Management Now Calls the Cycle to 2030, Then Sells Two Refineries at a Mid-Cycle Price

Results vs. Consensus

Published estimates for the quarter ranged from $24.47 to $26.01 of EPS. The table anchors on the Zacks $25.74, the same vendor we used for the second quarter's $14.52 and the figure carried by the most outlets. Against the $25.37 carried by several other outlets the beat was 6.1%. Against the top of the range the beat was 3.5%; against the bottom, 10.1%. The revenue consensus is less reliable, with one aggregator still at $33.56B on stale contributor estimates, so we use the $46.72B mark against total revenues and other income. Segment-level estimates come from a smaller panel of four analysts.

The shape of the beat is the story. It came from volume and from the Mid-Continent. Throughput beat by 4.7% and crude runs by 5.3%, while the system margin beat by only 2.1%. The West Coast, which the Street expected to be the highest-margin region by a distance, missed its margin estimate by 13.2%. In the first two quarters of our coverage the beat came from the crack outrunning the models. This quarter the models had the crack, and MPC beat them on barrels.

Year-over-Year Comparison

Two lines need context. The fall in equity-method income is a prior-year artifact: Q3 2025 included $738M of gains on the BANGL acquisition and the ethanol joint-venture sale, which is also why last year's GAAP EPS of $4.51 sits well above its adjusted $3.01. And the noncontrolling-interest line, MPLX's public unitholders, took 5.5% of consolidated net income against 29.5% a year ago. The minority claim is roughly fixed in dollars, so its share collapses whenever refining earns, which flatters MPC's per-share leverage to the crack in both directions.

Sequential Comparison

The R&M margin rose $9.37 per barrel and segment EBITDA per barrel rose $9.31, so essentially all of the margin improvement reached EBITDA. That is the cost discipline showing up. Capture fell 19 points in the same quarter, which is possible only because the indicator rose faster than MPC's realized margin. The company earned more and captured less, and both statements are true.

Nine-Month Comparison

Quality of the Beat

• Revenue: Price first, but volume contributed this time. Net refinery throughput rose 4.8% to 3,150 mbpd and refined product sales volume rose 6.5% to 4,030 mbpd, while the 10-Q attributes the R&M segment's $20.45 billion revenue increase primarily to an average refined product price $1.14 per gallon higher and to 247 mbpd more refined product sales. There is no acquisition contribution in the refining top line. The Q2 pattern, revenue up on falling volume, did not repeat.
• Margins: Entirely market-driven at the gross level, and better than that at the cost level. The 10-Q estimates the benchmark effect at about $9 billion of year-over-year margin and MPC's own variances against the benchmark at a net $700 million negative, so more than all of the improvement came from the crack. What MPC controlled went well: refining operating cost of $5.60 per barrel was flat year over year on 145 mbpd more throughput, and turnaround expense fell $106M.
• EPS: One adjustment, modest and defensible. A $123M pre-tax charge for a municipal tax dispute with the City of Carson covering several prior years, plus legal settlements from a single prior-year event, is added back, worth $0.34 a share after tax. The tax rate rose to 22.5% from 19.1%, the diluted share count fell 7.9%, and the minority share of net income fell to 5.5% from 29.5%. The last item is structural leverage, not engineering.
• Cash: Operating cash flow before working capital was $8.5 billion, the durable number. Working capital added another $2.3 billion, driven by income tax payables and the timing of the SPR crude exchanges. Both reverse: the taxes will be paid, and the 10-Q shows roughly 26 million barrels owed back to the reserve between April 2027 and July 2029, carried as a $1.9 billion liability.

Segment Performance

The mix moved further toward the cyclical half. A year ago refining and midstream each contributed about half of segment EBITDA (51.6% and 50.0%, with Renewable Diesel negative). This quarter refining is 82.2% and midstream 15.6%, against 76.6% and 20.5% in Q2. Midstream grew faster than at any point since our initiation, and its share still fell by about five points. That is the clearest single illustration of how far the refining margin has travelled from its own base: the stable half of the business is now a rounding error on the consolidated income statement.

Refining & Marketing

Segment adjusted EBITDA of $9,895M against $1,762M a year ago, on an R&M margin of $45.70 per barrel against $17.60. The benchmark did the work. The blended 3-2-1 crack the company publishes averaged $48.67 in the quarter against $17.66 a year earlier, and the 10-Q estimates the market-indicator effect at roughly $9 billion of year-over-year margin on its own. Crude capacity utilization was 100% against the roughly 94% implied by the August guide, on crude runs of 2,973 mbpd against a guided 2,820.

The disclosure that changed direction is capture, which fell to 93% from 112% in Q2. The CFO named the cause before anyone asked.

"Turning to slide 7, third-quarter capture was 93%. Secondary products were a material headwind, the largest we have seen on an absolute basis to capture." — Maria Khoury, EVP and CFO

The 10-Q puts a number on the aggregate. Against the prior-year quarter, crude mix, market structure, RIN prices, yields and other feedstock variances had "an estimated net negative effect of approximately $700 million" on segment adjusted EBITDA. For the first nine months the same line is a net positive of roughly $200 million. Purchased RIN expense alone rose to $654M from $212M.

Assessment: Ninety-three percent is below the 95% third-quarter average management volunteered in August, so the seasonal warning it gave was, if anything, too gentle. The cause is credible: when clean products rally harder than coke, sulfur, asphalt and fuel oil, every refiner's secondary barrels lag the indicator, and a 100%-utilization system makes more of them. What the quarter does not support is the reading we gave capture in August, that a structural step-up had occurred. Year-to-date capture is 101%, roughly the level the business ran at before the dislocation. The 112% was a quarter, not a regime.

Regional Performance

The regional scoreboard read well everywhere for a second consecutive quarter, and better than in Q2 in every region. Every region cleared $32 per barrel of segment EBITDA, against a FY2025 system average of $5.63.

Gulf Coast

Segment EBITDA of $32.77 per barrel against $4.50 a year ago, at 99% utilization. This is the region that carried the turnaround: $1.49 per barrel of planned turnaround expense against $0.60 a year ago, concentrated on conversion units, which is exactly the mechanism the CFO flagged in August as a capture headwind. Refining operating cost fell to $4.70 from $4.83, still the lowest in the system. Sour crude rose to 61% of the regional slate from 54%, and distillate yield rose to 519 mbpd from 460.

