Our Method
Most trading tools are built on an assumption: more data leads to better decisions. Give investors real-time dashboards, thousands of data points, configurable filters, and let them figure it out.
Decades of peer-reviewed research says that assumption is wrong.
We built Pinpoint Alpha on the opposite thesis. Here's the evidence behind our approach.
There's a critical difference between confidence and accuracy, and more information widens the gap between them.
Researchers at the University of Chicago ran three experiments testing what happens when decision-makers receive increasing amounts of data. The result: confidence rose significantly with each additional piece of information, but accuracy barely moved. People don't adjust for the fact that their brains can't actually process the additional inputs. They just feel more certain while being no more correct.
This shows up in real portfolios. A landmark study in The Journal of Finance tracked 66,465 real investor accounts over five years. The investors who traded the most, driven by confidence in the information they were consuming, earned 6.5 percentage points less per year than the market. More data led to more activity, and more activity destroyed returns.
A separate study found that when investors gained access to faster online trading tools and more market data, their returns dropped by over 5 percentage points compared to their own prior performance. Same investors, same market. Just more information access, and worse outcomes.
Sources
If more data isn't the answer, what is? The research points to depth over breadth.
A study in The Journal of Finance tracked mutual fund performance over 15 years and found that managers who concentrated their holdings in a few industries they knew deeply outperformed diversified managers by 1.2% per year on a risk-adjusted basis. A separate analysis from Harvard Business School took this further: fund managers' highest-conviction positions, the handful of stocks they knew best, outperformed the market by 2.8% to 4.5% annually. The rest of their holdings, the filler picked to meet diversification mandates, added nothing.
This pattern holds for individual investors too. A study of 66,465 household brokerage accounts found that investors who concentrated their portfolios in fewer stocks outperformed those with diversified accounts, particularly when they invested in names they were familiar with. The edge came from genuine information advantages built through repeated, focused engagement.
Economists at NYU and Columbia formalized this into a mathematical proof: rational investors who can choose where to direct their attention will naturally specialize in a small number of assets, because the more you know about something, the more valuable additional learning becomes. It's a virtuous cycle: focus creates knowledge, knowledge creates edge, edge rewards focus.
This is why Pinpoint Alpha tracks 15 tickers per subscriber. Not 50. Not 500. Fifteen names you choose, tracked deeply every single day, with institutional flow patterns that compound into a multi-week narrative you can't get from scanning a dashboard.
Sources
Here's a finding that surprises people: when MIT and NBER researchers studied equity analyst reports, they found that the written narrative analysis, the qualitative “here's what we think is happening” section, contained market-moving information that the raw numbers alone did not. When they included the narrative in their statistical models, the quantitative measures (earnings forecasts, price targets, ratings) lost their significance. The story wasn't decoration on top of the data. The story was the signal.
This aligns with broader research on how humans process complex information. A foundational study in the Journal of Personality and Social Psychology presented identical evidence to two groups: one received it as a structured narrative, the other as raw unorganized facts. Same information, different format. The narrative group made significantly more confident and decisive judgments. Raw recall of individual facts was the same across both groups. It was the organization into a coherent story that improved decision quality.
Nobel laureate Robert Shiller made this the centerpiece of his 2017 Presidential Address to the American Economic Association: narratives, not raw data, are what actually drive investment decisions. The human brain is built for stories. Fighting that wiring by presenting investors with data tables and dashboards is working against biology.
Every Pinpoint Alpha email is a narrative. Not a data dump, not a dashboard screenshot, not a list of numbers. A story about what institutional money did today, what it means in the context of this week, and what conditions would change the picture. Because the research says that's not just easier to read — it's a better signal.
Sources
This is the finding that convinced us the product should exist.
Researchers at the University of Chicago Booth School of Business fed complex corporate disclosures into an AI and compared the AI-generated summaries against the original full-length documents. The summaries were dramatically shorter. They were also better predictors of subsequent stock price movements. The AI synthesis didn't just save time. It produced a higher-fidelity signal than the raw source material.
Meanwhile, Wharton researchers reviewed 97 head-to-head comparisons of simple versus complex forecasting methods across 32 published papers. The result: not a single study found that more complexity reliably improved forecast accuracy. On average, complexity increased forecast errors by 27%. The simple models won because complex models overfit to noise, mistaking randomness for patterns and degrading when applied to new data.
Our pipeline ingests thousands of data points every evening: options chains, institutional trade prints, dark pool activity, volatility surfaces, open interest shifts. From there, it computes a focused set of metrics that the research literature has validated as genuinely informative. Then an AI synthesizes those metrics into a narrative analysis for each of your 15 tickers. The raw data goes in. A clear, actionable read comes out. The research says that compression isn't a loss of information — it's a gain in signal quality.
Sources
A field experiment published in Management Sciencetested how investors respond to three types of advice: pure human, pure AI, and human-curated AI. Investors were significantly more likely to follow recommendations from the human+AI combination than from either source alone, especially for higher-stakes decisions. The researchers found that having a human expert validate and contextualize AI-generated insights didn't compromise quality. It increased adoption.
There's a good reason for this. Separate research from Wharton found that people abandon purely algorithmic tools faster than human advisors after seeing equivalent errors, even when the algorithm objectively outperforms. This means that a pure AI product, no matter how accurate, faces a trust deficit that a human-curated product does not.
