🧠 Smart Money Consensus Report

The thesis in one line: When multiple elite hedge fund managers independently converge on the same stock, or when one manager with proven edge takes a big solo bet in their edge domain, that's actionable alpha. This report finds those moments.

1. Pick 2-3 signals from the top of the list. Don't try to buy everything — smart money's signals have ~50% hit rate historically. Quality > quantity.
2. Size based on entry quality. RIDING = full size. SMALL PREMIUM = 2/3 size. CHASING = 1/3 size or skip. LATE = skip.
3. Respect the cult penalty. If a name is flagged as cult, sizing should be smaller even if score is above threshold — the alpha is probably mostly gone.
4. Watch for exit signals next quarter. If 2+ signaling funds exit, that's your signal to reduce.
5. 13F is 45 days stale. Funds with high turnover (Druckenmiller 64%, Greenspan 52%) may have already rotated out. For them, treat signals as "what they bought 1-2 months ago" not "what they hold now."
6. This is not financial advice. Smart money is wrong ~40% of the time. Use it as one input, not gospel.

13F filings disclose holdings but not purchase prices. By the time they're public (45 days after quarter-end), the stocks may have moved significantly from where the smart money bought them. This tab tells you how much of the edge has been consumed before you place your order, plus a standalone valuation snapshot for each signal.

Important caveat: price context is a timing filter, not a signal itself. A high-conviction signal that's 🔴 stale is still worth watching — the thesis may extend for years. Conversely, an ⚫ below-cutoff signal needs careful due diligence: smart money got the falling knife, not you. Always combine Price Edge with signal score, entry quality, and Q/Q changes.

When 6+ funds hold the same stock, the cult penalty (0.5x or 0.75x) suppresses its signal score below the 3.0 emission threshold — because "everyone already owns it" isn't alpha. But if that crowded name has since fallen significantly, the setup changes. It becomes "17 funds own it AND the crowd hasn't bought the dip yet" — a genuinely contrarian entry on a name with deep smart-money conviction.

1. Raw score ≥ 4.0 — pre-penalty holdings conviction was strong
2. Cult penalty < 1.0 — 6+ funds hold it (was penalized)
3. Final score < 3.0 — was actually suppressed below emission threshold
4. Price fell ≥ 8% since 13F cutoff — the decay has made the entry attractive

Sorted by rescue attractiveness = raw_score × |price_drop_pct| / 10. Higher = stronger holdings-conviction combined with bigger price drop.

These are NOT auto-emitted signals. They're candidates for your manual review — names where the cult-penalty logic said "skip, crowded" but the price action has since re-opened the trade. Before acting:

1. Verify the thesis isn't broken — why did the stock fall? Earnings miss, management issue, regulatory problem, macro event? A macro-driven drop is an opportunity. A company-specific thesis break is NOT.
2. Check vs-SPY carefully — if it fell alongside SPY, it's a macro dip. If it fell while SPY rose, the market knows something specific.
3. Size conservatively — these are contrarian second-chance trades. Half the size you'd use for an emitted Tier-1 signal.
4. Expect the cohort to exit next quarter if thesis is genuinely broken — keep an eye on the next Q/Q diff. If 3+ suppressed-signal funds exit, the setup is dead.

The rotation into AI infrastructure + nuclear power + China is unmistakable. Seven sectors show net-positive institutional buying this quarter. The standouts:

1. AI Infrastructure (+16 net buys, ZERO sells) — CoreWeave, Nebius, Arista, and Broadcom absorb the capital rotating out of direct Nvidia exposure. This is the cleanest 'picks and shovels' thesis.
2. Power/Nuclear (+6 net buys, ZERO sells) — Vistra (VST), Constellation (CEG), plus speculative positions in SMR and OKLO. The AI-capex-needs-power thesis has gone mainstream.
3. China/EM (+6 net buys, ZERO sells) — Tepper ADD_BIG on BABA, PDD, KWEB, FXI. Sosin and Vinall join him in BABA. Druckenmiller's EWZ (Brazil) also counts here.
4. Industrials/Railroads (+3 net) — Kapadia's edge domain + Hohn's long-term monopolistic-business thesis converge on CNI, CP, UNP, NSC, CSX.

The losers:

1. Healthcare (-4 net) — 4 funds trimming (Druckenmiller NTRA/TEVA/INSM/VRNA + Mandel LLY). This contradicts the dry-run thesis that positioned healthcare as a major theme.
2. SaaS (-2 net) — Laffont EXIT SNOW, Kacher TRIM DDOG, Druckenmiller EXIT MDB. Multiple Tiger Cubs rotating out of high-multiple SaaS.

