Scorecard

Falsifiable stakes, graded automatically. Each prediction is minted with its resolution date fixed — up or down at the print, or beats the index over 1–3 months — and resolves against closing prices with no re-litigation. The stance ledger below grades every recorded call the same way.

Open predictions

Minted	Ticker	The claim	Conv	Resolves	Why
Jul 19	NTDOY	beats SPY over 3mo	3	Oct 19	The 'is ~11% sizing earned' falsifiable proxy on the #1 position: 8/06 print's 16.5M-unit sandbag beat + Sept hardware hike offsetting memory cost (the defined ADD trigger) plays out over the quarter, not the print night. SPY benchmark (non-tech comp for a JP consumer name).
Jul 20	RDDT	beats QQQ over 3mo	2	Oct 20	Beat-index half of JGC-411 (Sean 7/19). The alpha-basket 3mo-DRIFT stake, deliberately distinct from earnings-direction #112: the edge is the drift, not the print-night sign — Yipit-Q3 corroboration (momentum carries past the 7/30 print, the NTDOY quarter-shape), rising rate-of-change estimate revisions (BofA $180->$210, WF floor $176->$187, Needham $300 Conviction-Buy = a robust multi-month alpha factor pointing UP), + the #124 sizing-fork resolution over the quarter. Cheapest-growth in the basket (P/E/G 0.6, -36% off 52w high) blunts the de-rate tail. LOW CONVICTION (2) by design: beta amplification (~2-3x QQQ) is the dominant un-edged variance and print-independent Advance/Tencent supply overhang can cap the name regardless — adversarially verified as a ~53-55% lean, NOT a coin-flip (survives-thinly). QQQ = small-in-index/non-self-referential. 3mo-ONLY: a 1mo beat-index would be a levered duplicate of #112 (print-dominated). Alpha-basket AXON + DASH abstained (documented edge-grounds, not date).
Oct 9	FSLY	lags QQQ over 1mo	2	Nov 9	Upgrade jump (+6% pre-open 10/09) with the ~11/04 report inside the month. Fastly's own 7 such jumps since 2020 trailed QQQ over the month 5 of 7 (median -25); Sean's 8+ pt jump sales with a report inside the month trailed 7 of 8. Small samples; the next 10 days lean the other way (11 of 17 beat). Stake is on the month through the report, not a sell call; desk stance HOLD 3. Minted pre-open at the 10/08 close ($25.29), so the stake must also give back the ~6% gap: stricter than the study, which measured from the jump-day close.
Oct 7	DASH	beats QQQ over 1mo	2	Nov 9	Quick-sale forward test: Sean sold DASH (IBKR) on 2026-10-06, after holding it 8 days on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Oct 13	SOXX	beats SPY over 1mo	2	Nov 13	The 10/5-10/12 chip drop (SOXX -7.6%, still above its 50- and 200-day, +9.5% over 20 sessions) reads as profit-taking plus a financing scare, not a demand break: money rotated to platforms inside tech (XLK +$1.3B inflow). Base rate since 2019: 6 of 8 similar weeks beat SPY the next month (median +2.9 pts), one fell another 12.5%. Breaks if TSMC 10/15 guides 2027 capex near or below ~$80B.
Oct 13	SMH	beats IGV over 1mo	—	Nov 13	Desk forward call from 10/13: chips (SMH) beat software (IGV) over the next month. Prompted by Sean's 10/11 read that the long-software/short-chips swing was near a top, but this row does NOT grade that call: it starts at the 10/13 price, after software ran about 9.6% further ahead of chips, and that move counts neither for nor against him. Basis: since 2018, after comparable software-over-chips spikes, chips beat software the next month 8 of 10 times (median 3.8 points). Pair to #864 (SOXX vs SPY), which resolves the same day.
Oct 16	AVGO	beats QQQ over 1mo	2	Nov 16	Quick-sale forward test: Sean sold AVGO (Roth) on 2026-10-15, after holding it 3 days on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Oct 16	SMH	beats IGV over 1mo	3	Nov 16	Software-only selloff 10/16 (IGV ~-4%, SPY ~-0.3%). In 24 such days since 2016, SMH beat IGV over the next 20 sessions 16 times, median +4.4 points, even though IGV itself was up 19 of 24. The software dip usually recovers, but it isn't a rotate-into-software signal.
