MRVL

Two forward signals deteriorating: momentum collapsing and thesis has slipped to middling. This is exactly the CRWD-pattern our learnings flagged — holding through 60s scores. Cut on evidence; the setup from here is weak.

Win-probability reasonable at 68 but thesis rank is only middling and momentum at 31 signals the market is not confirming. Stock is 43% off highs — that's a broken chart, not a bargain setup. Catalyst is only sector-tailwind, not ticker-specific. Not enough forward edge to add.

Held name with a reasonable forward setup — Win-Prob 68 is respectable and the Meta custom silicon catalyst reads through to MRVL as a key ASIC partner. But thesis rank at 57 is middling and momentum is soft. Constructive but not core; hold at current view of the stock going forward.

Solid win-prob and decent risk quality on a custom ASIC play with hyperscaler customers, stabilizing after a deep drawdown. Thesis percentile is mid-pack in this batch of AI chip names. Constructive but not core; hold the existing position, don't chase.

Constructive but not core. Win-probability is the strongest signal — recovering from lows with Amazon/Google/Microsoft ASIC business intact, and HBM-waning narrative supports custom silicon relevance. Thesis rank at 57 is only middling versus batch. Hold-size rather than add; the recovery is real but not the highest-quality forward setup in this group.

Win-probability holds up at 68 with recent +6.68% weekly move and constructive consolidation. But thesis percentile at 57 is middling versus the AI-chip cohort here. Position_type is watchlist per input but risk specialist notes it's held — treating as hold-worthy on forward setup, not a strong add candidate.

Held name with decent win-probability and constructive fundamentals — custom ASIC design wins with hyperscalers intact. But momentum at 78 is getting extended and thesis has slipped to mid-pack as newer AI names rose. Not a deteriorating story, just no longer core. Hold the stock at moderate conviction; sizing is the Sizer's job.

Win-Prob at 68 is the strongest specialist signal here, and per our weighting rule this matters more than a middling thesis rank. Mid-range position (43% off highs), decent intrinsic quality, and TSMC/AI silicon catalyst backdrop supportive. Constructive but not core — held-size hold.

Two of four specialist signals are actively soft — thesis mid-tier, win-prob near coin-flip, and momentum the weakest in the batch. Catalyst mentions are indirect competitive-analysis pieces, not company-specific drivers. Fails both convergence gates for new_buy. Forward return picture is uninspiring.

Watchlist name with mediocre thesis, sub-coinflip win-probability, and weak momentum. Nothing here converges. The 'battle of AI compute' article naming isn't a catalyst — it's coverage. No reason to enter here; needs an earnings-beat catalyst before re-engaging.

Two specialists flag deterioration: win-probability slipped to 52 with negative momentum, and catalyst coverage lacks any MRVL-specific news (only NVDA-adjacent). Thesis rank dropped to 57 in batch context. Forward setup is the weakest of the held names here — trim on evidence, don't anchor to prior conviction.

Two specialists signal deterioration: thesis has slipped to middling percentile and win-probability is at 52 with negative momentum and a fresh 7% five-day drawdown. Sector tailwinds exist but the stock itself is not participating. Forward view here is weak — trim on evidence, not because of any P&L anchor.

Two specialist signals softening: win-probability slipping to 52 and momentum weak at 26. Thesis has dropped to mid-pack. Forward setup is fading — stock needs a catalyst to break higher and none is imminent. Not a broken thesis but clear deterioration versus the batch. Lean negative.

Two specialists signal deterioration: thesis dropped to mid-tier, win-prob at 52, momentum in bottom third with -3% 5d. No obvious near-term catalyst to break the drift. Forward view is weak regardless of what we're holding. Trim on evidence — this is exactly the pattern the learnings flagged (MRVL -16% memory).

Two specialists signal deterioration: Win-Probability is barely coin-flip and momentum has rolled over with the stock 28% off highs. Thesis has slipped into mid-pack and there's no direct catalyst driver. Forward view has weakened meaningfully — score the stock, not the trade.