What's NOT a consensus despite looking like one: Crowded names like GOOGL (17/20 funds), META (15/20), MSFT (8/20) fail to generate signals because the cult-stock penalty combined with the Tiger-Cub haircut drops their composite scores below 3.0. The formula is working as intended — it rewards distinctive, uncorrelated conviction, not ubiquitous holding.

Sector Flows: Net Positive

Sector Flows: Net Negative

What does a "good" signal look like?

A good signal has all of these: (1) Score ≥ 4.0, (2) 3+ funds converging, (3) style diversity ≥ 2 (so it's not just the Tiger Cubs herding), (4) Entry Quality = RIDING or SMALL PREMIUM (you're still close to the funds' cost basis), (5) not on the Cult Stock list. BABA this quarter is the textbook example: 4.90 score, 3 funds, 2 styles (macro + value + European value), RIDING entry, not crowded.

What's the "Cult Penalty" and why does it exist?

The cult penalty is a score multiplier that reduces the signal strength when too many funds hold the same name. 10+ funds hold → 0.5x (halves the score). 6-9 funds → 0.75x. ≤5 funds → no penalty. Why? Because if 17 of 20 elite funds already own GOOGL, the information is fully priced in — there's no edge in finding out they own it. You already knew. The penalty forces the formula to reward distinctive conviction, not ubiquitous holding. GOOGL with 17 holders dropped to 2.96 score (below the 3.0 threshold) this quarter despite being the most-held name — exactly the intended behavior.

What's the "Tiger Cub Haircut"?

"Tiger Cubs" are hedge funds founded by alumni of Julian Robertson's Tiger Management — Tiger Global (Coleman), Light Street (Kacher), Coatue (Laffont), Lone Pine (Mandel), Altimeter (Gerstner), SCGE (Wang), Slate Path (Greenspan). They share a similar investing style (high-growth tech), so they often pile into the same names. The haircut (0.67x) discounts signals where 3+ Tiger Cubs are the primary holders — because correlated managers agreeing isn't 3 independent data points, it's 1 style voting 3 times.

Why didn't GOOGL/META/MSFT signal? Everyone's buying them.

Exactly — "everyone's buying them" IS the problem. GOOGL is held by 17/20 funds, META by 15/20, MSFT by 8/20. After the cult penalty (0.5x) and style haircut (0.67x for GOOGL/META where 8+ Tiger Cubs hold), their composite scores fall below the 3.0 emission threshold. The formula is working correctly: a signal's value is in telling you something you don't already know. GOOGL at a trillion-dollar valuation with 17 smart-money holders is common knowledge, not alpha. The signal absence for these names is itself information — it tells you to look elsewhere for edge.

What's the difference between the 3 tiers?

Tier 1 (CONSENSUS) — 3+ funds hold. Safest because the independent agreement reduces single-manager risk. Within Tier 1, "Strong" = 4+ funds across 3+ styles; "Moderate" = 3+ funds across 2+ styles; "Weak" = 3+ funds same style (risky because they may be correlated).

Tier 2 (SOLO-EXPERT) — 1 fund, but that fund has explicit edge in the domain. E.g., Aschenbrenner going big on nuclear power is SOLO because his edge domain IS AI infrastructure / power. This can be as actionable as Tier 1 but higher variance.

Tier 3 (CROSS-DOMAIN) — 2 funds from DIFFERENT styles converging. Emerging signal. Often becomes Tier 1 next quarter as more funds discover the idea.

What does Entry Quality mean?

A proxy for "how much of the move have the funds already captured?" If most funds are NEW or ADD_BIG, the position is freshly built — you're close to their cost basis = RIDING. If they're all HOLDing a position built quarters ago, the stock may have already run = CHASING or LATE. Practical translation: RIDING signals can get full position sizing; CHASING should be half-sized; LATE should be skipped or watched for a pullback.

What's "Freshness" — why does fund turnover matter?

High-turnover funds (Druckenmiller 64%, Greenspan 52%) trade in and out aggressively — their signals are time-sensitive and may already be stale by the time you see them (13Fs are 45 days stale at filing). The 0.5x freshness multiplier for >50% turnover accounts for this. Low-turnover funds (Buffett 5%, Hohn 9%, CAS 6%) get a 1.2x boost because when they act, it's meaningful — they've rarely changed their mind in years.

What's "Conviction"?

The % of a fund's portfolio allocated to the position. A 10%+ position is a bet-the-fund conviction (2.0x multiplier). A 0.5-2% position is exploratory (0.6x). The multiplier reflects how much the fund itself believes in the name. Ackman holding UBER at 18.5% of his entire portfolio is a far stronger signal than one of 50 fund positions at 0.3%.

What does "Action" encode?

What the fund did this quarter vs. last: NEW (2.0x — opened a new position), ADD_BIG (1.5x — added >50% more shares), ADD (1.2x — added 20-50%), HOLD (1.0x — neutral), TRIM (0.5x — reduced 20%+), TRIM_BIG (0.2x — reduced 50%+), EXIT (-1.0x — fully out). NEW and ADD_BIG carry the highest signal because they represent active capital commitment in the current quarter.

