Research cutoff: Saturday, August 29, 2026.
U.S. price cutoff: Friday, August 28 regular-session close.
Horizon: roughly the next 1–15 U.S. trading sessions.
Weekend rule: U.S. cash markets are closed today, so Friday's close is the only price confirmation that counts.
Today's answer: one live watch, no clean buy
Friday did not give me a new catch-up trade. It gave me something more useful: a cleaner separation between an early laggard, a stock that has already caught up, and a stock whose giant gap is simply bad price action.
UPS is the only live catch-up watch I would carry into Monday. It gave back a little relative performance Friday, but it is still ahead of FedEx over the 3-, 5- and 10-session windows while remaining about 6.6 percentage points behind over 20 sessions. That is the short-term-turn / longer-term-gap structure this screen is built to find.
The catch is the same as yesterday: UPS itself still has not broken out. Friday's $105.33 close remains below the roughly $106.20 first repair level and well below the cleaner $108+ confirmation area.
Broadcom is no longer an early catch-up candidate. AVGO slightly fell Friday, but it strongly outperformed SOXX as the semiconductor basket sold off. It now beats SOXX over the 1-, 3-, 5- and 10-session windows. The only meaningful negative relative window left is the old 20-session gap. Under the freshness rule, that means the catch-up has already progressed too far to call this an early mean-reversion setup.
RIOT is the opposite problem. Its gap versus spot-Bitcoin exposure is enormous, but the stock collapsed Friday while Bitcoin held up much better. A bigger negative gap does not make the trade better when the candidate itself is still falling.
So the weekend conclusion is simple:
There is no clean U.S. catch-up buy today. UPS is the only live watch.
Candidate summary
| Name | Genuine relationship | What the screen sees | Status | What I need next |
|---|---|---|---|---|
| UPS | Direct parcel/logistics peer FedEx (FDX) | Positive 3/5/10-session relative turn while the 20-session deficit remains meaningful | Only live watch | Regular-session close above roughly $106.20, preferably $108+, then hold while the short-window relative turn survives |
| Broadcom (AVGO) | AI infrastructure / semiconductors vs SOXX, with NVDA as demand leader | Relative repair has already spread across 1/3/5/10 sessions; only the old 20-day gap remains | Graduated / too late to call early | Do not chase the old gap; reassess after Sept. 2 earnings |
| RIOT | Bitcoin economics vs IBIT, with an increasing AI/data-center component | Huge 10/20-session deficit, but absolute price is still breaking down | Rejected | First stabilize; then rebuild above roughly $21, with $22.20 a stronger repair area |
| Cisco (CSCO) | AI/data-center networking vs Arista (ANET) | Long-window lag exists, but Cisco's post-earnings margin compression is a real company-specific explanation | Rejected | Needs evidence margins/estimates are stabilizing before a relative-value thesis is clean |
I am not padding this to three "buys." One live watch is better than three forced ideas.
Relative performance through Friday's close
The figures below are candidate price return minus benchmark price return, in percentage points. Negative means the candidate lagged its benchmark over that window.
| Candidate vs benchmark | 1 session | 3 sessions | 5 sessions | 10 sessions | 20 sessions |
|---|---|---|---|---|---|
| UPS vs FDX | -0.17 pp | +1.03 pp | +1.47 pp | +1.92 pp | -6.57 pp |
| AVGO vs SOXX | +2.46 pp | +4.44 pp | +2.29 pp | +1.44 pp | -6.00 pp |
| RIOT vs IBIT | -5.98 pp | -9.01 pp | -4.74 pp | -23.37 pp | -29.03 pp |
| CSCO vs ANET | +0.86 pp | -3.39 pp | -4.57 pp | +0.16 pp | -13.56 pp |
These four rows tell four very different stories.
UPS is the useful one: the short windows have turned, but a meaningful older deficit remains.
AVGO has already moved through the transition. If I bought it now because the 20-day number is still negative, I would be using an old gap to justify chasing a move that has already happened.
RIOT has the largest gap, but the worst confirmation. That is exactly the sort of setup a naive mean-reversion screen gets wrong.