Oct 17	VRT	beats QQQ over 1mo	2	Nov 17	Quick-sale forward test: Sean sold VRT (IBKR) on 2026-10-16, after holding it 1 day on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Sep 20	CRWV	lags QQQ over 3mo	3	Dec 21	Read from the primary docs, not the headline: CoreWeave's 9/17 8-K lists the stack its new convert ranks beside - senior notes at 9.250/9.000/9.750/9.625/8.500% - and its own 424B5 use-of-proceeds says it is 'migrating our enterprise credit profile toward investment grade.' A company says that only when it isn't there. Two ticks later NVIDIA's CFO confirmed it from the supplier side: roughly a quarter of next year's business goes to AI labs 'for which we expect to leverage our balance sheet,' consumed by 'investment-grade customers OR THOSE THAT ARE BACKED BY ONE' - a clause that concedes the direct counterparty is not IG. The stake: funding a capex race at ~9% against hyperscaler competitors funding at IG is a structural cost-of-capital disadvantage that compounds, and ~$6.42B authorised inside three minutes (35M-share ATM + $3.0B convert + $500M option) is dilution landing now. So CRWV LAGS QQQ over 3mo. THE BULL HALF IS REAL AND I AM TAKING THE OTHER SIDE OF IT KNOWINGLY: >$25B of net new commitments between 6/30 and 8/11, contracted power ~3.7GW to ~4.2GW, and NVIDIA's backstop lowers default risk (though not cost of capital or dilution). A named articulate bull (RHouseResearch, 9/19) is full-port long CRWV against NBIS/IREN. Conviction 3/5 - deliberately not higher, because demand this contracted can outrun a financing disadvantage for longer than 3mo.
Sep 19	NFLX	lags QQQ over 3mo	2	Dec 21	Derived from the primary document, not the tape. Grading NFLX's 7/16 print gate on 2026-09-19 produced one concrete, issuer-sourced fact: Q2 actual revenue grew '13% year over year (+12% on a FX-neutral basis)' and the Q3 guide is 'revenue growth of 12% (or 11% F/X neutral)' — 1pp BELOW the trough, i.e. DECELERATION. The bull case required the opposite: MS (4/17) argued 2Q was the comp trough BY CONSTRUCTION, with the March-2026 US price hike rippling through on a 2-3 month lag, so re-acceleration was near-automatic. Netflix's own guide contradicted that in writing. This stake tests whether that falsification matters over a quarter against a QQQ carried by AI/semis. ⚠️ CONVICTION DELIBERATELY LOW (2/5) — the honest counter is that NFLX was already -21.3% YTD into the print, so the deceleration may be priced, and a cheap name with bad news out can outperform. I am staking the mechanism, not certainty. NOT a position call: zero shares, PITCH #11 stays at 1/5, stance PASS/WATCH unchanged. Minted as ONE defensible stake rather than three guesses after measuring that the desk mints too few falsifiable claims (108 of 336 gradeable ledger rows land neutral) - deliberately did NOT stake STX 10/26 or MU, where I have arithmetic on the GUIDE but no defensible view on the STOCK's direction.
Sep 25	SMH	beats SPY over 3mo	2	Dec 28	The chip run is a breadth surge (AI chips 1 of 11 above the 50-day on 9/14 to 9 of 11 on 9/24), and surges in this group haven't preceded fades: 26 past surges beat the S&P over the next month by +2.0 vs +2.1 on a normal day, 80% beat it over 3 months vs 71% normally. The chips-vs-everything split is at the 99.4th percentile, and in all 7 nearest analogs the chip lead widened over 3 months (median +8.6 pts vs +2.2 normal). Conviction low: 7 analogs, and today's basket lists lean toward past winners. SMH stands in for the desk's ai-chips basket (it also holds chip equipment and memory).
Oct 3	FSLY	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: FSLY led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	META	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: META led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	INTC	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: INTC led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	KLAC	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: KLAC led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	LRCX	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: LRCX led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	AXON	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: AXON trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	DASH	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: DASH trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	TXRH	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: TXRH trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	CMG	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: CMG trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 5	RSP	lags SPY over 3mo	2	Jan 5, 2027	S&P is narrow (SPY beat RSP by 5.2 pts over 63 sessions, 96th pct since 2003). In 13 past narrow episodes the big stocks kept leading 54% of the time (median +0.5 pts), so narrowness has not reliably reversed. A low-conviction lean that tests the 'narrowness unwinds' story.
Oct 12	TMUS	lags SPY over 3mo	2	Jan 12, 2027	After SpaceX's first spectrum buy (EchoStar, 9/08/2025) TMUS lagged SPY by 23.3 pts over 63 sessions; the 10/09/2026 Grain 800 MHz buy is the same buyer and a bigger threat (in-building low band, ground-network option). One confounded precedent (fall-2025 price war) vs TMUS's own extreme-day record (beat 6/12 at 60d), hence conviction 2.
Oct 12	PWR	beats SPY over 3mo	2	Jan 12, 2027	Power-to-site layer led SPY by 4-8 pts after the 9/24 notice (PWR +7.2 ex the 10/09 hurricane day). Record .4B backlog, FY26 outlook raised 7/30; storm work from Isaias adds Q4 revenue. Low conviction: 34x forward, report ~10/29-30 inside the window, and it trades as the AI-buildout factor (r=0.76 with FIX ex-market).
Oct 12	ORCL	lags SPY over 3mo	2	Jan 12, 2027	Layer split after the 9/24 Jupiter force-majeure notice: the building owners/renters lagged SPY 4-7 pts in 11 sessions (ORCL -6.9) while power-to-site names led. Jupiter's construction loan reportedly trades under 90c and the fuel-cell air permit is due by 11/23; Oracle carries the fixed-cost side of a power-gated build. Low conviction: short-horizon relative moves often reverse.
Sep 30	ADYEY	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +9.9% above its 52-week low with 2 of 6 Payments & fintech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	NU	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +6.3% above its 52-week low with 2 of 6 Payments & fintech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	NVO	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +9.0% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	PODD	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +0.5% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	SYK	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +2.2% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	INTU	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +8.0% above its 52-week low with 12 of 30 Software peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	IDXX	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +3.8% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Oct 1	APP	lags QQQ over 6mo	2	Apr 1, 2027	JGC-990 forward test: +0.0% above its 52-week low with 1 of 4 Ad tech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three). Qualifies only under the ad-tech/social basket split (PR #5382, 2026-10-01); the old joint Ads & social basket read it as a group low, so this stake also tests that split.
Oct 1	TTD	lags QQQ over 6mo	2	Apr 1, 2027	JGC-990 forward test: +1.0% above its 52-week low with 1 of 4 Ad tech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three). Qualifies only under the ad-tech/social basket split (PR #5382, 2026-10-01); the old joint Ads & social basket read it as a group low, so this stake also tests that split.
Oct 3	INTC	lags QQQ over 6mo	2	Apr 5, 2027	Repeat de-rating episode (11/2025, 7/2026): forward EPS +92% over 13 weeks while forward P/E fell 28%. Repeat episodes in ten chip/hardware stocks 2016-2026 trailed (median 6-month return +6% vs +17% baseline; exploratory, 17-18 cases).
Oct 3	LRCX	lags QQQ over 6mo	2	Apr 5, 2027	Second quarter within a year (4/24, 8/07) of forward EPS +12%+ with forward P/E -15%+. In ten chip/hardware stocks 2016-2026, repeat episodes returned a median +6% over 6 months vs +17% for any week (17-18 cases, exploratory split, not pre-registered).

