediately. And if Quantarded keeps outperforming in that environment too, the benchmark will be the first place where that becomes obvious.
Disclaimer
This newsletter is not financial advice.
All content is provided for informational and educational purposes only. Markets involve risk, including loss of principal. Past performance does not guarantee future results. Always do your own research.
Links
Website: https://quantarded.com
Subscribe: https://quantarded.kit.com/products/monthly-subscription?step=checkout