Cisco's gap is statistically interesting, but the reason for it is not mysterious. Its latest earnings produced a genuine margin debate that Arista does not share.
Seven-session walk-forward: the rules are getting stricter, not looser
I reconstructed the screen across the prior seven completed U.S. sessions — August 20, 21, 24, 25, 26, 27 and 28 — using only information available at each cutoff.
I am keeping the raw hit rate private. The useful part is what the failures and near-misses say about the process.
A few patterns keep repeating:
- RIOT has produced sharp one-day relative bounces that failed almost immediately. The second-session confirmation rule prevented those from becoming completed signals.
- Cadence became a legitimate short-window turn, then caught up quickly enough that the remaining old gap became stale. The freshness rule kept the screen from turning a successful early idea into a late momentum chase.
- UPS has gradually improved against FedEx, but the absolute breakout still has not arrived. The price-reclaim rule has kept us from buying the relationship just because the table looks attractive.
- Broadcom passed the semiconductor catalyst-day test on Thursday and still beat SOXX on Friday, but that success creates a new problem: it now outperforms the basket across 1/3/5/10 sessions. This is increasingly a completed catch-up, not a fresh one.
New refinement: candidate-event proximity freeze
The last seven sessions also expose one hole in the event rule.
We already freeze a new signal when a direct peer is about to report, because a peer's earnings can mechanically create or erase the relative gap.
From today, I am extending that logic to the candidate's own major binary event:
If the candidate's own earnings or similarly binary company event is within the next three regular sessions, a new catch-up signal cannot be completed solely on pre-event relative strength. Wait for the event and require post-event price/relative confirmation.
This matters immediately for Broadcom, which reports fiscal Q3 on September 2.
The reason is practical. A catch-up thesis says the market has temporarily underpriced a company relative to a real economic peer group. But an earnings report can change the company's own revenue, margin, guidance and valuation inputs overnight. Once that event is two sessions away, the upcoming company-specific information dominates the old peer gap.
This is a material change to the method, not a cosmetic wording update. It should reduce situations where the screen buys an apparently healthy reversion right before the company itself resets the relationship.
1. UPS: this is still the only live transition
Why the UPS / FedEx relationship is real
UPS and FedEx compete directly in parcel delivery, international logistics, business shipping and time-sensitive transportation. They share exposure to fuel, labor, trade volumes, economic activity, e-commerce and network utilization.
That makes relative divergence meaningful when company-specific explanations are not overwhelming the pair.
The business is improving, but it is not problem-free
UPS reported $22.8 billion of Q2 revenue and adjusted EPS of $1.76, then raised its full-year outlook to roughly $91.2 billion of revenue and $7.22 of adjusted EPS.
The company has also announced more than $2 billion of ongoing investment across International, Healthcare and Supply Chain Solutions, part of the push toward higher-quality revenue as it reduces lower-margin Amazon exposure.
There are still real reasons the stock trades at a discount to FedEx. The Amazon volume reset is not painless, transformation expenses remain heavy, and investors are still waiting for stronger proof that the network can turn lower volume into better margins.
Friday added another industry variable: Reuters reported that UPS and FedEx fuel surcharges have climbed above 24%, versus roughly 9% in 2021. Those surcharges can protect margins, but they can also pressure customers and shipping demand. That is a mixed catalyst, not a free profit lever.
What Friday changed
UPS closed Friday at $105.33, down slightly from Thursday's $105.68. FedEx also slipped, but by a little less.
That turned the one-session relative reading slightly negative again, while leaving the 3-, 5- and 10-session readings positive. The 20-session deficit remains about 6.6 percentage points.
That is still a good catch-up shape.