"Our ability to respond to market signals and maximize distillate yields was enhanced by the added flexibility from our Garyville investment that came online earlier this year." — Maria Khoury, EVP and CFO

Assessment: The Gulf Coast earned the least per barrel of the three regions for the first time in our coverage, and it is the only one where a controllable cause is visible: conversion-unit downtime that limits upgrading and pushes barrels into the low-value secondary pool. That is a timing cost, not a structural one, and management maintained its turnaround plan rather than deferring scope into a strong market. It also matters that the region running the heaviest, sourest slate in the system (61% sour) still held $32.77 per barrel through planned conversion downtime. Management is now steering the divestiture proceeds into this region, so its returns are about to matter more than its share of today's EBITDA.

Mid-Continent

The best region in the system at $36.00 per barrel, against $8.93 a year ago and $20.96 in Q2. Utilization was a record 101% and crude runs rose to 1,196 mbpd from 1,147. The cost line we flagged in August corrected cleanly: refining operating cost fell to $5.39 per barrel from Q2's $6.31, only $0.07 above the year-ago $5.32, and turnaround expense fell to $0.82 from $1.86 a year ago and $1.93 in Q2.

"In MidCon, we achieved record utilization of 101% and delivered $36 of adjusted EBITDA per barrel. Higher rates and maximized distillate yields were supported by our Robinson investment that came online in the second quarter." — Maria Khoury, EVP and CFO

Assessment: In August we said the Mid-Continent recovery was a price recovery, not an operating one, because volume fell and cost rose. This quarter answers that directly: volume up, cost back to last year's level, turnaround burden halved. We withdraw the reservation. The irony is the timing. In the same week that its Mid-Continent system posted the highest per-barrel EBITDA in the company, MPC agreed to sell two of its Mid-Continent refineries. We address that decision in Key Topics; the regional data is the reason it deserves scrutiny rather than applause.

West Coast

Segment EBITDA of $33.30 per barrel against $5.08, on the highest regional margin at $47.46 and a record 98% regional utilization. Turnaround expense almost vanished, at $0.30 per barrel against $2.54 a year ago. Heavy fuel oil yield fell to 32 mbpd from 53 and distillate rose to 227 mbpd from 202. Refining operating cost rose to $8.30 from $8.01, the only region where it went up materially.

"In the West Coast, we ran at 98% utilization, at record levels for this region, delivering over $33 of adjusted EBITDA per barrel." — Maria Khoury, EVP and CFO

Assessment: West Coast economics remain the most exposed to reversal. The commercial organization described advantaged California crude availability created by "recent closures" and import economics hampered by high freight and supply disruptions, both of which are features of the current dislocation rather than of the asset. The region earned a sixfold improvement on a cost base that is still the highest in the system and still rising. It is the clearest case in the portfolio of other people's problems becoming MPC's margin.

Midstream

Segment adjusted EBITDA of $1,882M against $1,709M, up 10.1%, the fastest growth since our initiation and an acceleration from 8.3% in Q2. This time the physical data supports the dollars: four of the five disclosed throughput metrics rose, against two of five falling in Q2.

"And we have no different expectation that our fourth quarter should actually be stronger than the third quarter. And I also shared earlier today on our MPLX call that we expect 2027 EBITDA growth to be stronger than we have in 2026." — Maria Khoury, EVP and CFO

Assessment: The arithmetic we set out in August now works. Nine-month Midstream EBITDA is $5,258M against $5,070M, up 3.7%. A 5% full-year outcome needs about $1,830M in the fourth quarter, 8.9% above the $1,680M of Q4 2025. Management says Q4 will exceed Q3's $1,882M, and if it merely matches Q3 the year lands at +5.8%. Two caveats survive. Natural gas processed still fell, which is the Rockies divestiture, and the release again lists rates, equity affiliates and acquisitions as drivers without splitting them. And the new item in the release is that incentive compensation is now a named cost in Midstream as well as in Corporate.

Renewable Diesel

Segment adjusted EBITDA of $254M against $(56)M, a $310M swing, and essentially flat on Q2's $258M. Renewable Diesel margin rose to $304M from $8M; other income, which includes the Martinez joint venture's equity earnings, rose to $54M from $25M. The release attributes the result to "a stronger margin environment, improved regulatory credit values, and higher throughputs."

Assessment: Nine-month segment EBITDA is $550M against a $(117)M loss a year earlier, and the commercial organization said early fourth-quarter conditions look like "more of the same." Nothing here changes our August read: this is a policy-window result in a segment management is deliberately not growing, and that is the right posture. The fourth quarter carries a $65M turnaround at Dickinson and 85% guided utilization, so the run rate will dip for a known reason.

Corporate

Corporate expense of $328M against $238M a year ago and a $260M guide, which included $30M of depreciation. That is a $68M miss on a guided line, the third consecutive miss, and the largest of the three. The difference from Q2 is that the release now explains it: "The variance was primarily due to higher expenses related to incentive compensation." The 10-Q sizes it at roughly $71M of performance-based compensation and bonuses within a $90M year-over-year increase.

"On the call, Chris mentioned that we could have a headwind or a tailwind coming from executive compensation, and that is partly because that executive compensation is tied to the performance of the MPC stock. So as it increases or decreases, there is some mark to market that is necessary." — Maria Khoury, EVP and CFO

Assessment: This confirms, in the company's own words, the mechanism we described at initiation. Part of Corporate cost is a mark-to-market on the share price, so the line rises in the quarters the stock rises and the guide, set three months earlier, cannot anticipate it. The disclosure improved, and we credit that. The economics did not: consolidated SG&A rose 27.5% to $1,100M, and the 10-Q attributes about $150M of the increase to incentive compensation across the company. The fourth-quarter guide is $300M including $60M of depreciation. Excluding depreciation that is $240M, which is $62M below the $302M the line just delivered on the same basis. We would model above it.

Product Yields: the Q2 Result Did Not Repeat

In August we called the collapse in heavy fuel oil yield the most important operating disclosure in the release, said it should update the thesis, and raised our clean-product yield assumption with high confidence. We did not hedge it as a single quarter. This was the quarter that tested it.

Heavy fuel oil rose back to 78 mbpd, 2.4% of system yield, from 29 mbpd and 1.0% in Q2. It remains below the 102 mbpd and 3.4% of a year ago. Distillate share held at 37.9%, identical to Q2 and above 36.3% a year ago. Sour crude fell to 45% of throughput from 48%.