Prediction track record

Kind	Right	Total	Hit rate	Avg alpha
Beats index (1mo)	0	1	0%	-3.6%
Earnings direction	3	8	38%	—

Resolved	Ticker	The claim	Outcome	Verdict
Sep 23	CTAS	UP on earnings	-2.5%	wrong

Stance ledger

794 graded (153 right · 174 wrong · 467 neutral) — 47% of decided calls right · 161 open

Called	Ticker	Call	Outcome	Verdict
Oct 11	LULU	stance · PASS	-2.8%	neutral
Oct 11	CRDO	stance · WATCH	0.0%	neutral
Oct 11	CMG	stance · EXIT	+5.4%	right
Oct 11	AMAT

Open predictions

Minted	Ticker	The claim	Conv	Resolves	Why
Jul 19	NTDOY	beats SPY over 3mo	3	Oct 19	The 'is ~11% sizing earned' falsifiable proxy on the #1 position: 8/06 print's 16.5M-unit sandbag beat + Sept hardware hike offsetting memory cost (the defined ADD trigger) plays out over the quarter, not the print night. SPY benchmark (non-tech comp for a JP consumer name).
Jul 20	RDDT	beats QQQ over 3mo	2	Oct 20	Beat-index half of JGC-411 (Sean 7/19). The alpha-basket 3mo-DRIFT stake, deliberately distinct from earnings-direction #112: the edge is the drift, not the print-night sign — Yipit-Q3 corroboration (momentum carries past the 7/30 print, the NTDOY quarter-shape), rising rate-of-change estimate revisions (BofA $180->$210, WF floor $176->$187, Needham $300 Conviction-Buy = a robust multi-month alpha factor pointing UP), + the #124 sizing-fork resolution over the quarter. Cheapest-growth in the basket (P/E/G 0.6, -36% off 52w high) blunts the de-rate tail. LOW CONVICTION (2) by design: beta amplification (~2-3x QQQ) is the dominant un-edged variance and print-independent Advance/Tencent supply overhang can cap the name regardless — adversarially verified as a ~53-55% lean, NOT a coin-flip (survives-thinly). QQQ = small-in-index/non-self-referential. 3mo-ONLY: a 1mo beat-index would be a levered duplicate of #112 (print-dominated). Alpha-basket AXON + DASH abstained (documented edge-grounds, not date).
Oct 9	FSLY	lags QQQ over 1mo	2	Nov 9	Upgrade jump (+6% pre-open 10/09) with the ~11/04 report inside the month. Fastly's own 7 such jumps since 2020 trailed QQQ over the month 5 of 7 (median -25); Sean's 8+ pt jump sales with a report inside the month trailed 7 of 8. Small samples; the next 10 days lean the other way (11 of 17 beat). Stake is on the month through the report, not a sell call; desk stance HOLD 3. Minted pre-open at the 10/08 close ($25.29), so the stake must also give back the ~6% gap: stricter than the study, which measured from the jump-day close.
Oct 7	DASH	beats QQQ over 1mo	2	Nov 9	Quick-sale forward test: Sean sold DASH (IBKR) on 2026-10-06, after holding it 8 days on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Oct 13	SOXX	beats SPY over 1mo	2	Nov 13	The 10/5-10/12 chip drop (SOXX -7.6%, still above its 50- and 200-day, +9.5% over 20 sessions) reads as profit-taking plus a financing scare, not a demand break: money rotated to platforms inside tech (XLK +$1.3B inflow). Base rate since 2019: 6 of 8 similar weeks beat SPY the next month (median +2.9 pts), one fell another 12.5%. Breaks if TSMC 10/15 guides 2027 capex near or below ~$80B.
Oct 13	SMH	beats IGV over 1mo	—	Nov 13	Desk forward call from 10/13: chips (SMH) beat software (IGV) over the next month. Prompted by Sean's 10/11 read that the long-software/short-chips swing was near a top, but this row does NOT grade that call: it starts at the 10/13 price, after software ran about 9.6% further ahead of chips, and that move counts neither for nor against him. Basis: since 2018, after comparable software-over-chips spikes, chips beat software the next month 8 of 10 times (median 3.8 points). Pair to #864 (SOXX vs SPY), which resolves the same day.
Oct 16	AVGO	beats QQQ over 1mo	2	Nov 16	Quick-sale forward test: Sean sold AVGO (Roth) on 2026-10-15, after holding it 3 days on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Oct 16	SMH	beats IGV over 1mo	3	Nov 16	Software-only selloff 10/16 (IGV ~-4%, SPY ~-0.3%). In 24 such days since 2016, SMH beat IGV over the next 20 sessions 16 times, median +4.4 points, even though IGV itself was up 19 of 24. The software dip usually recovers, but it isn't a rotate-into-software signal.
Oct 17	VRT	beats QQQ over 1mo	2	Nov 17	Quick-sale forward test: Sean sold VRT (IBKR) on 2026-10-16, after holding it 1 day on average. Since 2025, 63% of his sales within two weeks beat QQQ over the next month, but the rate swings with the market (1 of 10 in Jan-Mar 2026). A LAG here means the sale was right.