But the absolute chart has still not done the thing that matters most.
| Item | UPS plan |
|---|---|
| Observation area | Roughly $104–$106 |
| First useful trigger | Regular-session close above roughly $106.20 |
| Better confirmation | Push through $108, then hold while 3/5-session relative strength versus FDX stays positive |
| Invalidation | Fresh close below roughly $101–$102 weakens the transition materially |
| Likely catalysts | Better volume quality after the Amazon reset, healthcare/international expansion, network-cost leverage, positive estimate revisions |
| Main risks | Weak parcel demand, restructuring costs, fuel/labor pressure, tariffs/trade, Amazon and FedEx competition |
Analyst and community temperature
The sell side is constructive, but not unanimous. StockAnalysis currently shows a Buy consensus from 29 analysts with an average target around $116. Recent August calls range from Bernstein maintaining Buy with a $131 target to BMO and Wells Fargo staying at Hold.
That split is actually useful for this setup. The market sees upside if the turnaround works, but there is enough skepticism that the stock has not already priced in a clean recovery.
Community discussion is more cautious. Recent UPS employee/investor threads focus on lower domestic volume, the Amazon reset, transformation charges and uncertainty about how quickly operations normalize. I treat that as anecdotal operating color, not as evidence by itself.
Verdict: UPS is the only live catch-up watch. I am not buying the table; I am waiting for the stock. A $106.20 close would be useful, $108+ would be much cleaner.
2. Broadcom: a good catch-up can become a bad late entry
Broadcom is the most important methodological example today.
Thursday was exactly what I wanted: AVGO rose sharply after Nvidia's results, reclaimed $370, and beat the diversified semiconductor basket even though Nvidia itself was the sector leader.
Friday, semiconductor stocks weakened. AVGO closed around $368.79, but SOXX fell much harder. That left AVGO ahead of the basket over the 1-, 3-, 5- and 10-session windows.
Only the 20-session relative gap remains materially negative.
That means the early catch-up argument has mostly worked.
Why I am not chasing it
The temptation is obvious: AVGO still sits roughly six percentage points behind SOXX over 20 sessions, analyst sentiment is very bullish, and AI demand remains strong.
But the screen is specifically designed to catch early relative repair. Once the stock is already beating its basket across every shorter window, the stale-gap rule says the old 20-day deficit is not enough.
There is also now a dominant company event. Broadcom reports fiscal Q3 after the close on September 2. Benchmark reiterated a Buy rating Friday with a $545 target, and the broader Wall Street consensus remains strongly bullish. That is encouraging, but it does not make buying two sessions before earnings a clean peer-reversion trade.
| Item | AVGO read |
|---|---|
| Current status | Catch-up has substantially progressed; no longer an early laggard |
| Price reference | Roughly $365–$370 remains the recent repair area |
| Next decision point | Sept. 2 earnings, then reassess relative performance against SOXX |
| What would invalidate the repair | Sharp post-event rejection back below the recent repair zone combined with renewed SOXX underperformance |
| Main risks | Earnings/guidance reset, customer concentration, hyperscaler custom-chip competition, valuation, rates/tariffs |
Retail sentiment around Broadcom remains highly earnings-focused and polarized: plenty of AI enthusiasm, but also recurring concern about customer concentration, valuation and how much good news is already embedded. That is another reason not to turn a completed catch-up into a pre-earnings chase.
Verdict: graduated from the early-laggard list. Good setup, increasingly late entry.
3. RIOT: the largest gap is the easiest one to get wrong
RIOT is what happens when a relative screen forgets to look at the candidate itself.
Friday, RIOT closed around $18.99, down roughly 9%, while spot-Bitcoin exposure through IBIT fell much less. RIOT now trails IBIT by roughly 23 percentage points over 10 sessions and 29 points over 20 sessions.
Those are enormous numbers.
They are not a buy signal.
Why the relationship is getting more complicated
RIOT still has meaningful Bitcoin sensitivity through mining economics, treasury value and crypto risk appetite. But the company is also trying to become a much larger AI/data-center infrastructure story.
Its Q2 update included $174.2 million of revenue, up 14% year over year, and the company signed a 191 MW, 20-year data-center lease expected to generate roughly $9.1 billion of initial-term contract revenue. B. Riley kept a Buy rating after that deal, while Needham also raised its target.
That diversification is potentially valuable. It also means IBIT is becoming a less complete benchmark for RIOT than it was when Riot was a simpler Bitcoin-miner proxy.