Assessment: We overstated the Q2 yield result. With Q1 at 125 mbpd, Q2 at 29 and Q3 at 78, the 29 now looks like the outlier, a quarter with almost no conversion-unit downtime, rather than a new floor. The defensible claim is narrower: on a year-over-year basis the system is making less residual fuel and more distillate from a sourer slate, and the jet and distillate projects already in service are visible in the distillate line. That is real and it is modest. It is not the "decisive evidence" we called it in August, and the third bull pillar's status below reflects that correction.

Key Topics & Management Commentary

Overall Management Tone: The most confident call of the three we have covered, and the least specific where specifics mattered most. Management extended its macro horizon by three years, announced a divestiture, a dividend increase and a new authorization, and was precise on utilization, crude sourcing and project timing. Asked four separate ways about the cash pile, the proceeds and capital return, it again restated priorities rather than describing a method, and nobody asked it to reconcile a decade-long constructive call with the price it accepted for two refineries.

1. The horizon moved from "into 2027" to "the end of the decade"

In August management described "an enhanced mid-cycle environment through the end of the year and into 2027," and when invited to endorse a structurally higher mid-cycle for a decade, answered with 2027 and said the timing of resolution was unpredictable. Three months later, the same question did not need to be asked.

"We expect to remain in an enhanced mid cycle environment through the end of the decade with elevated margins through at least mid-2027" — Maryann Mannen, Chairman, President and CEO

The statement has two parts and they should be read separately. "Elevated margins through at least mid-2027" is a near-term call on the dislocation. "Enhanced mid cycle environment through the end of the decade" is a structural call that the post-dislocation base is above the old one. The supporting evidence offered was inventory and demand: diesel stocks below the five-year range, gasoline near the bottom of it, and no demand destruction in MPC's own system.

"But frankly, we have not seen really demand destruction across our system, particularly when you look at jet and diesel." — Maryann Mannen, Chairman, President and CEO

Assessment: This is the largest change in management's stated view across our three quarters of coverage, and it moves directly against the first bear point. A management team that refused a decade framing in August now asserts one. We take it seriously because this team has been conservative in its public macro language. We also note what it is not. It is not a margin level, a mid-cycle EBITDA figure, or a reversal scenario, and it arrived on the same morning management announced the sale of 5.9% of its crude capacity at a price the buyer says was negotiated before the conflict (Topic 3). A stated view and a transacted price are both evidence. The transacted price is the harder one to fake.

2. The outage measure changed for a third time

The most important number in our thesis has now been described three different ways in three quarters. In May it was roughly 6 million barrels per day, about 6% of global refined product capacity, attributable to the conflict. In August it was over 9 million barrels per day of total planned and unplanned downtime, roughly 4 million above historical norms. This quarter:

"Globally, there are 10 million barrels per day of planned and unplanned refining capacity offline. This is approximately 3 million barrels per day above its historical norms reflecting ongoing Persian Gulf disruptions and Ukrainian attacks on Russian infrastructure." — Maryann Mannen, Chairman, President and CEO

Total downtime rose by about 1 million barrels per day and the excess above norm fell by about 1 million. The implied historical norm therefore moved from roughly 5 million to roughly 7 million barrels per day in one quarter. That is plausibly seasonal, since autumn is turnaround season and the norm for October is higher than the norm for July, but management did not say so and nobody asked.

Assessment: On management's own current measure, the abnormal outage shrank by a quarter, from about 4 million to about 3 million barrels per day. That is a de-escalation, and it happened in the same quarter that management lengthened its horizon. The two can be reconciled (the incremental supply loss can fall while inventories stay depleted), but the reconciliation is ours, not the company's. Three quarters in, MPC still has not published a bridge between its outage measures or any estimate of what the business earns when the excess returns to zero.

3. The divestiture: what MPC sold, and at what price

MPC signed an agreement on November 2, announced the next morning, to sell its St. Paul Park, Minnesota and Mandan, North Dakota refineries, with related marketing and midstream assets, to CVR Energy for $4.2 billion plus working capital. The refineries carry 177 mbpd of crude capacity (105 and 72 mbpd), 5.9% of MPC's 2,986 mbpcd system. MPC receives $3.45 billion. MPLX receives $750 million, which the release describes as "a 9x multiple on forecast 2027 adjusted EBITDA of the associated logistics assets." Closing is expected at year-end.

MPC did not disclose what the assets earn. The buyer did, and its figures give two very different multiples for the same price. Adjusted to the buyer's accounting, which capitalizes turnaround costs and therefore flatters EBITDA relative to MPC's expensing, the business earned about $1.2 billion of EBITDA in the twelve months to September 30, 2026, $1.1 billion of it in the first nine months of 2026, and $506 million in FY2025. The $4.2 billion price is therefore roughly 3.6x trailing EBITDA and roughly 8.3x FY2025. Both multiples carry a caveat: the buyer's figures cover the refineries and the marketing business and do not mention the logistics assets for which MPLX receives $750 million, so they may set a price that includes those assets against earnings that do not. The buyer also said on its own call that negotiations began before the Iran conflict and that it underwrote mid-cycle rather than current earnings. The 10-Q gives the carrying value of the net assets, across the R&M and Midstream segments, as $1.8 billion, so the consolidated pre-tax gain is roughly $2.4 billion, part of it in MPLX's $750 million piece, and the CFO confirmed the refining proceeds are taxable.

"We acted on an opportunity to align our portfolio with our strategy at valuations that I think we deem to be very attractive." — Maryann Mannen, Chairman, President and CEO

The strategic rationale is coherent on its own terms: two small, isolated northern-tier plants with an average complexity of about 7.8, sold to an operator that wants scale in the region, with the proceeds directed to the Gulf Coast. Management has been talking about portfolio fitness since 2024, and these are the two plants in the system furthest from the Gulf Coast value chain it now emphasizes.

Assessment: On a normal year's earnings, 8.3x is a full price for two peripheral plants, and we do not think MPC sold badly. The more interesting point is what the price says about management's own cycle view. MPC transacted at a mid-cycle valuation, negotiated before the conflict, and handed the buyer all of the war upside plus whatever "enhanced mid cycle" follows it. A seller convinced that margins stay enhanced to 2030 had every reason to reprice or to wait; this one announced a sale at a pre-war number on the same morning it published its most constructive macro call. Both can be true, since a portfolio decision taken before the conflict can survive a change in the cycle. But the transacted price is the harder evidence, and it points to a management team that values a refining barrel at mid-cycle when it is the one being paid.