What's "Style Diversity"?

How many distinct investing styles are represented among the holders. 5 Tiger Cubs holding something = 1 style. Tepper (macro) + Sosin (value) + Aschenbrenner (AI infra specialist) holding something = 3 styles. More distinct styles = more independent conviction = higher-quality signal. This is what separates a genuine consensus from a crowd-trade.

Why 3.0 as the emission threshold?

Empirically chosen. Scores below 3.0 historically don't generate enough edge to justify action after transaction costs and taxes. Scores 3.0-4.0 are "emerging" (small position or watchlist). 4.0-5.0 are "strong" (meaningful position). 5.0+ are "elite" (max conviction).

What are the Exit Signals?

When 2+ funds who previously held a name fully EXIT in the same quarter, the formula emits a REDUCE signal. This is your early-warning for broken theses. Single-fund exits are shown in the "All Complete Exits" table as watchlist items. Q4 2025 had no multi-fund exits.

Why did Healthcare show as a losing sector when my last report was Druckenmiller-heavy?

The Druckenmiller-only dry run (Q4 2025, generated Apr 30 13:52) positioned Healthcare as the dominant theme because NTRA was his top position. That was accurate for Druckenmiller. But with all 20 funds visible, the aggregate picture shows 4 funds trimming healthcare (Druck TEVA/INSM/VRNA/NTRA + Mandel LLY) and NONE adding meaningfully. Healthcare flipped from "Druckenmiller's story" to "-4 net" across the cohort. This is the power of running 20 funds vs 1: it distinguishes idiosyncratic from consensus.

What about cult stocks that have FALLEN since the 13F? Aren't those interesting?

Yes, and that's exactly what the 💎 Hidden Value tab surfaces. The cult penalty assumes a crowded name's alpha is already priced in — which is true when the price keeps climbing. But if a 15-fund cult name has since dropped 10%, the "crowd has bought" logic reverses: the market has gotten off the name, and you can enter at a better price than the smart money got. The detector scans every position across all 20 funds (not just emitted signals) and flags any name where raw holdings conviction was ≥4.0, cult penalty suppressed it below threshold, and price has fallen ≥8% since cutoff. This quarter: MSFT (8 funds, -16.4%) and META (15 funds, -8.1%) both qualify — names the cohort loaded up on in Q4 that the market has since faded.

Why aren't Hidden Value names auto-emitted as Tier-1/2/3 signals?

Because they failed the original score criteria and the rescue logic is a post-hoc "second look." Auto-emitting them would mean the cult penalty isn't really a penalty — it'd just be a delay. Instead, Hidden Value is a manual-review candidate list: names worth considering but requiring extra due diligence (why did the stock fall? is the thesis intact? macro dip or company-specific?). Treat Hidden Value like a watchlist, not a buy list. Size positions half what you'd use for an emitted Tier-1 signal.

What's the "rescue score"?

Rescue score = raw_score × |price_drop_pct| / 10. It's a sort order, not a decision threshold. MSFT at raw 5.2 with -16.4% drop scores 8.52; META at raw 7.3 with -8.1% drop scores 5.94. Higher means stronger pre-penalty conviction combined with bigger decay. Use it to prioritize review order, not to decide what to buy.

How is today's stock price factored into the signal score?

It isn't — by design. The signal score is built entirely from 13F holdings data: share counts, portfolio %, conviction, cohort convergence, cult crowding. 13F filings don't disclose purchase prices, so there's no cost basis to reference. The score answers "how much did the smart money commit to this?" — not "is today's price a bargain?" Those are two different questions, and the report separates them intentionally.

Then how do I know if the current price still makes sense?

That's what the 💰 Price & Valuation tab exists for. It layers price context on top of the signal score — showing the 13F-cutoff price, today's price, the % change since cutoff (is the edge still there, or did the post-13F rally consume it?), vs-SPY relative performance, and a valuation snapshot (P/E, Forward P/E, EV/EBITDA, FCF yield, 52-week range position). Use Price Edge as a timing filter on top of the signal score.

What do the Price Edge bands mean?

For each emitted signal, the price change since 13F cutoff determines a band:

1. 🟢 Intact (±5%) — the stock hasn't moved much. You can still buy near smart money's average cost. Full-size position if other criteria check out.
2. 🟡 Decaying (+5% to +15%) — some edge has been consumed but a long-horizon thesis still has room. 2/3 size.
3. 🔴 Stale (+15% or more) — the post-13F rally captured the bulk of the alpha. You're late. 1/3 size or skip; wait for a pullback.
4. ⚫ Below cutoff (negative) — the stock has fallen since smart money bought. You can potentially enter at a BETTER price than they got. Aggressive buy — but verify the thesis isn't broken (why did it fall?).