Why Friday is still a rejection
Even allowing for the changing business mix, Friday's tape was poor. Coverage of the miner group noted that RIOT and MARA sold off sharply while Bitcoin itself held much firmer.
The old rule still applies: relative strength has to be accompanied by absolute stabilization. RIOT currently has neither.
| Item | RIOT plan |
|---|---|
| Status | Rejected, not a live catch-up candidate |
| First repair | Stabilize and rebuild above roughly $21 |
| Stronger repair | Reclaim roughly $22.20 and hold while relative performance versus IBIT improves |
| Invalidation | New lows / continued absolute breakdown make the gap irrelevant |
| Potential catalysts | Data-center lease execution, AI infrastructure milestones, Bitcoin recovery |
| Main risks | Bitcoin weakness, mining economics, execution/capex, financing, AI pivot timing |
Community discussion remains very excited about the AI/data-center lease, while analyst targets are substantially above the current price. That optimism is exactly why price confirmation matters: the market is currently refusing to reward the narrative.
Verdict: rejected. Gap size is not signal quality.
4. Cisco: this lag has an explanation
Cisco versus Arista is a real economic relationship. Both are exposed to enterprise/cloud networking, AI data-center connectivity and large infrastructure budgets.
The relative table also looks tempting: Cisco still trails Arista materially over the 20-session window.
But this is precisely where the company-specific filter has to do its job.
Cisco's latest quarter was strong on revenue and AI orders, yet the stock sold off because gross margin compressed to 66.3% and management's near-term gross-margin outlook disappointed investors. The market is not randomly forgetting Cisco while rewarding Arista; it is pricing a specific margin concern.
Several analysts still raised targets after the quarter, and Cisco's AI infrastructure orders remain a real bull point. But the margin debate is legitimate enough that the peer gap cannot be treated as unexplained mispricing.
Verdict: rejected from the catch-up list until margin/estimate pressure stabilizes. This may eventually become a good stock idea; it is not a clean relative-laggard idea today.
Other Friday laggards I am explicitly not promoting
Friday had plenty of dramatic declines. Most fail this strategy because the weakness has a clear reason.
- Marvell (MRVL): fell more than 10% as investors questioned the timing of revenue from its Google AI-chip agreement. That is a company-specific revenue-timing reset, not clean unexplained peer lag.
- PayPal (PYPL): dropped roughly 13% after reports that a potential acquisition consortium abandoned its pursuit. Obvious event-driven repricing.
- IREN: fell after earnings despite fast AI-cloud growth because adjusted EBITDA disappointed. Again, company-specific earnings information dominates the relative setup.
The screen should reject those names even when a peer or sector ETF is doing better. A laggard is not automatically mispriced.
Market backdrop: Friday made the hurdle higher
Wall Street ended lower Friday after Fed Chair Kevin Warsh reinforced the Fed's commitment to the 2% inflation target. Reuters reported the S&P 500 fell 0.25%, the Nasdaq 0.52%, and traders increased bets on a September rate hike.
That matters for this screen because higher-rate expectations can extend valuation pressure on growth stocks and also change the economic outlook for transports.
The next week has two important checkpoints:
- September 2: Broadcom earnings, plus other major corporate results.
- September 4: the August U.S. jobs report, a key input into the Fed's September decision.
Fund flows also turned more cautious: U.S. equity funds saw about $22.3 billion of net outflows in the week through August 26, the biggest weekly outflow since March.
None of that tells me to short the market. It tells me not to weaken the confirmation rules just because a laggard looks statistically cheap relative to a peer.
Bottom line
The interesting thing Friday was not a new laggard. It was the opposite: Broadcom stopped being one.
AVGO's relative repair is now broad enough that I would rather call the catch-up substantially underway than keep pointing at the old 20-session gap. With its own earnings on September 2, it also becomes the first clear example for the new candidate-event proximity freeze.
UPS is the only live watch. The relationship against FedEx still has the right shape: positive short and medium windows, a remaining 20-session deficit, intact enough fundamentals, and a plausible catalyst path. But the stock itself still has not closed above the repair zone.