4. Where the proceeds go: more refining capital, not less

Management was explicit that the sale does not signal retreat from refining. The proceeds are earmarked for the Gulf Coast, and the 2027 capital budget, due with the fourth-quarter results in February, is likely to be above the roughly $1 billion of annual refining growth capital of recent years.

"We really are not intending to reduce or shrink our refining footprint." — Maryann Mannen, Chairman, President and CEO

The project list already in flight was restated: the Galveston Bay 90 mbpd distillate hydrotreater, Garyville feedstock optimization (30 mbpd of incremental crude throughput) and Garyville export gasoline flexibility (10 mbpd), all for year-end 2027. The refining organization described further Gulf Coast work in "additional conversion, yield enhancements, and throughput." Management also restated the capital record: about 175 mbpd of capacity added over five years through capital already deployed.

Assessment: Selling 177 mbpd at roughly 8.3x normal-year EBITDA and redeploying into Gulf Coast conversion is a defensible swap if the new projects clear the 25% refining hurdle management has stated in prior quarters. For the fourth consecutive quarter, management has not published a realized return on a single completed project, and Garyville jet flexibility has now run for three quarters. The capital-allocation question we have asked since May has therefore become larger, not smaller. There is now a $3.45 billion pre-tax pool to deploy, on top of $12.3 billion of standalone cash.

5. Capture at 93%, and what the metric does and does not measure

Capture fell to 93% from 112%, below the 95% third-quarter average management gave in August. Management blamed secondary products, a headwind it called the largest in the company's history, plus weaker jet margins and seasonal inventory builds, partly offset by stronger clean-product margins and a wider diesel-gasoline spread.

"Year to date, frankly through the third quarter, we are at capture of 101%." — Maryann Mannen, Chairman, President and CEO

The more useful disclosure came from the commercial organization and concerned the definition, not the number.

"However, I want to remind you that those differentials they are referenced in our market indicator. As well as backwardation is referenced in our marketing indicator. Unlike others, they put that in their capture results." — Rick Hessling, Chief Commercial Officer

Assessment: This is an important clarification for anyone who used capture, as we did, to measure MPC's crude-sourcing edge. If the heavy-sour and inland differentials sit in the indicator, a wider WCS discount raises the indicator and the margin together and leaves capture roughly unchanged. Capture therefore measures execution against a crude-adjusted benchmark, not the value of the crude advantage itself. That makes the 93% a cleaner execution miss than it looks, and it means the third bull pillar needs evidence other than capture. Early fourth-quarter data did not help: the secondary-product headwind "through October" was the same as in the third quarter.

6. One hundred percent utilization, with the turnaround plan intact

Crude utilization reached 100%, a third-quarter record, with records in the Mid-Continent (101%) and the West Coast (98%). It was done without deferring maintenance.

"With total throughput of nearly 3.2 million barrels per day, our refineries ran at 100% utilization. We achieved this while executing our planned turnaround activity, maintaining our commitment to safe and reliable operations." — Maria Khoury, EVP and CFO

Answering a question on costs and the 2027 turnaround outlook, the CEO added a point nobody had asked for directly: "We did not delay or postpone any turnaround activities in 2026." The full-year refining turnaround plan of $1.35 billion was reaffirmed, and the refining organization said 2027 should be a lighter turnaround year than 2026.

Assessment: This is the strongest operating quarter MPC has reported in our coverage. Utilization beat the guide by six points while turnaround expense came in within $4M of plan, which is evidence that the reliability capital of the last several years is producing availability rather than just being described. The commitment we flagged most strongly in August, that the $1.35 billion turnaround plan not be quietly cut, was met: nine-month spend of $1,099M plus a $230M fourth-quarter guide is $1,329M. A lighter 2027 turnaround year is also a mechanical tailwind to next year's throughput.

7. The heavy-crude edge widened again

Our May upgrade test asked whether the crude advantage survives narrower light-heavy differentials. For the third consecutive quarter the test could not be run, because differentials widened again. The commercial organization reported a Hardisty print of roughly $32 under WTI on the morning of the call, against about $14 in August.

"So it continues, that differential continues to widen out. ... I continue to think it may widen out further." — Rick Hessling, Chief Commercial Officer

The structural detail matters more than the print. About 90% of MPC's crude comes from the U.S. and Canada, the waterborne remainder moves on time charters fixed before the conflict, and the fourth quarter is expected to set a record for heavy Canadian runs. Venezuelan and SPR barrels landing on the Gulf Coast and increased Enbridge apportionment were named as further pressure on competing heavy grades.

Assessment: Favourable, again, and still untested in the direction that matters. A $32 discount is a large tailwind to the absolute margin. It also widens the gap between today's economics and the normalized differential we asked about. The Venezuelan and SPR flows behind part of it are policy decisions, and the SPR flow reverses on a schedule: the 10-Q shows MPC must return roughly 26 million barrels to the reserve between April 2027 and July 2029 against about 21 million received.

8. Twelve billion dollars of standalone cash, and a "targeted cash framework"

Consolidated cash and short-term investments reached $13.6 billion, of which $1.3 billion sits at MPLX. MPC standalone cash is $12.3 billion, against a "roughly $1 billion" operating need stated in August, and before $3.45 billion of pre-tax sale proceeds. Working capital was another $2.3 billion source of cash in the quarter.

"Working capital was a $2.3 billion source of cash for the quarter primarily driven by income tax payables and the timing benefit of the SPR crude exchanges." — Maria Khoury, EVP and CFO

The prepared remarks introduced a new phrase without defining it: "We are maintaining balance sheet discipline. With a targeted cash framework that allows flexibility while supporting return to capital." The 10-Q supplies a piece we asked for in August: the SPR return obligation is carried as a derivative liability of $1.9 billion, $590 million of it current. Income tax payables are the other named working-capital source, and taxes paid in the first nine months were $1,804M against $130M a year earlier.

Assessment: The cash that is genuinely committed can now be roughly sized: about $1.9 billion against the SPR return, a tax payable that will be settled, and the tax on a $2.4 billion gain. Even after those, MPC standalone cash will be several multiples of the stated operating need at year-end, and the fourth quarter is guided at near-full utilization. A "targeted cash framework" is precisely what we have asked for since May. Naming it without stating the target extends a series of non-answers that began in May, now in a more frustrating form, because the company has told us the framework exists.

