What We Measure, How We Measure It, and What We Don’t Yet Claim
Every number on this page is tracked forward, in public, before we know the outcome. It’s easy to build a strategy that looks great against history. It’s much harder, and much more honest, to publish a call before you know how it turns out and let the result speak for itself. No backtested fiction. No cherry-picking. We show you what worked and what didn’t.
TOP20 — Our Flagship Systematic Portfolio
Live20 stocks selected by our core quality ranking, equally weighted, rebalanced every two weeks. TOP20 demonstrates how the ranking translates into real, disciplined portfolio decisions — tracked forward in time, not backtested.
Holdings
20
Weighting
Equal weight
Rebalance
Biweekly
Since
August 2026
1W Return
—
4W Return
—
Since Inception
—
Alpha vs SPY
—
Alpha vs MAG7
—
TOP20 launched in August 2026. Its track record is short, honestly, because it’s new — we’d rather tell you that plainly than pad this page with a longer-looking number that doesn’t reflect what you’d actually have experienced as a subscriber. All results are tracked forward in time — not retroactively fitted.
TOP20 vs SPY vs MAG7 (since inception)
Equity curve chart will be displayed here as data accumulates.
Performance assumes execution at next market open after rebalance signal. No transaction costs included. Past performance does not guarantee future results. The model portfolio is not investment advice.
What Makes BambooSignal Different
Most services stop at a list. BambooSignal goes further.
A portfolio, not just a ranking
Rankings are translated into a real, rules-based portfolio with defined holdings, weights, and rebalance logic. This is how institutional investors think — not as isolated trades, but as a system.
Everything tracked forward in time
Every ranking and every portfolio decision is recorded at the time it is made — not retroactively fitted to favorable outcomes. Forward tracking is the only honest way to measure edge.
No cherry-picking or backtest bias
The examples and evidence on this page include both strong outcomes and modest ones. The edge is statistical, not anecdotal. That distinction matters.
How TOP20 Is Built
TOP20 is deliberately simple to explain — the discipline is in the rules, not in complexity.
Pure quality ranking
Holdings are selected purely by the daily five-pillar research ranking across 6,600+ equities — the same process described in our methodology. No discretionary overrides.
Rebalanced every two weeks
TOP20 is re-evaluated on a fixed biweekly schedule, not daily. This is a deliberate turnover control — rankings are inputs to the process, not trade orders to act on every time they shift.
An exit buffer that prevents unnecessary turnover
A holding isn't sold the moment it slips a few ranks. It's retained as long as it stays within a meaningful buffer, and only exits once it falls further. This prevents selling a position simply because it slipped from rank 18 to rank 22.
A sector cap, and no forced backfilling
New entries are capped per sector to prevent concentration. And if fewer than 20 stocks genuinely qualify on a given rebalance, TOP20 runs under-filled rather than lower its bar to hit a round number.
From portfolio to signal-level proof
Signal Edge — Insider Sell Detection
The portfolio above is built from signals. But do those signals actually have edge? The data says yes. The standout: insider sell detection across 1,034 Form 4 signals.
56.6%
Directional accuracy at 5 days
59.7%
Directional accuracy at 10 days
p < 0.0001
Statistical significance
914 signals evaluated at the 10-day horizon. z-scores: 4.2 (5d), 5.9 (10d). These results suggest the signal captures genuine information asymmetry — not noise.
Not All Signals Are Equal
Intellectual honesty requires distinguishing between what has been statistically validated, what shows high conviction based on observed patterns, and what depends on context.
Insider Sell Detection
Statistically validated across 1,000+ observations with p < 0.0001. The signal demonstrates measurable directional edge at both 5-day and 10-day horizons, improving with time — suggesting it captures real information asymmetry rather than short-term noise.
Cluster Insider Buying / Strong Momentum + Multi-Factor Convergence
Individual examples show strong follow-through (PARR cluster buy: +25.7% at 10d). Sample sizes are smaller than Tier 1, so the edge is observed but not yet statistically confirmed at the same level of rigor. These signals draw on well-documented academic factors (insider conviction, momentum persistence).
Emerging Leaders / Strong Momentum Stocks
Early-stage detection of institutional accumulation and momentum signals. Performance depends on market regime and entry timing. Individual examples (MU, CAVA) show strong results, but these signals require user judgment for position management.