Middling scores across the board with meaningful short-term weakness and no company-specific catalyst. Forward setup is neutral-to-soft. Not a conviction name from here; holding rather than exiting because the AI custom-silicon thesis remains intact, but the score reflects mediocrity, not our cost basis.

Momentum has rolled over hard (-9.84% week) and thesis rank has fallen to mid-pack. Two forward signals softening at once — technical break and mediocre win-prob — while valuation stays demanding. Forward view has deteriorated relative to when this became a core position. Not an exit, but conviction is negative-lean from here.

Thesis has slipped to mid-pack, momentum is flat-to-negative, and the stock is showing near-term distribution (-9.84% week). No MRVL-specific catalyst, just sector-adjacent read-throughs. Forward setup is neutral-to-negative from here; conviction fades.

Forward view has softened — thesis percentile mid-pack, meaningful near-term price weakness, stretched multiple needs execution. Not a breakdown yet, but no longer a leadership name. Neutral hold-level conviction; let the Sizer decide whether to trim on portfolio grounds.

Held name showing meaningful near-term weakness with negative momentum and mid-tier thesis. Catalyst piece is supportive but framed as contrarian. Not an active deterioration on multiple specialists — just tepid convergence. Hold as a lower-conviction position; Sizer decides trim depth.

Two specialists are flagging weakness: momentum is deteriorating hard (-9.84% 5d) and win-probability sees vulnerability at 84x PE with negative technicals. Thesis has slipped meaningfully from prior batches. This is deterioration on the forward signals — score the stock, not the trade. Trim.

Setup is deteriorating on the margins: momentum is soft, no MRVL-specific catalyst, and the thesis percentile has drifted to mid-pack. Not broken enough to force an exit — AI custom silicon narrative still intact — but conviction is neutral-to-soft from here. Held status justifies hold rather than trim; the Sizer can decide whether to lighten.

Past winner where thesis rank has cooled and momentum has weakened, but the catalyst read (hyperscaler ASIC contracts) directly confirms the original bull case. Hold; not a name to add on today's forward view without a clean breakout.

Forward signals are middling across the board — mediocre thesis rank, sub-60 win-prob, and weak momentum with a 5-day pullback. Not a breakdown but not a name I'd add. Hold and let the Sizer decide whether to trim.

The MRVL breakout that delivered our win earlier has clearly rolled over — momentum 39, thesis slipped to 57, down 5% in 5 days, no direct catalyst. Win-prob 58 is mediocre. This is not deterioration hard enough to force an exit (thesis intact, risk quality decent), but the forward setup has weakened materially. Hold, don't add.

Momentum still confirming and the custom-silicon/ASIC narrative remains the right side of the AI infrastructure trade. Thesis percentile has slipped relative to fresher names and win-prob is only 58, but the forward setup is still constructive. Hold-quality conviction, not add-quality.

Mid-tier conviction here. The custom-silicon thesis is alive and the stock is in a normal pullback rather than a breakdown — no two-specialist deterioration signal. Forward view remains constructive but not core. Hold.

Win-probability is the highest in the batch tied with AVGO at 72, and momentum is confirming with a 7%+ weekly move into breakout territory near $330. Per our learnings, AI-hardware names breaking 52w highs on hyperscaler ASIC narratives are exactly the pattern to trust. Thesis percentile slipped to 60 as other names rose, but the forward setup from here is strong. Hold and let it work.

Approaching 52w high on a 7% weekly move with real ASIC win narrative — the kind of price/catalyst combo we've learned to lean into. Win-probability is the strongest of the held names and the setup mirrors our prior MRVL winner. Forward expected return remains attractive.

Win-probability is the highest in the batch tied with AVGO, and price action is constructive heading into breakout territory at 52w highs. Thesis percentile slipped as newer AI names rose, but the forward setup — momentum confirming, catalysts in supportive narrative — keeps conviction strong. Pattern-match to prior MRVL winner.