9. Capital return: $3.0 billion at the market's average price, and a new $5 billion

MPC repurchased 8.49 million shares for $3.0 billion in the quarter, at an average of $354.17. The quarter's volume-weighted average price was about $353.21, so the company paid essentially the market average. The board added a $5 billion authorization, taking capacity to $8.1 billion as of September 30, and raised the quarterly dividend 10%. The monthly pattern in the 10-Q is the part that matters for our second bear point.

Dollars deployed roughly doubled from July to August as the share price rose about 15%, and stayed near that level in September at a price another 17% higher.

"We would continue to say that our capital allocation priorities have not changed." — Maryann Mannen, Chairman, President and CEO

Assessment: In August we wrote that whether Q3 intensity tracked price or cash would be the direct test of our second bear point. It tracked cash, and therefore price. The company spent the least in the cheapest month and about twice as much in the two dearer months. Execution against the tape was neutral: no timing skill, no timing damage. The September tranche, bought at $399.42, sat 1.7% under water at the November 4 close of $392.68. That is too small a loss to call the policy wrong. It is, however, the first evidence that the pattern can cost money, and the new $5 billion authorization was announced at a share price 64% above our initiation price.

10. MPLX: a third 12.5% raise, and $3.1 billion a year to the parent

MPLX raised its distribution 12.5% for a third consecutive year, and MPC now expects $3.1 billion a year of distributions from it. Harmon Creek III began operations in August, taking Northeast processing to 8.1 Bcf/d and fractionation to 800 mbpd. Titan sour gas treating is expected to exceed 400 MMcf/d by year-end. MPLX and partners took final investment decision on the Solitude pipeline system, 4.5 Bcf/d of additional Permian gas egress. MPLX will redeploy its $750 million of sale proceeds in the Permian and Marcellus.

"MPC now expects to receive $3.1 billion of annual distributions from MPLX more than covering MPC's annual dividend and capital expenditures a differentiator in our industry." — Maryann Mannen, Chairman, President and CEO

Assessment: The coverage claim checks out on disclosed figures. Nine-month dividends paid were $864M, which with the 10% raise implies roughly $1.27 billion a year, and the 2026 standalone capital outlook is $1.5 billion; $3.1 billion covers both. Next year is the caveat: a higher 2027 refining capital plan would take part of that cushion. The market's verdict on MPLX this quarter ran opposite to the parent's. MPLX units fell 5.0% between our August note and the print while MPC rose 36.7%, and fell 2.4% on the print day.

11. The cost ledger: refining held, the stock-linked lines did not

Refining operating cost came in at $5.60 per barrel, exactly on the guide and one cent above the prior year, on 145 mbpd more throughput. It is the first quarter of our coverage in which this line met its guide. Distribution cost was $5.98 per barrel against roughly $5.97 implied by the dollar guide, essentially on target per barrel and $83M over in dollars because volume was higher. The lines that missed were the ones tied to the share price: Corporate by $68M, and incentive compensation named for the first time as a Midstream cost driver.

"Refining operating costs were $5.60 per barrel for the third quarter of 2026, versus $5.59 per barrel for the third quarter of 2025." — MPC Q3 2026 earnings release

The refining organization framed the posture as "cost discipline" rather than cost cutting, and credited the reliability spending of recent years for holding the line.

Assessment: Our third bear point combined two different risks, and this quarter separated them. Plant-level cost drift is contained: refining operating cost met its guide, the Mid-Continent overrun from Q2 reversed, and the fourth-quarter guide holds at $5.60. The pro-cyclical overhead risk materialized and is now explicitly explained: consolidated SG&A rose $237M year over year, about $150M of it incentive compensation. Neither is large relative to $11.7 billion of quarterly EBITDA. The second, though, is a cost that rises with the share price, and a $5 billion buyback authorization at today's price makes it larger in the same direction.

Guidance & Outlook

MPC guides a short list of R&M cost and volume inputs one quarter at a time, which makes each guide a falsifiable commitment. Below is how the August guide landed, followed by the new fourth-quarter outlook.

Grading the Third-Quarter Guide

The best guide-versus-actual card of our coverage on the lines that matter. Volume beat by 5% and the refining cost line was met to the cent for the first time in three quarters. The distribution dollar overrun is volume, not rate. The one clean miss is Corporate, and it is the stock-linked line. The distribution-cost dollar figure is implied from the disclosed per-barrel rate and actual throughput, because MPC guides that line in dollars and reports it per barrel.

Fourth-Quarter 2026 Outlook

Implied quarter-over-quarter shape: crude runs fall 4.8% as conversion-unit work moves to the Gulf Coast and West Coast, partly replaced by 73 mbpd more purchased blendstocks, so total throughput falls only 2.2%. Turnaround expense drops $64M. Distribution cost per barrel rises to roughly $6.09 because the dollar guide is flat on lower volume. The CFO flagged that the conversion work "will limit our ability to upgrade certain products and create a headwind to capture," the same language used for the third quarter, which came in at 93%.

Full-year turnaround: nine-month refining turnaround spend of $1,099M plus the $230M fourth-quarter guide is $1,329M, against the $1.35B plan management reaffirmed on the call. The August question is answered: the plan was not cut.

Divestiture timing: the sale is expected to close at year-end, so the fourth-quarter guide still includes St. Paul Park and Mandan. Every guide from the first quarter of 2027 should be read against a system about 177 mbpd smaller, with Mid-Continent throughput and its per-barrel metrics re-based.

Guidance style: volume-conservative and now cost-accurate on the plant lines. MPC has met or beaten its crude-run guide three quarters running, by 153 mbpd this time. Refining operating cost, which we told readers in August to model $0.05 to $0.10 above the guide, met it. We drop that adjustment for the plant line and keep it, in larger form, for Corporate.

Analyst Q&A Highlights

Why sell these two refineries, and why at this price

The call opened on the divestiture, with a two-part question: how the assets were chosen and whether management was comfortable with the valuation. The answer addressed strategy at length and valuation in a single clause, and did not engage with the earnings the assets had just generated.

Q: "So maybe you could just talk about the process why you identified these assets and how that fits into the broader strategy? And did you get comfortable with the valuation around this monetization as well?"
— Neil Mehta, Goldman Sachs

A: "We acted on an opportunity to align our portfolio with our strategy at valuations that I think we deem to be very attractive. Glad that we are able to divest to a company that is looking to grow and also operate safely and reliably."
— Maryann Mannen, Chairman, President and CEO

Assessment: "Very attractive" is a claim about price, and every earnings figure for the assets came from the buyer: roughly 3.6x trailing EBITDA and 8.3x FY2025 under the buyer's accounting. Management offered no figure of its own, which is the one number that would let an investor test the claim. The strategic case is fine. The valuation case was asserted, not shown, and it was not reconciled with the decade-long cycle call made earlier on the same call.

