Q4 Stock Winners And Losers: Intel, Valero, Risk Reset

I kept staring at an old notebook from January and laughing at the numbers I had circled in pencil. A chip name in the low forties. A refiner that still looked like a gap-and-go story tied to one headline. A cement producer that was supposed to ride falling mortgage rates. A solar manufacturer that sat on a list and still felt wrong. Nine months later the first two have more than doubled, the last two have quietly slipped off the tape, and the fourth quarter has opened with the ten-year yield parked near a level last seen in 2007. If you still treat January’s lines as living risk, you are managing a ghost.

That is the awkward gift of a momentum screen. Names qualify on price and relative strength, and they fall off the same way. Nobody votes them in. What a human still has to decide is which chart deserves a write-up, what the setup actually is, and where the idea is simply wrong. Those levels get drawn before the ending is known. Three quarters into a wild year is a decent moment to grade the ink.

Why A Mid-Year Grade Matters More Than A Fresh Pick

Lists built by the market do not care about your narrative. They care whether a stock is still acting like a leader. As of the first Monday in October, a broad best-stocks screen still held about 145 names. That count is large enough to hide a lot of mediocrity and small enough that the real leaders stand out if you bother to rank them. The interesting work is not crowning a new favorite every Monday. It is going back to four older write-ups and asking a ruder question. Did the level you set still tell you anything?

I’ve found that most post-mortems cheat. Winners get a victory lap. Losers get a footnote. A fair grade does both in the same sitting, with the same rules. Two names on this tape did nearly everything a trend follower could have asked. Two others broke the rules early, or never earned a fresh long in the first place. The split is the lesson.

A stop that sits 40 percent below the last print is not risk management. It is a souvenir.

The backdrop makes the split sharper. The ten-year Treasury closed September 28 at 5.24 percent, right on the prior 2007 closing high. The central bank hiked in September, and markets were still pricing another move in October. Inflation had not cooled in any way that comforted bond desks. A conflict involving Iran was in its seventh month. Spending on artificial intelligence kept lurching higher every quarter, which helped some industrial and chip stories and did nothing for rate-sensitive building names. Football fans already know the cliché. The fourth quarter is where a lead either becomes a win or turns into a story you do not want to retell.

What The Screen Actually Rewards

Price first. Relative strength second. Story third, and only if the first two are already behaving. That order annoys people who want a thesis before a chart. It also keeps you out of a lot of elegant ideas that never trend. A name can have a backlog, a buyback, and a tidy slide deck and still be the wrong long if the 50-day is rolling over and closes keep failing.

Perhaps the most interesting aspect of this year’s tape is how far the winners sat from the herd’s favorite theme. Yes, one champion is a processor company tied to data-center demand. The other champion is a refiner. Not a model lab. Not a power-purchase agreement dressed up as software. Just crude in, fuel out, and a margin that exploded when supply routes jammed. Once you study price instead of headlines, those pockets stop looking like accidents.

Intel: The Rebuild The Tape Finally Paid For

Intel designs and manufactures processors for personal computers and data centers, and it is trying to stand up a foundry business that makes chips for other firms. The name was written up twice, once in late January and again in late April. From the first note the stock is up about 163 percent. From the second, about 39 percent. Year to date the gain sits near 214 percent. That is not a gentle rerating. That is a stock that left the station and then left the station again.

January did not look like a celebration. Fourth-quarter revenue had printed at $13.7 billion, down 4 percent, and first-quarter guidance of $11.7 billion to $12.7 billion missed what the street wanted. The only number that mattered on the chart was roughly $40. Hold the low forties. Respect a rising 50-day. The uptrend was intact even if the press release was ugly. By April the problem had flipped. A 24 percent one-day jump on first-quarter results shoved the relative strength index to 82. Stretched, yes. Broken, no.

Guidance for the second quarter sat at $13.8 billion to $14.8 billion. The company delivered $16.1 billion, up 25 percent from a year earlier. That was a seventh straight quarter above expectations. Non-GAAP earnings of $0.42 more than doubled the guided figure. The data-center and AI segment grew 59 percent to $6.3 billion and ran at a 40 percent operating margin. Management said server CPU demand was outrunning industry supply and expected the pinch to last into the following year. You can argue about how long a shortage narrative survives. You cannot argue that the print was soft.