RIOT is rejected despite the giant Bitcoin gap. The absolute chart is too weak, and the company's AI pivot is making IBIT a less complete benchmark anyway.
Cisco is rejected because the lag has a real company-specific explanation: margins.
So the weekend call remains:
No clean U.S. catch-up buy yet.
The one level I care about most on Monday is UPS above roughly $106.20, with $108+ the stronger proof. If that happens while the 3- and 5-session relative turn versus FedEx stays positive, the screen will finally have a cleaner actionable transition.
Educational disclaimer: This report is for informational and educational purposes only. It is not personalized financial advice, investment research, or a recommendation to buy or sell any security. Relative relationships can break permanently rather than mean-revert.
Sources
Market and next-week catalysts
- Reuters, Aug. 28 — Wall Street ends lower after Warsh reaffirms inflation fight: https://www.reuters.com/business/sp-500-nasdaq-futures-slip-after-tech-rally-warshs-speech-awaited-2026-08-28/
- Reuters, Aug. 28 — Jobs report and Broadcom results are next hurdles: https://www.reuters.com/business/wall-st-week-ahead-jobs-report-broadcom-results-pose-next-hurdles-stock-market-2026-08-28/
- Reuters, Aug. 28 — U.S. equity funds post biggest weekly outflow since March: https://www.reuters.com/business/us-equity-funds-post-biggest-weekly-outflow-since-march-ahead-nvidia-earnings-2026-08-28/
UPS / FedEx
- Reuters — UPS raises full-year revenue forecast: https://www.reuters.com/business/ups-raises-full-year-revenue-forecast-2026-07-28/
- Reuters — transport fuel surcharges rise amid Iran conflict: https://www.reuters.com/business/energy/iran-war-drives-us-transport-fuel-surcharges-also-industry-profits-2026-08-28/
- StockAnalysis — UPS analyst forecasts: https://stockanalysis.com/stocks/ups/forecast/
- Yahoo Finance — UPS vs FedEx operating comparison: https://finance.yahoo.com/markets/stocks/articles/fedex-corporation-fdx-growing-faster-205403027.html
- UPS historical prices: https://shareprices.com/us/ups/history/
- FedEx historical prices: https://shareprices.com/us/fdx/history/
- Reddit / UPSers — post-Q2 operating concerns: https://www.reddit.com/r/UPSers/comments/1v928ml/why_the_stock_price_fell/
Broadcom / semiconductors
- TipRanks — Benchmark bullish ahead of Sept. 2 earnings: https://www.tipranks.com/news/broadcom-stock-avgo-why-benchmark-is-bullish-ahead-of-q3-earnings
- StockAnalysis — Broadcom analyst forecast: https://stockanalysis.com/stocks/avgo/forecast/
- StockAnalysis — SOXX overview and holdings: https://stockanalysis.com/etf/soxx/
- Investing.com — Broadcom historical data: https://www.investing.com/equities/avago-technologies-historical-data
- Reddit / BroadcomStock — investor AI-revenue discussion: https://www.reddit.com/r/BroadcomStock/
RIOT / Bitcoin
- Yahoo Finance — RIOT and MARA sell off while Bitcoin holds firmer: https://finance.yahoo.com/markets/crypto/articles/mara-riot-sink-6-while-152002594.html
- Yahoo Finance — B. Riley keeps RIOT at Buy after $9.1B AI lease: https://finance.yahoo.com/technology/ai/articles/b-riley-keeps-riot-buy-133607142.html
- RIOT historical prices: https://stockanalysis.com/stocks/riot/history/
- Reddit / RiotBlockchain — community discussion: https://www.reddit.com/r/RiotBlockchain/new/
Cisco / Arista
- Yahoo Finance — Cisco drops on gross-margin fears despite strong quarter: https://finance.yahoo.com/markets/stocks/articles/cisco-drops-7-gross-margin-143127630.html
- StockAnalysis — Cisco overview/history: https://stockanalysis.com/stocks/csco/
- StockAnalysis — Arista overview: https://stockanalysis.com/stocks/anet/
Community sources are anecdotal and included only for sentiment and operating color, not as trade validation.