"Below cutoff" — that sounds too good. Isn't smart money catching a falling knife?

Sometimes yes. A ⚫ signal demands extra due diligence. Ask: did anything fundamentally change since the filing? Earnings miss? Management shake-up? Regulatory issue? If the thesis is intact (macro event pulled everything down, unrelated sector rotation, a one-quarter disappointment in a multi-year story), then yes, you're getting a better entry than three elite funds got. If the thesis is broken (the company's story changed), then the funds will likely exit next quarter's 13F — and you'd be stepping into a dying position. The ⚫ band is the highest-signal, highest-risk category. Always cross-reference with valuation metrics and recent news.

What's "vs SPY"?

The signal's price change since 13F cutoff minus SPY's price change over the same period. Positive = outperforming the market (the thesis is working). Negative = underperforming (either the market moved and the signal didn't, or the signal moved down more). A stock up 8% while SPY was up 20% looks "intact" in absolute terms but is actually lagging badly — vs SPY catches this.

How do I read the valuation metrics?

Rough guide:

1. Trailing P/E: <15 = cheap, 15-25 = normal, >30 = pricey. Missing/None = company is not profitable.
2. Forward P/E: generally preferred over trailing because it uses next-year estimates. Same bands, usually lower than trailing for growing companies.
3. EV/EBITDA: <10 = cheap, 10-15 = normal, >20 = pricey. Negative EV/EBITDA means the company isn't profitable at the EBITDA level — common for early-stage growth names like NBIS (Nebius) but a major red flag for mature companies.
4. FCF Yield: free cash flow divided by market cap. >6% = cheap on cash generation, 3-6% = normal, <2% = growth premium, negative = burning cash.
5. 52-week range position: 0% = sitting at 52w low, 100% = sitting at 52w high. <30% = washed out (potential bounce candidate). >70% = near highs (momentum or euphoria).

These are just starting points — valuation is contextual (a 35 P/E is cheap for a 30%-grower, expensive for a 3%-grower). The tab gives you the numbers; you have to apply the context.

Where does the price data come from?

Yahoo Finance via the yfinance Python library. Free, no API key, rate-limit-tolerant for single-run batches. Prices are fetched at report-generation time, so each regenerated report reflects the latest close. The cutoff date is the last trading day of the reporting quarter (e.g. Dec 31 2025 for Q4 2025).

Why didn't my signal's valuation show certain metrics?

yfinance's info dict doesn't always populate every field. Common reasons: (1) ADRs or foreign-listed names have partial data, (2) recently-IPO'd companies have short histories (CRWV shows "None" P/E because it had trailing losses and yfinance's P/E logic won't return a negative number), (3) company is in a transition (pre-revenue, pre-profit, or post-restructuring). The table shows "—" for missing fields. Cross-reference with a second source (Finviz, Seeking Alpha) if you need the number.

What kill-switches exist?

Nine kill-switches prevent bad signals during extraordinary conditions: (1) VIX > 40 = advisory only, no new BUY signals. (2) Manager retirement = exclude fund 4 quarters. (3) Fund AUM drops >50% QoQ = 0.5x weight. (4) SEC enforcement on signal name = refuse BUY. (5) Delisting/bankruptcy risk = refuse BUY. (6) Late/amended 13F = flag. (7) Pending M&A/tender = special situation, no new BUY. No kill-switches are active for Q4 2025.

How often does this run?

Quarterly, 3 days after the 13F filing deadline:

• Q4 2025: Feb 17, 2026 (this report — dry-run)
• Q1 2026: May 18, 2026 (next scheduled automated run)
• Q2 2026: Aug 17, 2026
• Q3 2026: Nov 19, 2026

What can this NOT see?

13F filings disclose US equity long positions only. Not visible: (1) short positions — funds may be long GOOGL + short META as a pair trade, and we'd only see the GOOGL. (2) Put options used as hedges. (3) Total-return swaps (Archegos-style leverage). (4) Non-US holdings (a European fund's Japanese stocks are invisible). (5) Private positions (pre-IPO). (6) Cash positions and leverage. Always assume the picture is incomplete, especially for long/short funds like Jericho.

What if a fund files an amended 13F later?

Amended filings get flagged but don't retroactively change prior reports unless the change is >20% of the position. The next quarterly run picks up any amendments automatically.

Is this financial advice?

No. Smart money is wrong ~40% of the time — elite hedge funds blew up on Archegos, Silicon Valley Bank, Chinese EVs, various WeWork-style bets. This report is a data-driven starting point for your own research, not a trade execution system. Always do your own due diligence, size positions within your risk tolerance, and never bet more than you can afford to lose.