The Summer Giveback And The Second Leg

The stock ran as high as $142 earlier in the year, then handed back roughly a third of the June-era gains by August. The summer low landed in the low $80s near the end of July. A higher low followed in September. From there price pushed back above the 50-day and cleared the $100 to $105 ceiling that had capped every August rally. September 21 added another leg, a 12 percent jump tied to hopes that a popular consumer AI agent would lift CPU demand. Hope is not a model. A gap that holds is.

Into the first week of October the stock sat near $119. The 50-day had turned back up around $101. The 200-day rested near $81. Friday’s session pushed to $126 intraday and reversed to close at $119, just under a recent high near $127. That $127 area is the near-term ceiling. The relative strength index had cooled to about 61 after a late-September push above 70. The July washout had dragged that same oscillator into the low 30s. A full flush, then a repair, then a second leg. Anyone who wanted out had a window. At 61, momentum is no longer stretched enough to forbid another attempt at $127.

Nothing on that chart looks broken to me. Extended versus January, obviously. Broken, no. The next hard date is the third-quarter report on October 22, against guidance of $15.8 billion to $16.8 billion in revenue and $0.38 in earnings per share. Beat that and the ceiling gets tested. Miss it and the gap from September 21 becomes the first argument.

Either the gap holds or it does not. You will know which soon enough, and you should judge it on the close.

A chart rule worth keeping

Resetting Intel Risk So The Stop Still Means Something

Original stops from the forties are now so far below the quote that they tell you nothing useful. They are history. Fresh lines have to sit where the current trend would actually fail.

• Traders can use $110 as the line in the sand. That is the open of the September 21 gap candle, the first price paid after the air pocket. A close back below it means the gap is filling and the September move has lost urgency.
• Investors can anchor to $100. The rising 50-day sits on top of the $100 to $105 band that capped August and then launched September. A weekly close under $100 says the second leg failed.
• Intraday noise does not count. Whipsaws love round numbers. Closing prices are the only vote that matters.

In my experience, people hate resetting a stop upward because it feels like moving the goalposts. It is the opposite. You are refusing to pretend that a 2025 or early-2026 level still describes 2026 risk. If $110 breaks on a close, the trade thesis from September is wounded. If $100 breaks on a weekly close, the investment thesis from the second leg is wounded. Those are different clocks. Mixing them is how small losses become furniture.

Valero: The Refiner That Became A Horse

Valero is one of the largest independent petroleum refiners in the world. It turns crude into gasoline, diesel, and jet fuel, and it runs renewable diesel and ethanol businesses alongside the core plants. The January note landed right around fresh Venezuela headlines. Energy names gapped. Among the three major independent refiners, this one looked like the cleanest horse. Investors were told they could stay long above a $155 to $160 support zone. Traders were pointed at a rising 50-day near $172 rather than the bottom of a gap near $177, because gaps love to whipsaw the impatient. That 50-day now lives around $350. Read that sentence again if it did not land.

From the January write-up the stock is up about 113 percent. Year to date the gain is closer to 142 percent. Venezuela turned out to be the opening act of a much larger oil-price shock. Since late February, closure of the Strait of Hormuz during the Iran conflict has disrupted something on the order of 20 percent of the world’s oil supply. Brent spiked toward $118 in late March, fell back near $70 by July 1, then climbed toward $109 in early September as attacks on shipping and energy infrastructure resumed. Product markets do not need a permanent spike to print spectacular refining profits. They need tight capacity and inventories that are already thin.

On the second-quarter call, management described global refining capacity as tight and product inventories well below historical averages. The numbers matched the tape. Net income was $3.7 billion, or $12.62 a share, against $714 million, or $2.28, a year earlier. Revenue rose to $44.5 billion from $29.9 billion. Refining operating income more than tripled to $4.5 billion. The company returned $2.6 billion to shareholders in the quarter, a 59 percent payout ratio, and the board authorized a fresh $5 billion buyback in July. That is not a story stock hoping for a multiple. That is cash showing up.

