Crowded Longs, Failing AI Leadership & the 90% Win-Rate Lie

Market / Macro Context

• The S&P 500 and NASDAQ are described as being at or near all-time highs (mid-August), with the comment that “markets have had some strong runs this year.”
• Despite strong index performance, the hosts argue that AI-related / “crowded” leadership is not fully reclaiming losses from the July liquidation/bottom period.
• US bonds are noted as particularly weak (“general market bonds have been particularly weak US bonds”).
• Central theme: even if the overall trend is still up, relative leadership has changed, so positioning should shift from aggressive to neutral/cautious.

AI Trade / Equity Leadership & Stock Selection

• The hosts discuss the idea that stocks that didn’t rally when indexes rallied may be vulnerable if the index pulls back.
• They emphasize comparative strength:
  • If a stock/sector is lagging while the market is near highs, the lag may be a warning flag, not a “buy-the-dip” signal.
• Caterpillar (CAT) is cited as an “AI trade” / AI buildout beneficiary via industrial exposure.
  • Observation: CAT can’t reclaim above its 50-day moving average, even as the broader market reaches highs.
• Other AI/semis/AI-adjacent tickers discussed in a “lagging vs leading” context include:
  • Nvidia (NVDA) (noted with a major news item below)
  • Intel (INTC)
  • Arm (ARM)
  • AMD
  • Astera Labs (ticker not specified)
  • Onsemi? / NBIS (spelled as “NBIS” in subtitles)
  • Onto Innovation (ticker not provided)
  • Dell (ticker not explicitly stated)
  • Flex (ticker not provided)
  • A “K10 / K Corp” (subtitles show “Kay Corp / Kay 10 Corp”; ticker not clearly stated)
  • SanDisk is mentioned (no ticker shown in subtitles)
• “Old leaders” concern:
  • Some previous leaders fell in July and are now only partially rebounding, implying rotation / leadership failure risk (“you’re stuck in old merchandise”).

Positioning / Crowding Signals (Strategies Implied)

• Focus is placed on positioning and crowdedness.
• A “crowded trade” example is described as long Dow / short Nasdaq, but the main discussion recommendation leans the opposite way:
  • “Long NASDAQ, short Dow”
  • Framing also suggests that if shorting, Dow is a place to play the short side.
• Rationale:
  • Dow positioning is described as “out of control”, with speculators selling into the NASDAQ rally.
• VIX as a complacency/risk gauge:
  • Framework mentioned: if VIX is below its lower Bollinger Band, it can be an early “ignore/bullish” signal, but otherwise is still a concern.
  • The hosts note that entering band conditions has historically preceded pauses.
  • They connect this with complacency and multiple AI complex stocks still lagging.
• General risk posture:
  • Prefer to reposition out of stocks that “whacked you in July” and remain behind, rather than chase them merely because the overall market is up.

Major News / Numbers Mentioned (AI Funding & Growth)

• Anthropic: revenue up approximately 1,400% year-over-year (as stated).
• Nvidia: forming a $500 billion consortium to fund AI buildout.
  • Characterized as mostly a memorandum of understanding (i.e., less than fully committed capital).
• These developments are treated as positive, but the hosts still question whether equities have fully rebuilt leadership.

Commodities: Copper as the “Most Crowded AI Trade”

• Copper futures are singled out as the most crowded trade.
• Explicit positioning claim:
  • “This is the longest speculators have been in copper ever.”
• Setup described:
  • A reversal day is mentioned; on a day with positive news about Congo (subtitles indicate 7th largest exporter and/or ~7% share), copper gapped up, tried to break out, then failed.
• Recommendation / caution:
  • Copper is presented as a high-failure-risk setup due to crowding (“Everybody’s long”).
  • The hosts describe copper as useful for hedging AI longs or short exposure, calling it a “favorite AI short,” while also stressing no guarantees.

Copper Ticketers / Instruments Explicitly Mentioned

• Copper (futures): “copper” / “copper chart” / “copper futures” (instrument name; no ticker)
• COPX: copper-related ETF
• Freeport-McMoRan: FCX
• “Southern Copper”: ticker not explicitly shown (commonly SCCO, but not stated in the summary)

Hedging / Derivatives Caution

• A discussion/disclosure emphasizes that hedges and spread trades can fail.
  • Example: shorting software vs long AI hardware where the hedge doesn’t behave as expected, leading to a blowout.
• Core caution:
  • Don’t “play derivatives” as if hedged; increasing size under hedging assumptions can raise blow-up risk.

“90% Win-Rate Lie” / Performance-Metrics Debate (Framework)

• The hosts critique marketing claims that highlight “high win rates,” arguing win rate alone is misleading without considering win/loss distribution.
• Key ideas:
  • High win rates may come from small wins paired with rare but large losses (an asymmetric payoff problem).
  • Martingale is cited as a classic example: very high probability of small wins but eventual ruin due to sequence/tail risk and bet limits.
• Methodological points shared (implied framework):
  • Relative strength filter: confirm whether individual stocks are participating in index rallies; if not, ask what happens on a pullback.
  • Positioning / crowding filter: use signals like VIX (including Bollinger band behavior) and COT/crowded positioning thinking (referenced explicitly for copper).
  • Trade expectation sanity check: don’t anchor on win percentage; focus on risk/reward and frequency of big losses.
  • Process-based improvement: improve edge by learning from winning vs losing trades, and reduce trade frequency when needed (the speaker notes win rate rises as trades reduce).

Key Performance Metrics and Numbers (Trading/Statistical Claims)

• Win-rate discussion:
  • One host states a win rate ~38% over time.
  • Claims a return-to-drawdown > 3:1 (“over 3 to one”).
• Critique of “90% win rate” sellers:
  • Allegation that many sellers compute win rate using closed trades, excluding positions that haven’t been closed when they eventually blow up (described as a “scam” in subtitles).

Explicit Recommendations / Positioning Calls

• Equity positioning bias:
  • Shift from aggressive to neutral/cautious due to leadership not fully restored after July.
  • Prefer staying out of the Dow (if the goal is avoiding crowded/lagging areas), and tilt toward NASDAQ leadership.
• Copper:
  • Treat copper as a possible short / hedge candidate based on crowding, while emphasizing it is not guaranteed.
• Portfolio/strategy behavior:
  • Don’t chase names because the broader market bounced.
  • If a stock/sector is no longer leading, stop playing it and rotate into stronger relative performers.

Disclosures / Disclaimers

• Subtitles include “no guarantees” language regarding the copper setup.
• No verbatim “not financial advice” disclaimer appears in the provided subtitles, but the hosts repeatedly stress uncertainty (e.g., “we don’t know…”, “not predicting…watch what they do”).

Mentioned Tickers / Instruments

Indices / References

• S&P 500
• NASDAQ
• Dow (DJIA)
• Russell (referenced as “Russell”)
• S&P (referenced as “Spies”)

Equities / ADRs / Stocks

• NVDA, CAT, INTC, ARM, AMD, FCX, NBIS (unclear mapping)
• Additional company names mentioned without clear ticker mapping in subtitles:
  • SanDisk
  • Astera Labs
  • Onto Innovation
  • Dell
  • Flex
  • “K10 / K Corp” (not clearly mapped)

ETFs / Commodities

• COPX
• Copper (futures implied)

Presenters / Sources

• Matt Caruso (co-host)
• Jason Shapiro (co-host)
• Podcast/guest/source mentioned: Daniel Katylo (“Diary of a CEO” podcast)
• Context mention: OpenAI (in an AI safety discussion)
• Mentioned entity: Stanley Druckenmiller (misspelled as “Ducken Miller” in subtitles)