Intel · Strategies
Strategies.
How-to explainers for the strategies Closelook actually uses — covered calls, the barbell, position sizing, hedging, rotation. Each piece pairs the mechanics with current market data and a Closelook-portfolio application. Diary, not advice.
-
The Request Becomes the Transaction
Cloudflare has put a price tag on machine traffic. The rails are real. The demand is not yet proven.
HTTP 402 waited thirty years for a buyer that could use it. Cloudflare's Monetization Gateway prices machine access via x402 — the rails are real, the demand is not. The pricing question turns out to be a positioning question: what do you own that a machine cannot get anywhere else?
-
The Rotation Is Inevitable. The Date Is Not.
Capex → opex → beneficiaries is how every infrastructure cycle resolves — and the market prices the handoff before any income statement confirms it. The strategy view for August through early 2027: which of our three indices leads in each phase, which sectors inside them carry the leadership, and how the book migrates along the calendar.
The switch from AI capex to AI opex to AI beneficiaries is not a hypothesis — it is the standard resolution of every infrastructure build in market history. What is uncertain is when the market prices each leg and along which path. This is the forward map, August 2026 to early 2027, told through the three-index chain — Rubin builds, the Agentic Ecosystem operates, the Agentic Winners use. Short term, the base case is a semi and capex bounce; medium term, the bid bifurcates — average capacity loses it, constraint-solvers keep it. The operating layer is the first index-level bet on agentic uptake; the beneficiaries hold the biggest potential returns from the most punished bases, and are the one layer where only selection works. Phase by phase: the catalysts, the sector leaders, the migration schedule. And the one way this cycle really is different: the capex cycle re-arms with each silicon generation — Hopper, Blackwell, Rubin, Feynman — so the build narrows instead of disappearing.
-
The Agent Is Not the Product. The Orchestrator Is.
Legacy enterprise software is moving from seat-based applications to agentic execution — and every incumbent is racing for the same seat: the control plane that assigns, permissions and meters the work. There are fewer seats than claimants.
The decisive contest in agentic enterprise software is not who builds the smartest agent — models and agents are becoming interchangeable. It is who orchestrates them: which platform decides what agents do, gives them context, controls their permissions and gets paid for the work. SAP, Microsoft, Salesforce, ServiceNow, Oracle and Workday are all racing toward the same architecture, and enterprises will not keep six control planes. The field, the phases, the coalitions — and the knockout ahead.
-
Forward-Deployed Engineers: The Human Infrastructure Layer of Agentic AI
The AI trade is moving from model capability to operating deployment. The clearest public, high-frequency proxy for that shift is not a chip metric — it is a job title.
The forward-deployed engineer is the human bridge between what a frontier model can do and what an enterprise actually runs. How fast that bridge is being built — across the labs, hyperscalers and IT-services firms — is one of the earliest public reads on agentic AI adoption, and the leading indicator behind a new Closelooknet index.
-
Rolling Up and Out: Raise the Cap Without Closing the Trade
You sold the call, the stock ran, and now it's about to be called away. You don't have to let it go — and you don't have to pay to keep it. Done right, the roll is cash-neutral: it lifts your cap, buys you time, and leaves the premium you already banked in your pocket.
You sold the call, the stock ran, and now it's about to be called away. You don't have to let it go — and you don't have to pay to keep it. Done right, the roll is cash-neutral: it lifts your cap, buys you time, and leaves the premium you already banked in your pocket.
-
When You Think the Rally Pauses, Sell Calls — Not Your ETF
Selling a winning core position triggers tax, transaction cost, and the risk you're wrong about getting back in. Writing calls against it monetises the pause you expect while you keep the position. With semis up 126% and overbought, here's the overlay — and when to upgrade it to a collar.
Selling a winning core position triggers tax, transaction cost, and the risk you're wrong about getting back in. Writing calls against it monetises the pause you expect while you keep the position. With semis up 126% and overbought, here's the overlay — and when to upgrade it to a collar.
-
From Apple to Nebius: Matching the Strategy to the Stock and the Vol
Six pieces, one decision. Here's how to route from a name and a view to the right structure — and let the volatility regime, not your mood, set how aggressive the strike. The map for the whole series.
Six pieces, one decision. Here's how to route from a name and a view to the right structure — and let the volatility regime, not your mood, set how aggressive the strike. The map for the whole series.
-
Win Small, Often. Lose Big, Once. The Risk Math Nobody Prices Until They Meet It.
The premium seller's edge is real — and so is the tail that eventually erases a year of it in an afternoon. The standard risk toolkit can't see that tail, because it was built on a curve markets don't obey. With the fear gauge near its lows, this is exactly when the bell curve lies to you.
The premium seller's edge is real — and so is the tail that eventually erases a year of it in an afternoon. The standard risk toolkit can't see that tail, because it was built on a curve markets don't obey. With the fear gauge near its lows, this is exactly when the bell curve lies to you.
-
Gold: Not What It Used to Be
A high-beta momentum asset with positive correlations with global equities
Since autumn 2025 gold has often traded like a high-beta momentum asset — tightly, positively correlated with global equities rather than hedging them. A look at the correlation-regime shift, the cross-asset funding mechanics driving it, and what it means for portfolio construction.
-
Why the Same Greeks Are Calm on Apple and Lethal on Nebius
The Greeks aren't abstract — they're the dials behind every covered call and put in this series. The surprise is that the one everybody watches, delta, is nearly identical on Apple and Nebius. It's the other four that explain why the same trade is a yield on one name and a grenade on the other.