Why The Non-Tech Winner Matters

I am doubly interested in this call because it has nothing to do with the year’s dominant data-center theme. Phillips 66, Marathon Petroleum, and Valero have sat near the top of large-cap energy performance for most of the year. The third of that trio was still more than 25 points ahead of the next energy name in the big index, APA, when this review was written. When a price screen works, it can look like a magic trick. How did you find that? You watched what was already leading, then you refused to abandon it because it was not fashionable.

It will not always work this cleanly. That is why the risk lines exist. Valero spent the spring chopping between roughly $230 and $265, broke out in July, and has not come back to test its 50-day since. Late September brought the first real scare. After tagging a high near $418, the stock was hit on the heaviest volume in months and dropped toward $365 within a few sessions. Buyers absorbed it. Price recovered to about $406, with the 50-day rising near $355 and the 200-day near $260. The $418 high is the level that still needs a clean clear.

The oscillator told the same story in fast motion. It peaked near 80 in mid-September, fell to 50 on the selloff, and climbed back toward 66 as the stock repaired. A full momentum reset in a handful of sessions, without damage to the trend. Profit-takers were flushed. The chart did not need weeks of therapy. Healthy, if you can stand a word that sounds soft next to a refiner.

1. Traders can use $365 as the stop. That is the late-September low, where the selloff ran out of sellers. A close beneath it turns a shakeout into something meaner.
2. Investors can anchor near $350. The rising 50-day around $355 sits on the mid-August consolidation. Two separate reads, same neighborhood.
3. A weekly close under that zone puts the July trend in question. You get the answer before a small loss becomes a round trip.

Could crude roll over if shipping lanes reopen and the risk premium vanishes? Of course. Refining margins are not a pension. They are a spread, and spreads mean-revert when the scare fades. That is an argument for the stop, not an argument for pretending the trend is imaginary while it is still making higher lows. The tape has been kinder to this group than to almost any other pocket outside the obvious technology leaders. Respect it until it stops deserving respect.

The table is a scoreboard, not a promise. Levels move. Trends end. The point is that the winners needed new lines because the old ones had become decorative, and the losers had already answered the original question.

CRH: When The Rate Handoff Never Arrived

CRH produces cement, asphalt, and other construction materials, with most of the business in North America. The January thesis was a falling-rate handoff into materials. Mortgage rates had just dipped under 6 percent for the first time in five years. Traders were given a rising 50-day near $120. Investors were given $105, the bottom of an August gap, with a plain sentence attached. Below that level you are wrong. Move on.

The stock broke with the homebuilders as yields vaulted. The ten-year started the year around 4.15 percent and sat above 5.2 percent by early October. A widely followed 30-year mortgage average hit 7.03 percent as of September 24. Existing-home sales fell to a one-year low of 3.98 million in August. CRH felt it in the Americas building-solutions segment, where second-quarter revenue fell 2 percent and adjusted EBITDA fell 8 percent on divestitures and subdued residential demand. Outside commentary pointed at weak U.S. cement sentiment, worry about an infrastructure funding cliff, and skepticism toward recent deal-making.

In June the company agreed to buy Arcosa for $150 a share in cash, a deal still pending at the time of this review, at roughly $8.5 billion of enterprise value. Buybacks were paused after the announcement. The operating business did not collapse. Second quarter was a record on several lines: revenue of $10.8 billion, up 6 percent; earnings of $2.21 a share, up 14 percent; full-year earnings guidance reaffirmed at $5.60 to $6.05. At about $83.60 the stock traded near 14 times the midpoint of that guidance. Earnings held up. The multiple did not. The name left the leadership list on March 5 and was down about 33 percent from the January note, about 31 percent year to date.

This is the part people skip because it is boring and it saves accounts. The $120 trader line and the $105 investor line both broke within roughly two months. Taking a stack of small losses quickly will always beat locking capital into a donkey for months. I do not enjoy that metaphor. I also do not enjoy watching a “quality compounder” narrative talk someone out of a level they wrote down in January. The chart showed the $105 shelf for a reason. Once it went, the rate thesis was no longer a thesis. It was a wish.