Sep 20	CRWV	lags QQQ over 3mo	3	Dec 21	Read from the primary docs, not the headline: CoreWeave's 9/17 8-K lists the stack its new convert ranks beside - senior notes at 9.250/9.000/9.750/9.625/8.500% - and its own 424B5 use-of-proceeds says it is 'migrating our enterprise credit profile toward investment grade.' A company says that only when it isn't there. Two ticks later NVIDIA's CFO confirmed it from the supplier side: roughly a quarter of next year's business goes to AI labs 'for which we expect to leverage our balance sheet,' consumed by 'investment-grade customers OR THOSE THAT ARE BACKED BY ONE' - a clause that concedes the direct counterparty is not IG. The stake: funding a capex race at ~9% against hyperscaler competitors funding at IG is a structural cost-of-capital disadvantage that compounds, and ~$6.42B authorised inside three minutes (35M-share ATM + $3.0B convert + $500M option) is dilution landing now. So CRWV LAGS QQQ over 3mo. THE BULL HALF IS REAL AND I AM TAKING THE OTHER SIDE OF IT KNOWINGLY: >$25B of net new commitments between 6/30 and 8/11, contracted power ~3.7GW to ~4.2GW, and NVIDIA's backstop lowers default risk (though not cost of capital or dilution). A named articulate bull (RHouseResearch, 9/19) is full-port long CRWV against NBIS/IREN. Conviction 3/5 - deliberately not higher, because demand this contracted can outrun a financing disadvantage for longer than 3mo.
Sep 19	NFLX	lags QQQ over 3mo	2	Dec 21	Derived from the primary document, not the tape. Grading NFLX's 7/16 print gate on 2026-09-19 produced one concrete, issuer-sourced fact: Q2 actual revenue grew '13% year over year (+12% on a FX-neutral basis)' and the Q3 guide is 'revenue growth of 12% (or 11% F/X neutral)' — 1pp BELOW the trough, i.e. DECELERATION. The bull case required the opposite: MS (4/17) argued 2Q was the comp trough BY CONSTRUCTION, with the March-2026 US price hike rippling through on a 2-3 month lag, so re-acceleration was near-automatic. Netflix's own guide contradicted that in writing. This stake tests whether that falsification matters over a quarter against a QQQ carried by AI/semis. ⚠️ CONVICTION DELIBERATELY LOW (2/5) — the honest counter is that NFLX was already -21.3% YTD into the print, so the deceleration may be priced, and a cheap name with bad news out can outperform. I am staking the mechanism, not certainty. NOT a position call: zero shares, PITCH #11 stays at 1/5, stance PASS/WATCH unchanged. Minted as ONE defensible stake rather than three guesses after measuring that the desk mints too few falsifiable claims (108 of 336 gradeable ledger rows land neutral) - deliberately did NOT stake STX 10/26 or MU, where I have arithmetic on the GUIDE but no defensible view on the STOCK's direction.
Sep 25	SMH	beats SPY over 3mo	2	Dec 28	The chip run is a breadth surge (AI chips 1 of 11 above the 50-day on 9/14 to 9 of 11 on 9/24), and surges in this group haven't preceded fades: 26 past surges beat the S&P over the next month by +2.0 vs +2.1 on a normal day, 80% beat it over 3 months vs 71% normally. The chips-vs-everything split is at the 99.4th percentile, and in all 7 nearest analogs the chip lead widened over 3 months (median +8.6 pts vs +2.2 normal). Conviction low: 7 analogs, and today's basket lists lean toward past winners. SMH stands in for the desk's ai-chips basket (it also holds chip equipment and memory).
Oct 3	FSLY	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: FSLY led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	META	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: META led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	INTC	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: INTC led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	KLAC	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: KLAC led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	LRCX	beats SPY over 3mo	2	Jan 4, 2027	JGC-1005: LRCX led the S&P by 19-32 pts in the month into Q3 reports. The top fifth by run-in beat SPY over the next month 53% of the time (55% even after a DOWN report day), ahead of the bottom fifth in all five half-years.
Oct 3	AXON	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: AXON trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	DASH	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: DASH trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	TXRH	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: TXRH trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 3	CMG	lags SPY over 3mo	2	Jan 4, 2027	JGC-1005: CMG trailed the S&P by 17-19 pts in the month into Q3 reports. Across 1,656 reports (200 names, 2024-26) the bottom fifth by run-in beat SPY over the next month 42% of the time (40% after an UP report day). Report + month after falls inside 3 months.