Pinpoint Alpha is AI-generated and human-curated. The data pipeline and analysis engine are algorithmic. The interpretation rules, signal prioritization, and quality control are designed and maintained by a human who trades the same markets and uses the same data daily. This isn't a black box that spits out “bullish” or “bearish.” It's a system built by someone who understands what these signals mean in practice, using AI to scale that understanding across 15 tickers every day.
Sources
The framework
The Pinpoint Conviction Score is a four-component framework that produces a single committed read on the market every trading day. The four components combine into one signed score from −100 to +100, mapped to eight regime labels — from Strong Bull through Sideways through Strong Bear.
The framework is anchored by 30+ years of trend-following research. Each pillar maps to a published, validated piece of the trend literature; the integration logic and hysteresis rules are calibrated against ten years of SPY data.
Faber's quantitative work on tactical asset allocation shows that price relative to a long-term moving average is the single most robust predictor of forward returns and drawdown risk across asset classes.
Mean-reversion research demonstrates that extreme deviations from short-horizon moving averages predict near-term mean-reversion pressure. The further price stretches above (or below) trend, the more fragile the move becomes.
Weinstein's stage analysis defines a market in four phases by the slope of its long-term moving average. A flat or rising slope reduces drawdown probability sharply; a downward-sloping average is the strongest signal that supply has overtaken demand.
CBOE multi-timeframe research and Detrick's LPL trend studies show that trend signals which align across short, medium, and long horizons are materially more durable than signals that fire on a single timeframe in isolation.
Honest scope
The score is built on a deliberate trade-off. By design:
We will keep you long through entire bull legs, even when daily flow data looks contradictory underneath. The 22-day rally in late April / early May 2026 is the canonical example: the score held the same bullish read every day while skeptics churned out underneath-divergence warnings.
We will get you out before sustained drawdowns. When a real regime change happens the score downgrades within 5–7 sessions — in practice catching the move at roughly −4% to −7% off the absolute top, before the bulk of a −10%+ decline. In the COVID crash of 2020 the score flipped bearish at SPY −7.7% from peak, well before the −34% drawdown that followed.
We will NOT call you out two days before a top. No framework reliably does. The Conviction Score is explicitly designed to be slow at absolute turns. If you want a tool that calls every wiggle, this is not it.
We add yellow flags before regime change. Volume non-confirmation, RSI divergence, VIX divergence, and put/call complacency are surfaced as cautionary yellow flagsalongside the score. They don't flip the directional read on their own — they downgrade conviction and trigger explicit warnings in the email.
The framework optimizes for net trend capture minus drawdown survival — not for predicting tops. Over ten years of backtested SPY data (2,313 sessions) the score would have spent 71% of all days correctly committed to the prevailing trend.
The score committed to a bullish read at SPY $676 on April 8 and held that read for 22 consecutive trading sessions as the index rallied to $738. Throughout the run, three yellow flags fired in total — volume non-confirmation on two days, a brief RSI divergence on one — each surfaced explicitly in the daily email. The directional read never flipped, because none of the four pillars degraded enough to warrant a regime change. This is exactly the kind of run the framework is built to keep subscribers in.
| Period | Detail | Outcome |
|---|---|---|
| 2018–19 trend | Dec 17, 2018 → Apr 1, 2019 · 153 sessions | 63% of sessions held the bullish label. SPY +19.0% over the run. ~24% of sessions read off-trend. |
| 2020 COVID drawdown | Feb–Mar 2020 · regime flipped within 6 sessions of top | Score downgraded to bearish before the bulk of the −34% peak-to-trough decline. |
| 10-year SPY base rate | 2,313 sessions · 2014–2024 | 71% of all sessions correctly committed to the prevailing trend — matches the long-run base rate of SPY trading above its 200-day moving average. |
| May 2026 (live) | Apr 8 → May 8, 2026 · 22 sessions | Held a bullish read for 22 consecutive sessions while SPY rallied +8.5%, $676 → $738. |
Backtest results reflect what the score said, not what any individual subscriber earned. Past framework readings do not guarantee future regime accuracy.
Sources
15 tickers, not 500. Because concentrated attention outperforms scattered attention by 1-4% annually, and the research proves that rational investors should specialize.
Daily narrative, not a dashboard. Because narrative structure produces better decisions than raw data presentation, and analyst narratives contain market-moving information that numbers alone do not.
AI synthesis with human curation. Because AI summaries outperform original source documents, but human oversight earns the trust that sustains long-term use.
Fewer signals, higher conviction. Because complexity increases forecast error by 27%, and simple models outperform in uncertain environments like financial markets.
Weekly carry-forward analysis. Because multi-day persistence is where the real signal lives, and no single-day data point is as informative as a week-long institutional narrative.
The result: you spend 5 minutes scanning your daily email instead of 45 minutes interpreting a dashboard. The research says that's not a shortcut. It's a better process.
Want to see the specific signals we compute and track? Explore our Signal Glossary to see every metric, what it measures, and why it matters.
Pinpoint Alpha is an analytical newsletter, not a financial advisor. We provide data-driven institutional flow analysis, not investment recommendations. Past signal accuracy does not guarantee future results. Always do your own research before making investment decisions.