CRH scorecard, plain version:
  January idea: falling rates, materials handoff
  Trader line: 50-day near $120
  Investor line: $105 gap shelf
  What happened: both levels failed by early spring
  Business: record quarter, multiple compressed
  Lesson: earnings can be fine and the stock can still be wrong

Could the Arcosa deal eventually look clever if infrastructure spending stabilizes and residential demand stops shrinking? Maybe. That is a different trade from the one that was written in January, and it does not get to borrow the old stop. A new setup would need a new base, a reclaim of intermediate moving averages, and a yield backdrop that stops punishing anything with a long construction cycle. Until then, a cheap multiple on a broken chart is a value trap wearing a hard hat.

First Solar: On The List, Not Ready For A Long

First Solar makes thin-film solar modules in the United States and is one of the few module makers that does not lean on Chinese supply chains. The late-January note carried a headline energy that the stock could soon break out toward highs not seen in two decades. It also carried a warning. The name was on the leadership list and still not ready for a fresh long. Two conditions were named. Reclaim the 50-day with stronger closes and a turning oscillator. Or retest a rising 200-day around $200 and prove it. Neither was a suggestion to chase.

The fundamental postcard was real enough. A backlog then cited at $16.4 billion, 54.5 gigawatts running through 2030, plus an expected 50 percent earnings-growth year. Backlogs are not cash in the door, and they shrink when projects slip. By the autumn review the backlog had eased to 45.1 gigawatts, worth about $13.6 billion. Still large. No longer the same postcard.

Price did the more interesting thing. The stock eventually tagged an all-time high of $320.95 on June 3, about 32 percent above the January write-up price. Then it rolled over and was removed from the list a final time in late June. Into early October it sat roughly 45 percent below that June high, down about 27 percent from the January note and about 32 percent year to date. On September 24 it fell 10.3 percent in a session as higher borrowing costs weighed on projects that need heavy upfront capital, with the ten-year near 5.2 percent. Rates, politics, and a higher cost of materials all had a turn at the microphone. The chart did not need a culprit. It needed a buyer, and the buyer did not show up at the levels that would have made the January caution unnecessary.

If inclusion on a leadership list were enough, you could wrap the screen in a fund and charge a quiet fee. It is not enough.

After the January post the stock kept falling into spring, left the list, bottomed somewhere between March and May, then staged a short, sharp summer rally into that June high. So what. The job of a setup note is not to capture every point of upside in every name that eventually twitches higher. It is to say when a qualified stock is still a bad entry. Sometimes a formation looks like a breakout waiting to happen and is actually a chart that has not finished disappointing people. Writing “not today” is a position. It just does not show up in a performance trophy until the alternative path reveals itself.

I’ve found that readers remember the winner they missed and forget the loser they were talked out of. The second memory is worth more. First Solar is a clean teaching case because the fundamental story was easy to like and the technical posture was easy to distrust. Both things were true at once. Holding both in your head is the whole craft.

Rates Did The Sorting, Not The Headlines

Look at the four names as a rate experiment and the pattern gets less mysterious. Intel’s demand story lived in server CPUs and a supply squeeze. Higher yields did not cancel a data-center buildout that was already funded. Valero’s profit story lived in a physical disruption and tight product inventories. A 5 percent ten-year does not refill a diesel tank. CRH and First Solar both needed cheaper capital somewhere in the chain, either at the homebuyer or at the project sponsor. They did not get it.

Recent market commentary has kept circling the same trio: inflation that refused to cool, a policy rate path that still pointed up into October, and a geopolitical premium in energy that flickered rather than vanished. You do not need a forecast to trade that. You need to notice which business models flinch when the ten-year kisses 5.24 percent and which ones do not. Mortgage rates back above 7 percent are not a footnote for cement and asphalt tied to residential work. They are the thesis.

• Chip supply tightness can coexist with expensive money if the buyer is a hyperscale operator already committed to capacity.
• Refining spreads can widen when a shipping lane closes even if the bond market is miserable.
• Building materials feel the mortgage rate before they feel the infrastructure bill.
• Solar projects feel the cost of capital before they feel the backlog slide.

None of that is destiny for the fourth quarter. A ceasefire, a yield retreat, or a soft inflation print could reorder the deck. The point of writing levels before the ending is that you do not have to guess the macro speech. You watch the close against the line you already chose.