Signal Examples — Real Data, No Cherry-Picking
The following examples represent actual signals generated by BambooSignal across three signal pillars: insider activity, momentum, and institutional accumulation. They include both strong outcomes and modest ones — because that is what honest reporting looks like.
GOOGL
InsiderInsider Sell Warning
Mega-cap. Insider sell signal preceded a multi-day decline in one of the most liquid names in the market.
RCL
InsiderInsider Sell (founder family)
Founder-family selling detected before a sustained move lower. The signal preceded broad market awareness by days.
MTSI
InsiderInsider Sell (corporate)
Corporate insider selling preceded a sharp decline that accelerated over the following two weeks.
PARR
InsiderInsider Cluster Buy (5 insiders)
Five insiders purchased within a short window. The strongest cluster buy signal in the dataset, with exceptional follow-through.
WAB
InsiderInsider Buy (executive)
Executive open-market purchase. Modest but consistent positive returns aligned with the insider conviction signal.
FIG
InsiderInsider Buy (executive)
Executive buying detected early. Returns accelerated from 5-day to 10-day, suggesting the signal captured a developing trend.
MU
MomentumMomentum / Emerging Leader
Identified via momentum and emerging leader signals. Strong short-term follow-through across multiple observation windows.
CAVA
MomentumMomentum / Emerging Leader
Momentum signal detected early-stage trend acceleration. The stock continued to lead its sector after identification.
Returns are measured from signal date. Alpha is calculated vs. benchmark (SPY) over the same period. Past performance does not guarantee future results.
How This Translates to Real Returns
Edge does not compound through a single trade. It compounds across many signals, applied consistently over time. BambooSignal’s value operates at three levels:
Avoiding losses
Insider sell signals with 59.7% accuracy at 10 days help identify names where informed participants are reducing exposure. Avoiding or de-risking these positions — even occasionally — protects capital that would otherwise erode returns.
Improving decision quality
When multiple signal pillars converge — insider conviction, momentum confirmation, and fundamental quality — the probability of a favorable outcome increases. The system does not replace judgment; it improves the inputs to your judgment.
Capturing asymmetric moves
Cluster insider buying (like PARR's five-insider convergence at +25.7%) represents the kind of asymmetric opportunity that can meaningfully impact portfolio returns. These setups are rare, which is precisely why they are valuable — and why systematic detection matters.
No guarantees. No promises of specific returns. The advantage is probabilistic and portfolio-level — exactly how institutional investors think about edge.
Important Considerations
Transparency builds trust. Here is what you should know:
Not every signal works. Individual signals will produce false positives. The edge is statistical — it appears across many observations, not in any single trade.
Market conditions matter. Signal performance can vary across market regimes. A sell signal during a broad market panic has different context than one during a calm uptrend.
The observation window is still limited. While 1,034 sell signals provide statistical confidence (p < 0.0001), longer observation periods across more market cycles will strengthen or refine these findings.
The model portfolio has a short live track record. Portfolio-level performance will become more meaningful as weeks and months of forward data accumulate.
Buy signals have smaller sample sizes. Cluster insider buying and momentum signals show promising individual results, but have not yet accumulated the sample depth needed for the same level of statistical validation as sell detection.
The bottom line: Despite these caveats, the insider sell detection signal remains statistically significant at p < 0.0001 across 1,000+ observations with improving accuracy over longer horizons. The model portfolio translates these signals into real holdings, tracked forward in time. This is a real, measurable edge — not a backtest artifact.
What This Means for You
Most investors encounter insider selling data days or weeks after the filing — if they encounter it at all. By that point, the move has often already happened. The same is true for momentum shifts and accumulation patterns: by the time they appear in headlines or screeners, the highest-probability entry window has passed.
BambooSignal’s value is not prediction. It is timing. The daily report surfaces these signals while they are still forming, which connects directly to real portfolio decisions:
Avoiding bad entries — seeing insider sell signals before committing capital to a deteriorating name
Increasing conviction — when insider buying, momentum, and quality converge on a stock you are already researching
Acting earlier than consensus — positioning before the signal becomes widely recognized and priced in
The question is not whether these signals exist. The data above confirms they do. The question is whether you see them in time to act on them.
See the Evidence as It Forms
Access the daily report and see the process as it runs — before opportunities become consensus. The edge is in timing, not hindsight.