The Greeks aren't abstract — they're the dials behind every covered call and put in this series. The surprise is that the one everybody watches, delta, is nearly identical on Apple and Nebius. It's the other four that explain why the same trade is a yield on one name and a grenade on the other.
-
Sell to Own, or Sell to Never Own — The One Decision Behind Every Put
A sold put is one instrument that does two opposite jobs. Whether assignment is your win or your failure — and which strike expresses it — is the whole game. The same put pays 9% on Apple and 41% on Nebius, and the gap tells you exactly what you're being paid for.
A sold put is one instrument that does two opposite jobs. Whether assignment is your win or your failure — and which strike expresses it — is the whole game. The same put pays 9% on Apple and 41% on Nebius, and the gap tells you exactly what you're being paid for.
-
The Buyer Needs Three Things to Go Right. The Seller Needs One.
Most option buyers lose and most option sellers win — and it isn't luck, it's the structure of the bet. With the fear gauge near its lows and money pouring into cheap calls on the AI run, here's the math that says be the house.
Most option buyers lose and most option sellers win — and it isn't luck, it's the structure of the bet. With the fear gauge near its lows and money pouring into cheap calls on the AI run, here's the math that says be the house.
-
The Barbell Architecture — Why the Middle Is Decay
Two structurally orthogonal poles, zero in the middle. A portfolio shape structurally consistent with the Bessembinder evidence — and where every "balanced" alternative tends to optimise the wrong object.
Most portfolio architectures still in use — Markowitz mean-variance, 60/40, Risk Parity, Dividend Aristocrats, Defensive Equity — were built for a return distribution that no longer dominates. The Bessembinder evidence (64,738 stocks, 43 markets, 31 years) shows compound returns are positively skewed: roughly 2% of firms create 100% of public-market net wealth, the median individual stock loses to cash, the median listed stock in Germany destroyed 31.9% of capital over three decades. A portfolio shape structurally consistent with this evidence is a barbell — a strategic pole of 66-75% in the broadest cap-weighted index vehicles (QQQ + VEU), a tactical pole of 25-34% in signal-driven thesis exposure (Rubin/HALO/Euro-AI/AW25 + Pattern 03), and zero in the middle. A caveat the piece itself flags: the Bessembinder math assumes public markets capture wealth creation, and the current Databricks/OpenAI/Anthropic/SpaceX trend points the other way. This is Part 1 of two; Part 2 covers sizing math and rebalancing rules per leg.
-
Sizing and Rebalancing the Two Legs — Decision Rules for the Barbell
The architecture is the easy part. Sizing the QQQ-vs-VEU split, weighting the four Closelooknet indices inside the tactical pole, choosing the cadence on each leg, and recognising the five failure modes that quietly collapse a barbell back into a balanced portfolio.
Part 1 set the architecture: a strategic pole of 66-75% in QQQ + VEU, a tactical pole of 25-34% in Rubin/HALO/Euro-AI/AW25 + Pattern 03, zero in the middle. Part 2 covers the sizing math and the rebalancing discipline. The strategic pole splits between QQQ and VEU on a roughly 60/40 internal weight, drift-tolerant, not tactically rebalanced. The tactical pole is signal-driven: enter when a wave is in Dawn or Early-Ramp under the Money Temperature framework, weight across the four index sleeves by current-cycle conviction, exit when the wave matures. Single-name Pattern 03 exposure rides on top of the index allocations within the tactical pole's budget. Five failure modes — tactical drift past 34%, premature exit on tactical drawdowns, the "one more story stock" bloat, the 60/40-rebound-to-balance temptation, and single-name concentration without exit discipline — are where most real-world barbell implementations actually break.
-
Covered Calls for Premium Harvest — Quantum, AI, and the High-Beta End
When implied vol is 80%+, a 30-day call sells for 5–10% of the underlying. The income looks irresistible. The drawdown risk explains why most people lose money trying.
On a high-beta name with 80% implied volatility — quantum-computing stocks like IONQ, RGTI, QBTS, recent AI IPOs, small-cap thematic plays — a 30-day call struck 10% out of the money can pay 5–10% of the underlying in premium. That's a 60–120% theoretical annualised yield on the option overlay alone. The catch: the same names move 30% in a week in either direction. This is Part 2 of two; Part 1 covers the slow, conservative end. Here the question is not 'is the premium worth it' (it usually is, mathematically) but 'can you survive the realised-vol path that justifies the implied vol you're selling.'
-
Covered Calls for Income — the Dividend Aristocrats Approach
Stack a 10% out-of-the-money 6-month call premium on top of a 2–4% dividend yield. The conservative end of covered-call writing — how the math works, when it pays, what it costs.
A covered call on a Dividend Aristocrat — say KO, JNJ, PG — at a 10% out-of-the-money strike with 6 months to expiry is the slow, conservative end of options income. The premium is modest, the cap on upside is wide, the underlying is famously dull. Combined with the existing dividend yield, the total cash-return profile lands in the 6–8% annualised range on low-vol blue-chips, with a 10% cap on capital appreciation over each six-month window. This is Part 1 of two; Part 2 covers the opposite end of the spectrum — high-beta quantum and AI darlings where the premiums are massive but the drawdown risk is too.
16 explainers on file.
FAQ · from the current data · as of 2026-08-01
Strategies — Q&A
What is Strategies?
Rule-based strategy notes — posts, not papers — each with the rule, the record and the caveat. Hub: /strategies/.
How often does Strategies publish, and how many editions are there?
As strategies are formalised; 16 editions are archived on this page, the latest "The Request Becomes the Transaction" dated Aug 1, 2026.
Is this investment advice?
No. Closelook publishes an investment research diary — human-reviewed, AI-supported — not investment advice. Past performance is no indication of future performance.