How much heavy Canadian crude the system can run, and where WCS goes

The follow-up asked about crude sourcing after the sale and the WCS differential, which had widened beyond even bullish expectations. The commercial answer was the most specific of the call and pointed to further widening.

Q: "So your views on the market and your perspective on how much WCS you can run through the system?"
— Neil Mehta, Goldman Sachs

A: "So we are really leaning into the WCS story. ... In fact, here in the fourth quarter, we expect to set records on how much heavy Canadian we process."
— Rick Hessling, Chief Commercial Officer

Assessment: A record heavy Canadian run rate in the fourth quarter, at a differential about $18 per barrel wider than in August, is a meaningful near-term margin support. It is also additional exposure to the variable we most want to see normalize before upgrading: the heavy differential. The answer did not address how much heavy capacity leaves the system with the Mid-Continent sale, and the question was not repeated.

Whether 2027 holds up, and whether demand is cracking

A question on the refining outlook asked whether the gasoline pullback was seasonal and whether there were early signs of demand destruction. Management said no to the second and yes to the first, and anchored its 2027 view in inventories.

Q: "So as you enter 2027, how are you looking in terms of the overall refining macro? And have you seen any early signs of any kind of demand destruction here?"
— Manav Gupta, UBS

A: "But frankly, we have not seen really demand destruction across our system, particularly when you look at jet and diesel. ... So we remain quite constructive over 27."
— Maryann Mannen, Chairman, President and CEO

Assessment: The answer was constructive and conditional at once. The CEO named the "timing of, you know, resolution on the Iran conflict" as the variable, and the commercial organization added that "everything you see is headline driven here lately." The next session demonstrated that point at MPC's expense. What management did not do is size the downside, which remains the gap in an otherwise improving disclosure set.

What happens to a cash balance that could pass $20 billion

The sharpest capital question of the call put a number on the problem: standalone cash already above $12 billion, a strong fourth quarter, and sale proceeds arriving at year-end. The answer restated priorities and pointed to Gulf Coast reinvestment.

Q: "Your cash on the balance sheet could balloon to over $20 billion by the end of this year, and that is a lot of cash to hold on to the balance sheet. So I am just trying to understand the possible uses of this amount of cash on the balance sheet."
— Manav Gupta, UBS

A: "We would continue to say that our capital allocation priorities have not changed."
— Maryann Mannen, Chairman, President and CEO

Assessment: The same answer as in May and August, to a question that has grown each time. The $20 billion figure is the questioner's estimate, not management's, and it does not net out the SPR liability or taxes on the gain. Management did not correct it, did not offer its own year-end figure, and did not define the "targeted cash framework" from its own prepared remarks. A company with a framework could have answered with a number.

Whether the proceeds defend the stock or fund deals

A follow-up on the proceeds noted that growth capital has run around $1 billion a year, so the $3.45 billion cannot all go to organic projects, and asked whether it would support the stock in a downturn or fund acquisitions. Management answered on capital spending first and returned to buybacks only when it realized the question had not been answered.

Q: "So should we expect some of that cash to be used to defend the stock if the market turns? Does it open the aperture for inorganic activity?"
— Jason Gabelman, TD Cowen

A: "We really are not intending to reduce or shrink our refining footprint. We really want to be sure that our refining footprint meets the demand and we believe the Gulf Coast is a place where capital is warranted there."
— Maryann Mannen, Chairman, President and CEO

Assessment: Two signals, one by omission. The 2027 capital budget will be larger than the recent run rate, and the inorganic half of the question went unaddressed. The idea of holding cash to buy stock "if the market turns", which is the closest thing to a valuation-sensitive buyback anyone has proposed on these calls, was not taken up. The pattern in the 10-Q shows why: this company buys when it has cash, not when the stock is down.

What is in the fourth-quarter capture outlook

A question on fourth-quarter capture asked for the puts and takes: wider WCS, freight, backwardation, secondary products and the conversion-unit turnarounds. The answer gave a one-month read and a definitional caveat that matters more than the read.

Q: "Can you just walk through some of the puts and takes on capture relative to 3Q?"
— Joe Laetsch, Morgan Stanley

A: "So we will continue to watch that for the next 60 or so days. But early on, we are seeing very similar numbers that we saw on the secondary side in 3Q."
— Rick Hessling, Chief Commercial Officer

Assessment: One month into the fourth quarter, the largest capture headwind in the company's history had not eased, and conversion-unit turnarounds are guided to add to it. Management's counterweight is that the fourth quarter is "typically our strongest quarter on the things that we can absolutely control." Read together, we would model fourth-quarter capture below the 101% year-to-date figure, not above it.

Whether the sale proceeds are taxable

A short exchange established that the refining proceeds are fully taxable, which matters for how much of the $3.45 billion actually reaches the balance sheet.

Q: "I would imagine the sale price is well in excess of your cost basis"
— John Royall, Piper Sandler

A: "So when you think about the taxable implications from a refining perspective, you will expect that those proceeds will be taxable."
— Maria Khoury, EVP and CFO

Assessment: A direct answer to a direct question. With a carrying value of $1.8 billion against a $4.2 billion price, the consolidated pre-tax gain is about $2.4 billion, part of it in MPLX's $750 million piece. Management did not size the tax. If the whole gain were taxed at the 22.5% effective rate MPC recorded this quarter, the leakage would be roughly half a billion dollars, an upper bound for MPC's share. Net proceeds are therefore closer to $3 billion than to the $3.45 billion headline, which is still a large addition to an already unanswered cash question.