Oct 5	RSP	lags SPY over 3mo	2	Jan 5, 2027	S&P is narrow (SPY beat RSP by 5.2 pts over 63 sessions, 96th pct since 2003). In 13 past narrow episodes the big stocks kept leading 54% of the time (median +0.5 pts), so narrowness has not reliably reversed. A low-conviction lean that tests the 'narrowness unwinds' story.
Oct 12	TMUS	lags SPY over 3mo	2	Jan 12, 2027	After SpaceX's first spectrum buy (EchoStar, 9/08/2025) TMUS lagged SPY by 23.3 pts over 63 sessions; the 10/09/2026 Grain 800 MHz buy is the same buyer and a bigger threat (in-building low band, ground-network option). One confounded precedent (fall-2025 price war) vs TMUS's own extreme-day record (beat 6/12 at 60d), hence conviction 2.
Oct 12	PWR	beats SPY over 3mo	2	Jan 12, 2027	Power-to-site layer led SPY by 4-8 pts after the 9/24 notice (PWR +7.2 ex the 10/09 hurricane day). Record .4B backlog, FY26 outlook raised 7/30; storm work from Isaias adds Q4 revenue. Low conviction: 34x forward, report ~10/29-30 inside the window, and it trades as the AI-buildout factor (r=0.76 with FIX ex-market).
Oct 12	ORCL	lags SPY over 3mo	2	Jan 12, 2027	Layer split after the 9/24 Jupiter force-majeure notice: the building owners/renters lagged SPY 4-7 pts in 11 sessions (ORCL -6.9) while power-to-site names led. Jupiter's construction loan reportedly trades under 90c and the fuel-cell air permit is due by 11/23; Oracle carries the fixed-cost side of a power-gated build. Low conviction: short-horizon relative moves often reverse.
Sep 30	ADYEY	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +9.9% above its 52-week low with 2 of 6 Payments & fintech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	NU	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +6.3% above its 52-week low with 2 of 6 Payments & fintech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	NVO	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +9.0% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	PODD	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +0.5% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	SYK	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +2.2% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	INTU	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +8.0% above its 52-week low with 12 of 30 Software peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Sep 30	IDXX	lags QQQ over 6mo	2	Mar 30, 2027	JGC-990 forward test: +3.8% above its 52-week low with 4 of 11 Health care peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three).
Oct 1	APP	lags QQQ over 6mo	2	Apr 1, 2027	JGC-990 forward test: +0.0% above its 52-week low with 1 of 4 Ad tech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three). Qualifies only under the ad-tech/social basket split (PR #5382, 2026-10-01); the old joint Ads & social basket read it as a group low, so this stake also tests that split.
Oct 1	TTD	lags QQQ over 6mo	2	Apr 1, 2027	JGC-990 forward test: +1.0% above its 52-week low with 1 of 4 Ad tech peers under their 200-day. In 2025-26 none of Sean's 8 buys at a low alone beat QQQ over six months (flat at one and three). Qualifies only under the ad-tech/social basket split (PR #5382, 2026-10-01); the old joint Ads & social basket read it as a group low, so this stake also tests that split.
Oct 3	INTC	lags QQQ over 6mo	2	Apr 5, 2027	Repeat de-rating episode (11/2025, 7/2026): forward EPS +92% over 13 weeks while forward P/E fell 28%. Repeat episodes in ten chip/hardware stocks 2016-2026 trailed (median 6-month return +6% vs +17% baseline; exploratory, 17-18 cases).
Oct 3	LRCX	lags QQQ over 6mo	2	Apr 5, 2027	Second quarter within a year (4/24, 8/07) of forward EPS +12%+ with forward P/E -15%+. In ten chip/hardware stocks 2016-2026, repeat episodes returned a median +6% over 6 months vs +17% for any week (17-18 cases, exploratory split, not pre-registered).

Kind

Right

Total

Hit rate

Avg alpha

Beats index (1mo)

-3.6%

Earnings direction

38%

—

Resolved

Ticker

The claim

Outcome

Verdict

Sep 23

CTAS

UP on earnings

-2.5%

wrong

Called

Ticker

Call

Outcome

Verdict

Oct 11

LULU

stance · PASS

-2.8%

neutral

Oct 11

CRDO

stance · WATCH

0.0%

neutral

Oct 11

CMG

stance · EXIT

+5.4%

right

Oct 11

AMAT