Win-probability is the highest tier-2 in the batch and price is pressing the 52w high with strong 5-day momentum — the classic breakout-confirming setup from our winners playbook. Thesis percentile has compressed as newer names rose, but the forward setup from here remains strong. Solid hold.

Win-probability is the highest weight in our framework and Marvell's at 72 with a strong recent move toward 52wk-high breakout territory. Thesis percentile slipped to 60 but the forward setup on custom silicon and hyperscaler capex remains intact. Constructive hold.

Win-probability leads the call here — strong 5d move and breakout territory with ASIC/optical DSP earnings revisions accelerating. Thesis percentile has slipped relative to batch but the forward setup is constructive. Hold the position; let the breakout play.

Win-probability 72 plus strong momentum approaching 52w high — the exact pattern our biggest winners showed. Thesis percentile is only middling but the forward win-rate signal and price action are what matter here. Constructive setup; let it run.

Win-prob and momentum both supportive on the custom-silicon ramp. Thesis percentile has slipped as fresher AI infra names rose, but the forward setup from here is still strong. This is exactly the MRVL pattern from learnings — hold the winner, let it work.

The custom ASIC ramp (Trainium, Google TPU exposure) plus strong momentum (5d +7.27%) and supportive 'Beyond Nvidia' narrative matches our winning pattern: AI-infra hardware breaking higher on named hyperscaler design wins. Win-Probability solidly 70. Forward setup remains strong; how much to size is the Sizer's call.

Best momentum in batch outside ARM/TSM, with a credible custom-silicon ramp story (Trainium, TPU) and supportive sector catalysts. Win-prob plus the price action mirror the playbook that worked. Forward setup remains strong.

Highest momentum in batch with the stock breaking out and a clear forward custom-silicon ramp story (Trainium, TPUs). Win-prob 70 and rising analyst revisions. This is exactly the breakout-AI-infra pattern our recent learnings tell us to lean into. Forward setup remains strong.

Strong momentum confirming a real ASIC design-win story with Trainium/TPU exposure. Win-probability solid and forward catalysts through 2025 are concrete. Thesis percentile slipped as newer infra names rose but the forward setup remains attractive.

Custom ASIC ramp (Trainium, TPU) intact with a specific forward product catalyst on optical interconnects sampling this year. Momentum confirming near highs, analyst revisions trending up. Forward setup remains strong even after the run. Thesis pct slipped only because newer names crowded the top; the AI-silicon ramp story is unchanged.

This is the exact pattern we got right before — AI-infra silicon with hyperscaler design wins and confirming momentum. Thesis percentile drifted to 60 but forward setup remains strong: custom silicon ramp through 2025, positive revision trajectory, 14% below highs (not pinned). Hold conviction stays high.

Win-prob has collapsed and the technical setup is broken with a severe 5d move down. Thesis percentile has also slipped to mid-pack. Two specialists effectively signal deterioration — forward setup no longer justifies the held weight. Trim.

Two specialists signaling deterioration: win probability collapsed to 38 and there's no direct Marvell catalyst — only sector-tailwind inference. The -16.7% 5d move is a real technical breakdown on a 91x PE name. Forward view from here is weak regardless of prior winner status. Trim.

Protect-winner rule applies: +36% from entry. Only one specialist (win_prob) signals active breakdown; thesis, catalyst, and risk remain constructive. Per core principle 1, do not cut conviction below 75 without 3-of-4 breakdown. Hold the winner, let it consolidate.

Per protect-winners rule, MRVL up 38% gets conviction floor of 75 unless 3 of 4 specialists signal breakdown — here only win_probability is clearly negative; thesis, catalyst, and risk remain constructive. Recent 5d weakness is concerning but performance and momentum percentiles remain elite. Hold the winner; do not trim on one bad week.

Protect-winner rule applies: up 30% from entry. However, only one of four specialists (win_prob at 38) signals active breakdown, and the breakdown is technical not thesis-based. Performance and momentum scores remain strong. Per rule 1, do not cut below 75 unless 3 of 4 specialists deteriorate — we're at 1. Holding the winner; sizer can manage size given recent volatility.