What They're NOT Saying

1. What the divested refineries earn, from MPC's side: The release gives a price and, for MPLX's share, a forward multiple. It gives no earnings for the refining assets. The only figures in the public domain are the buyer's: about $1.2 billion of trailing EBITDA and $506 million for FY2025, under an accounting policy that capitalizes turnarounds. MPC described the valuation as very attractive without publishing the number that would let anyone test it.
2. What "targeted cash framework" means: Introduced in the prepared remarks, not defined, not quantified, and not referenced in any of four capital questions. It is the closest management has come to the framework we have asked for since May, and it arrived without its content.
3. Any reversal scenario, third quarter running: Management lengthened its horizon to "the end of the decade" and named the Iran conflict's resolution as the swing variable. It still declines to say what the business earns when the excess outage returns to zero.
4. A bridge between three outage measures: About 6 million barrels per day conflict-attributable in May; over 9 million total, about 4 million above norm, in August; 10 million total, about 3 million above norm, now. The implied norm moved by about 2 million barrels per day in one quarter without comment.
5. The size of the secondary-product drag: Called the largest in the company's history, and still not quantified in capture points or dollars for the third consecutive quarter. The 10-Q's roughly $700 million aggregate variance bundles it with crude mix, market structure, RINs and yields.
6. The tax on the sale, and net proceeds: Confirmed as taxable, not sized. On a roughly $2.4 billion gain the difference between gross and net is material to any capital-return arithmetic.
7. Why Corporate depreciation doubles: The fourth-quarter Corporate guide includes $60M of depreciation against $30M in the third-quarter guide and $26M actually recorded. The change was not explained.
8. Realized returns on completed projects, fourth quarter running: Garyville jet flexibility, El Paso yield improvement and Robinson product flexibility are all in service. The company is about to raise refining capital and still has not published an outcome against its stated hurdle for any of them.
9. The organic Midstream growth rate: Rates, equity affiliates and acquisitions are listed again without a split, now alongside a divestiture drag and incentive compensation. The 10.1% headline is good; its composition remains unrecoverable from the disclosure.

Market Reaction

• Pre-print setup: MPC closed at $430.39 on November 2, entering the print up 164.6% year to date against +10.2% for the S&P 500, up 119.8% over the trailing twelve months, and up 1.9% over the trailing 30 days. The 52-week closing range was $162.63 to $484.92.
• The peak came three weeks early: The record close of $484.92 was set on October 14. On October 27 the stock fell 7.4% in a sector-wide session driven by U.S.-Iran diplomatic signals and a diesel cargo that is the first U.S. import of a Russian petroleum product since 2022. The pre-print close sat 11.2% below the record.
• Reaction session (November 3): MPC opened at $430.00, a 0.1% gap down, traded a $414.06 to $435.00 range, and closed at $427.34, down 0.7% or $3.05. Volume was 2.8M shares against a 2.3M 30-day average, a 1.2x day. Reported premarket strength of roughly 2% on the beat did not hold.
• Against the index and peers: The S&P 500 rose 0.4%. Valero rose 1.6%, HF Sinclair 1.5% and PBF 3.6%; Phillips 66 fell 1.1% and Exxon Mobil 1.2%. MPLX units fell 2.4%. The buyer, CVR Energy, fell 13.1%, on dilution from the equity offering that funds the purchase.
• Following session (November 4): MPC fell 8.1% to $392.68, after opening at $417.90 and trading as low as $379.10, on 4.0M shares, 1.7x average. The S&P 500 rose 1.6%. Valero fell 5.8%, HF Sinclair 7.8%, PBF 6.7%, Phillips 66 4.5% and Exxon Mobil 4.5%. The trigger was the midterm result: with Democrats taking control of Congress, the market priced an earlier end to the Iran conflict, and crude and gasoline futures sold off hard.
• Positioning after two sessions: The November 4 close sits 19.0% below the October 14 record and 10.9% above the $354.17 average price MPC paid for its third-quarter repurchases.

The print itself barely moved the stock, and that is the first thing to explain. An EPS beat of 4.6% against the Zacks mark is the smallest of our coverage in percentage terms, against 22.1% in Q2 on the same vendor and more than double the Street in Q1. The Street had caught up with the crack. What the call added at the margin was mildly negative and forward-looking: the largest secondary-product headwind in the company's history persisting "through October," fourth-quarter crude runs guided 4.8% lower, and conversion-unit turnarounds flagged again as a capture headwind. The divestiture, the dividend increase and the new authorization did not earn a premium. MPC lagged three of the four refining peers we track on a day crude fell and products firmed.

The following session matters more for the thesis than the print. A political event with no company news took 8.1% off the stock, while the S&P 500 rose 1.6%. That is a 9.7-point underperformance in one day, larger than the reaction to a $26.93 quarter in either direction. Together with October 27, it means two sector-wide sessions within seven trading days removed roughly 15% from MPC on news about the probability of the conflict ending, not about the company. We read this as the market's own estimate of how much of MPC's value is rented. It is the same question management declines to answer, and the market answered it twice in two weeks.

The honest scorecard on our rating belongs here too. From our August note to the reaction close, MPC returned 36.7% against -2.2% for the S&P 500, a 38.9-point underperformance for a Hold in one quarter. Even after November 4 the gap is 26.2 points. Since our May initiation the stock is up 50.7% against 5.9% for the index. We correctly identified the type of risk, and the last two weeks are a demonstration of it. We were wrong about how long the rent would run before the market began to discount its end.

Street Perspective

Debate: Is MPC cheap on forward earnings, or are forward earnings war earnings?

Bull view: The bull case being made on the Street is that MPC trades around five times forward earnings after the November 4 drop, less than half its five-year average, against a management team that has now called an enhanced mid-cycle through 2030. Published consensus for 2027 sits roughly in the low-to-mid $60s of EPS, broadly flat on 2026, and the stock is pricing a sharp reversal that management, inventories and the absence of demand destruction all argue against.

Bear view: The bear camp contends that the forward multiple is computed on earnings the conflict created, and that two sessions within seven trading days have shown what happens when the market assigns even a moderate probability to its end. Some desks argued before the print that the share price already discounted the mid-cycle margins expected in 2030; on that view, there is nothing left to pay for if the war premium goes.

Our take: Neither side's multiple is the right one, because the earnings base is the variable. Using the company's own crack sensitivity, $392.68 is about 12x the EPS MPC would earn at a $25 blended 3-2-1 crack, which is roughly two-thirds above the first nine months of 2025 and a quarter below the first nine months of 2026. In other words, the current price already assumes something like the "enhanced mid cycle" management described, and only that. It does not assume the war. That is fair value for a Hold, not a bargain, and the November 4 session suggests the market is pricing roughly the same thing.

Debate: Was the CVR sale good capital allocation, or the war sold at a pre-war price?

Bull view: A growing consensus view is that the sale affirms a strategy of concentrating capital in complex, integrated core assets. MPC gets $4.2 billion, roughly 8.3x the plants' FY2025 EBITDA on the buyer's figures, books a consolidated gain of about $2.4 billion over carrying value, and redeploys into the Gulf Coast, where it runs its lowest-cost plants.

