Quantarded Weekly Signals #001 — Week 52, 2025

Welcome to the second issue of Quantarded.

Last week was about setting expectations and explaining how to read these signals. This week is more interesting, because it forces us to do something harder: be honest when the data is not pointing clearly in one direction.

Part of that may simply be where we are in the calendar. Late December tends to be a structurally different period for markets. Participation drops, liquidity thins, portfolios are rebalanced or closed out, and attention becomes more fragmented. Historically, this is a time when short-lived moves are more common and conviction tends to fade faster.

That does not invalidate the signals, but it changes how they should be read: in periods like this, disagreement, mixed sentiment, and lower persistence are often the signal.

This week, Quantarded processed 20 House trade disclosures filed during the week, 146,996 Reddit comments analyzed and 20,680 stock ticker mentions detected and classified.

How to read this newsletter

As a reminder, each issue looks at three different layers.

1. Reddit sentiment, which captures where retail attention and conviction are clustering right now.

2. House trade disclosures, which help us understand whether positions are being built or reduced over time, not whether someone made a single trade.

3. Performance tracking, where we look at what happened after publication, without backfilling or reinterpretation.

None of these is a trading system on its own. The value is in how they reinforce or contradict each other.

Reddit picks — mixed signals this week

Week 52 is a good example of why Quantarded exists.

At first glance, there are several tickers with high activity and strong scores. But once you look at directionality and dispersion, the picture becomes more nuanced.

Some signals are constructive, others are clearly negative, and overall conviction is more fragmented than last week.

Here are the most relevant Reddit signals, with that context in mind.

$TSLA: SELL, high confidence

$TSLA stands out as the strongest directional signal of the week, but on the negative side. Discussion volume was high, and sentiment skewed decisively toward selling. This is not a subtle signal, but it is also not a unanimous one.

$META: SELL, low confidence

$META lands on the SELL side this week, though with lower conviction than $TSLA. Sentiment skewed negative, but discussion was more dispersed. This reads more like skepticism than outright rejection.

$AXTI: BUY, low confidence

$AXTI shows a clear BUY signal driven by concentrated positive sentiment rather than sheer volume. This is the kind of ticker that would be easy to miss in raw counts, but stands out once imbalance and consistency are taken into account.

$ONDS: BUY, low confidence

$ONDS shows a modest BUY signal. Confidence is lower, but sentiment imbalance is still sufficient to justify a directional call. This is a weaker signal, included for completeness rather than strength.

High volume, low conviction (worth watching)

• $SLV produced the highest composite score of the week. Mentions were frequent and sentiment skewed positive, but the imbalance was moderate rather than extreme. This reads as broad, constructive interest rather than a sharp consensus.

• $TSLA deserves a second mention here, precisely because of its volume. Even beyond the SELL signal, it generated one of the highest comment counts of the week. That level of activity often precedes continued volatility, even if direction is contested.

Taken together, Week 52 looks mixed. There are actionable-looking signals, but fewer cases where everything aligns cleanly.

If last week answered “what looks unusually strong,” this week answers a different question: where is attention fragmented, and where is sentiment actively disagreeing?

House trades — activity without conviction

House trade disclosures this week tell a much quieter story.

There were multiple filings, but they are dominated by small position changes, often well below the $100k threshold that Quantarded uses on the website and UI to flag more meaningful activity.

You will notice that many trades in the $1k–$15k range, a few repeated large allocations and limited evidence of sustained accumulation in a single name.

A few examples help illustrate the point.

• Rudy C. Yakym III (R) disclosed a BUY in $3-MONTH, MATURE for $15,001. This looks more like routine cash or short-duration positioning than a directional equity bet.

• Tim Moore (R) disclosed small BUY trades across $CBRL and $VZ, each around $15k. Even combined, these do not show strong consolidation in any single asset.

• David J. Taylor (R) disclosed both a SELL in $LLY and a BUY in $HD, each at roughly $1k. These are too small to infer anything meaningful on their own.

The important takeaway is not that “nothing happened,” but that nothing consolidated.

This is exactly the kind of week where House data is useful by what it does not show. There is activity, but no clear build-up of exposure that would justify a strong signal.

Performance review — how Week 51 actually played out

This is the first week where we can look back. Last week’s picks showed strong early momentum, followed by gradual fading as the week progressed.

At the portfolio level:

• End of week return: +2.34%
• Weighted end of week return: +2.38%

The pattern is worth calling out. Early momentum was strong, on day one these picks made +6.22%, but it did nto persist.

This reinforces something important about these signals: they are better at identifying attention and short-term positioning shifts than sustained trends (for now).

Knowing the algorithm — how the sausage is made

Recency and why timing matters

I want to start a small recurring section that pulls back the curtain a bit, not to dump internals, but to show that this is not hand-waving. This week’s behavior is a good excuse to explain one of the most important ideas in Quantarded: recency weighting.

Not all Reddit mentions are treated equally. A mention from five minutes ago does not carry the same informational value as one from three days ago. In fast-moving environments, information decays.

That is not a vibes-based claim. It is a standard approach in time-series analysis, usually modeled with exponential decay, the same family of ideas behind EWMA/EMA smoothing (Box et al., Time Series Analysis: Forecasting and Control).

In Quantarded, we model this decay using a half-life of 12 hours. Half-life means that an event that is 12 hours old contributes half as much as it did when it was created, 24 hours old contributes a quarter, and so on (Hyndman & Athanasopoulos, Forecasting: Principles and Practice).

Formally, if age_hours is the time since the event happened, the recency weight is:

Or, in exponential form (which is what we actually use in SQL):

Recency is applied before aggregation, not after. That means late-week “new information” can outweigh early-week noise, even if total volume looks similar. This is why signals often look strongest early and fade if attention does not persist.

This behavior is intentional. It makes the system sensitive to changes in attention, not just raw popularity.

That is the recency story in Quantarded: a simple decay model with an interpretable half-life, widely used in forecasting, signal processing, and finance (Oppenheim & Schafer, Signals and Systems).

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 and consider consulting a licensed financial professional.

Links

Website: https://quantarded.com

Subscribe: https://quantarded.kit.com/products/monthly-subscription?step=checkout