MRVL is up 25% from entry — winner protection rule applies. Only one specialist (win_prob) signals breakdown; thesis, catalyst, and risk_quality remain constructive. Recent 5d weakness is real but doesn't meet the 3-of-4 breakdown bar to cut a winner. Hold the position; do not chase further but do not rotate out of a +25% AI infrastructure name into choppy alternatives.

Up 22% from entry — winner protection rule applies. Only one specialist (win_prob at 38) is flashing breakdown on the -16.7% 5d move; thesis, catalyst, and risk all remain constructive or neutral. Per our learning that MRVL-type winners deserve top-tier weight, we hold conviction in the high-70s despite the technical wobble. Do not rotate out of a +22% winner on momentum noise alone.

PROTECT WINNERS rule applies: held position up 45%+, so conviction floor is 75 unless 3 of 4 specialists signal active breakdown. Here only win_prob (38) is clearly negative; thesis (60), catalyst (60), and risk_quality (70) are constructive-to-neutral. Recent technical pullback is profit-taking after a massive run, not thesis breakdown. Hold the winner.

Winner up 28.5% from entry — PROTECT rule applies. Thesis percentile slipped to 60 as newer names rose, but no specialist signals breakdown; win_prob still 68 and price consolidating near highs. Custom ASIC thesis intact. Hold, do not trim a working winner for marginal rotation.

Classic winner to protect — up 28% from entry with momentum and performance at the top of the batch. Win-prob 68 and bull thesis intact. Do not let lower thesis_pct (60) drive a trim of a clear winner. Hold.

Textbook winner: +40% since entry, momentum at 93rd percentile, performance at 88, all four specialists constructive. Per core principle #1, do not reduce conviction below 75 on a +20%+ winner without 3-of-4 breakdown signals — and we have zero. Protect this. Let it run.

Textbook protect-the-winner: held position up 53.6% with momentum at 98th percentile and direct positive catalyst (Jensen calling out MRVL on AI infra). Per core principle 1, hold conviction high. Sizing concerns are for the portfolio sizer, not me.

MRVL up 49.5% since entry — squarely in protect-winner territory per Principle 1. Three of four specialists constructive (win_prob 68, catalyst 65, momentum 96, performance 93); only thesis percentile slipped to 60 because the watchlist refreshed, not because the AI ASIC story broke. No active breakdown signal. Hold high conviction. This is exactly the lesson from Micron/Palo Alto: keep scores pinned high on confirmed leaders.

MRVL up 36% since entry — a clear winner. Three of four specialists constructive: win-prob, momentum, performance all strong, catalyst supportive. Thesis percentile dropped only because newer AI names rose in relative ranking, not because the MRVL thesis broke. Core operating principle #1: protect winners. Do not trim.

Held winner near 52w highs with strong momentum and constructive win-prob. Thesis percentile moderate (60) but the custom ASIC story is intact. Protect the position; no reason to trim given performance and momentum.

MRVL is the canonical 'protect the winner' case from our learnings — held at +24.7%, win_prob 72, performance percentile 78, catalyst directly aligned (optical/HBM/networking). All four specialists constructive. No reason to trim.

Held winner up 24.7% since entry — winner-protection rule applies, floor at 75. All four specialists constructive, direct ticker coverage in catalyst flow, momentum strong. This is exactly the pattern we want to ride. Protect.

MRVL is a proven winner up 23% from entry with three of four specialists constructive. Per protect-winners rule (>20% gain, no breakdown signal), conviction stays well above 75. Custom ASIC/AI networking thesis intact, momentum strong. Hold the winner.

MRVL is a +22% winner — protect-winner rule applies. Win-prob 72, momentum and performance both in upper quartile, near 52wk high breakout setup. No specialist signals breakdown. Lack of direct catalyst today is noise; this is exactly the kind of winner our learnings tell us to run, not trim.