Bear view: The bear camp contends that the price is roughly 3.6x trailing EBITDA, that the buyer itself says it negotiated before the conflict and underwrote mid-cycle earnings, and that MPC therefore handed over the war upside for nothing. Meanwhile the proceeds, net of tax, add another $3 billion or so to a cash balance that management cannot explain.

Our take: The strategic logic is sound and the normal-year multiple is fair, so we do not score the sale as a mistake. What it does is reveal how management values a refining barrel when it is the one being paid: at mid-cycle, not at the enhanced cycle it describes on the call. The real test is redeployment. A swap from 8.3x normal-year earnings into Gulf Coast projects works only if those projects clear the stated 25% hurdle, and MPC has not yet published a realized return on any completed project.

Debate: Is the buyback disciplined, or pro-cyclical by design?

Bull view: Some sell-side desks argue that $3.0 billion executed at essentially the quarter's average price, 2.8% of the market capitalization in one quarter, a 10% dividend increase and $8.1 billion of fresh capacity are exactly what shareholders should want from a company generating this much cash. Even after its two-day drop, the stock trades about 11% above the average purchase price, and nobody on the Street is asking the company to buy less.

Bear view: The bear camp contends that the monthly data settles the question the company will not answer. MPC spent $601M in July at about $292 and roughly twice that in each of August and September at $353 and $399, so intensity rose with the price. The September tranche is already under water at the November 4 close, and the new $5 billion authorization arrived within three weeks of the record close.

Our take: The bear is right about the pattern and the bull is right about the outcome so far. A company that buys when cash arrives will, at a refiner, buy most heavily near the top of the cycle, and this quarter showed it doing exactly that. The fix is the one management has now half-announced: a "targeted cash framework" with a number in it. Until the number exists, we treat the second bear point as materializing rather than emerging.

Model Implications

Aardvark Labs does not yet maintain a published financial model for MPC. As in August, the table carries forward the driver set established at initiation and marks where this quarter moved each one, so the fourth quarter can be graded against a stated starting point.

The Valuation Bracket

At the November 4 close of $392.68, MPC's market capitalization is roughly $108.0B on 275M shares, and consolidated enterprise value roughly $127.4B against $33,022M of debt and $13,609M of cash and short-term investments. That is 4.8x trailing-twelve-month adjusted EBITDA of $26,441M and 10.7x FY2025's $11,956M. On earnings, the shares trade at 6.5x nine-month adjusted EPS annualized, 7.8x trailing-twelve-month adjusted EPS of $50.38, and 36.7x FY2025's $10.70. The range between 6.5x and 36.7x is wider than August's 8.1x-to-29.2x bracket: the share price rose against a fixed FY2025 base, while annualized earnings rose faster than the price.

A more useful frame converts the debate into the one variable that decides it. The 10-Q gives a sensitivity of about $1,125M of annual R&M EBITDA per $1 per barrel of blended 3-2-1 crack. After tax at 22.5% and on 275M shares, that is roughly $3.17 of annual EPS per $1 of crack. Anchoring on the third quarter annualized and holding capture, volume and share count constant gives the bracket below. It is illustrative, not a forecast: it ignores Midstream growth, buybacks and the divestiture.

The bracket calibrates reasonably against reported results: at the nine-month average crack it gives about $57 of annual EPS, against $60.85 for the nine months actually annualized. Its message is that the share price is roughly a 12x multiple on a $25 crack. That is about two-thirds above where the crack ran in 2025 and well below where it runs now. Put differently, the stock already discounts a lasting enhanced mid-cycle and a small part of the war premium. It does not discount the third quarter.

Valuation impact: The business case is about where it was in August: better operations and cost control, a weaker capture and yield read, and a stronger Midstream. The share price is 26% higher than at our August note even after November 4, and the bracket says the price already assumes most of what management's new macro call asserts. We do not see an asymmetric entry point in either direction.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation in May, unchanged in wording, scored against this quarter's print, call and 10-Q.

Overall: thesis unchanged in direction, weaker in composition. The operating pillars that management controls, regional leadership and MPLX, are stronger than in August. The pillar we upgraded in August on yield evidence is weaker than we said, and we correct ourselves on it. The capital-allocation risk moved from a concern to an observed pattern. The cycle risk is unchanged in our tags and visibly larger in the share price, which has now had two sector-wide repricing days in seven sessions.

Action: Hold. First, the cost of the call. From our August note to the reaction close, MPC returned 36.7% against -2.2% for the S&P 500, and since the May initiation the stock has beaten the index by 44.8 points even after November 4. That is a large miss and we do not argue it away. Our upgrade conditions have changed once, and the history belongs on the record. In May we set three and required two: a Mid-Continent recovery above the FY2025 R&M average of $5.63 per barrel with continued regional leadership, a stated repurchase framework, and evidence that the crude edge persists at narrower light-heavy differentials. In August, after the Mid-Continent condition was met, we restated them as three alternatives, any one sufficient: a stated capital-return framework, evidence the crude edge holds through a normalized differential, or a price closer to the mid-cycle management concedes exists. We keep the August set unchanged and grade both. The Mid-Continent condition was met again, at $36.00 per barrel, the best region in the system. A stated capital-return framework was not met: the framework was named, not stated. Evidence that the crude edge holds at a normalized differential was not testable, because WCS widened to roughly $32 under WTI. A price closer to the mid-cycle was not met: even after November 4 the stock is 26% above our August note price, and our bracket says it already discounts a lasting enhanced mid-cycle. That is one of the two the May rule required and none of the three August conditions. To pre-state how we would net a mixed outcome: a stated cash target with a return cadence, published with the February capital plan, would count as the capital-return condition met even if paired with a larger capital budget. A blended crack sustained near $25 with the stock materially below today's level would satisfy the price condition. On the downside, the May Underperform trigger, an R&M margin compressing toward the FY2025 average of $16.87 per barrel while repurchase continued near the current price, did not fire: the margin was $45.70. We add a new Underperform trigger alongside it, not in its place: we would move to Underperform if repurchase continued at the third-quarter pace into a blended crack that had already fallen toward $25, because that is the second bear point turning from a pattern into a cost. We would not add to a refiner at 5.9x to 6.7x the published 2027 consensus in the week the market began pricing the end of the conflict that produced them. We would not sell one that generated $8.5 billion of operating cash flow before working capital in a single quarter, about 8% of its market value, while the dislocation lasts.