How To Read A Monday List Without Worshipping It

A weekly leadership list is a menu, not a meal. Sector ranks and industry ranks tell you where relative strength is clustering. The top five by relative strength tell you who is currently impossible to ignore. A sector spotlight is useful only if you then open the chart and ask whether the entry is mature, early, or already exhausted. Blind allocation to every name that qualifies is how a good screen becomes a bad product.

In my experience, three filters keep a Monday scan honest. First, is the stock above a rising intermediate average, or is it merely famous? Second, did the last pullback hold a prior breakout, or did it slice through it on heavy volume? Third, is the fundamental print confirming the price, or are you long a multiple that has already done the work? Intel cleared the second and third filters after a scary summer. Valero cleared all three and then survived a September volume slap. CRH failed the first filter in March and never got it back. First Solar failed the entry filter in January even while it still qualified.

Simple screen habit: qualify on strength, enter on structure, exit on the close you wrote down.

That sentence is short on purpose. Most bad trades are long explanations attached to a level the trader already violated. If you need a paragraph to justify staying under your line, you are no longer in the trade you planned. You are in a new trade that has not been planned.

Earnings Season Is The Next Stress Test

Intel reports on October 22 against a revenue band of $15.8 billion to $16.8 billion and earnings guidance of $0.38. The bull case wants another beat in the data-center segment and language that supply stays tight into next year. The bear case wants a guide-down, a margin slip, or a foundry timeline that suddenly needs more patience than the multiple can stand. The chart case is simpler. Hold $110 on a closing basis through the print and the September gap remains a feature. Lose it and the second leg is being audited in public.

Valero does not have the same immediate date circled in this note, but the macro calendar is its earnings call. Any credible reopening of constrained shipping, or a sharp retreat in Brent from the low $100s back toward the summer trough near $70, will test whether $365 was a shakeout or a preview. Refiners can stay bid on tight inventories even if crude cools, right up until product stocks rebuild. Watch inventories, not just the headline barrel.

For the losers, earnings are not a rescue plan. CRH can print another tidy quarter and still fail to reclaim $105 if mortgage demand stays stuck near a 3.98 million sales pace. First Solar can talk about a domestic supply chain and still fail if project finance remains expensive and the backlog keeps leaking gigawatts. A good quarter inside a bad trend is a trade for someone else, after a base exists.

Position Size Is The Part Nobody Frames

A doubled stock creates a sizing problem that the original note never had. If Intel or Valero has become a large slice of a portfolio because price did the compounding, the new stop is not only a chart level. It is a dollar risk. A 8 percent dip from $119 to $110 is a different amount of money than an 8 percent dip was in February. The same is true from $406 down to $365. Winners that are not trimmed or re-anchored quietly turn into concentrated bets. Concentration is fine if you chose it. It is sloppy if it happened while you were on vacation from the position.

One approach I like is mechanical and a little dull. When a name doubles from the entry note, either harvest a portion back to the original dollar risk or rewrite the stop so the remaining shares risk a predefined slice of the account. Do not do both by accident and call it a strategy. The September shakeout in the refiner was a gift for anyone who had been waiting to tighten. Buyers showed up at $365. That is information. It is also a place to decide whether you still want the full size if that low fails next time.

Losers teach the other half. CRH under $105 was not a place to “average down because the multiple is 14.” Averaging down is a strategy when the original thesis is intact and the level was noise. It is a habit when the level was the thesis. First Solar under the 200-day, after a failed summer spike, was not a coupon. It was a stock that had already shown you the rally and the failure. Missing the June high is not a reason to buy the autumn low out of spite.

A Fourth-Quarter Checklist That Fits On One Card

You do not need a new personality to trade the last quarter of a year like this. You need fewer open questions. Here is the card I would actually keep next to the screen.

1. Rewrite stops on anything up more than 80 percent from the original note. Decorative levels do not count.
2. Separate trader closes from investor weekly closes. They answer different questions.
3. Treat list inclusion as a qualifier, not an order ticket.
4. If a name broke both published levels, it is off the book until a new base exists.
5. Into earnings, decide the reaction level before the release, not during the first three minutes.
6. Check whether the business flinches when the ten-year is above 5 percent. If it does, size it like a rate trade.
7. Keep a pocket of leadership that is not the crowd’s favorite theme. This year that pocket was fuel.

Short list. Harder to follow than it looks, mostly because a green quote makes people generous with rules and a red quote makes them inventive. The inventors are the ones who will explain, in December, why $100 was “never a real level” after they watched a weekly close beneath it.

What I Would Watch Before The Next Monday Scan

First, Intel’s relationship with $119 to $127. A push through the recent high on a close, with the oscillator still nearer 60 than 80, would say the second leg has another chapter. A failure that fills the September 21 gap would say the chapter already ended and the 50-day near $101 is the next conversation. Second, Valero versus $418. Clearing that high after a volume shakeout is the definition of a trend that digested supply. Losing $365 would say the digestion was incomplete. Third, the ten-year itself. A sustained break above the September 28 close of 5.24 percent keeps pressure on anything that needs a mortgage or a project loan. A retreat back under 5 percent would not automatically repair CRH or First Solar, but it would at least stop adding weight to the same side of the boat.

Fourth, and this one is easier to ignore, breadth inside the 145-name list. If leadership narrows to a handful of chip and energy names while the rest of the screen goes stale, the index can still rise and the average holding can still hurt. Relative strength is a ranking. Rankings get lonely at the top when the middle goes quiet. I would rather see the list rotate than worship two champions and call it a market.

There is also a humility clause. The same process that caught a refiner and a processor early can miss the next leader because the chart was not ready on the day you looked. That is acceptable. Capturing every point was never the assignment. Staying wrong after the level breaks is the expensive version of being thorough.

The Human Part Of A Mechanical List

Screens feel objective because a formula put the name on the page. The write-up is where a person still has to say something that can be falsified. Intel above $40 in January was falsifiable. Valero above the rising 50-day was falsifiable. CRH below $105 was falsifiable, and it was falsified. First Solar “not today” was falsifiable too, in the other direction. If the stock had reclaimed the 50-day and held, the caution would have been early, and a later entry would have been the honest update. It did not. The caution aged well even though a summer spike tried to embarrass it.

That is as close as this business gets to a fair exam. You publish the level before the move finishes. You do not edit the level after the fact and claim you always meant the other number. When the move is large enough that the old level is silly, you publish a new one and you say why. Traders get the gap and the shakeout low. Investors get the moving average that used to be resistance and is now the floor. Everyone gets the same instruction about closes. Intraday heroics are how accounts get chopped up by their own reflexes.

If you want a single takeaway from three quarters of this tape, take this. Strength clustered in a processor rebuild and in independent refiners, far from a pure story about software agents, and weakness clustered in businesses that needed cheaper money to keep the multiple alive. The list noticed before the essays did. Your job in the fourth quarter is not to invent a new theme. It is to keep the winners on a leash that still jerks, and to leave the broken rate stories alone until the chart, not the hope, invites you back.

A Last Pass Over The Numbers That Actually Matter

Intel near $119, 50-day near $101, 200-day near $81, trader line $110, investor line $100, report date October 22. Valero near $406, 50-day near $355, 200-day near $260, trader line $365, investor line $350, prior high $418. CRH near $83.60, off the list since March 5, old lines $120 and $105 already gone, guidance still $5.60 to $6.05. First Solar about 45 percent under a June high of $320.95, backlog smaller than the January postcard, entry never granted. Ten-year at 5.24 percent on September 28. Mortgage average at 7.03 percent on September 24. Existing-home sales at 3.98 million in August. Brent’s path from roughly $118 to $70 to $109 is the refiner’s weather report.

Memorize the lines, not the adjectives. Adjectives age badly. A close does not. If the fourth quarter wants to make or break the year, it will do it at those prices, on those dates, with the same unsentimental habit that put 145 names on a list and then removed the ones that stopped leading. I would rather be early on a reset stop than late on a story I already graded.

None of this is a personal recommendation, and none of it knows your taxes, your time horizon, or how you sleep when a refiner drops $50 in a week. It is a record of what price already did, where the original ideas were wrong, and where the surviving trends would have to fail before the grade changes. The clock is running. The levels are written. The rest is just the close.