Research cutoff: Friday, August 21, 2026 at roughly 12:30 BST / 07:30 ET.
U.S. price cutoff: Thursday, August 20 regular-session close.
Horizon: roughly the next 1–15 U.S. trading sessions.
Premarket moves do not count as confirmation in this screen.
Today's answer: UPS has the right shape, but there is still no clean buy
Today produced an interesting trap in the numbers.
Fortinet and F5 suddenly look much better relative to Palo Alto Networks and CrowdStrike over one, three and five sessions. If I only looked at the spreadsheet, both could appear to be entering the catch-up phase.
They are not.
FTNT fell again Thursday to $150.79. FFIV fell for a sixth consecutive session to $377.79. Their apparent short-window relative strength mostly comes from cybersecurity leaders falling even harder — and CrowdStrike had a stock-specific leadership headline in the mix. That is not the kind of confirmation I want to buy.
UPS is more interesting. It is now beating FedEx over the short windows while remaining materially behind over 10 and 20 sessions. That is the transition shape this screen is built to detect. But UPS itself is barely moving, and the company still has legitimate operating reasons for a discount: domestic volume pressure, costly restructuring and investor skepticism about how quickly the post-Amazon network reset turns into durable margin improvement.
So the screen is narrow again:
- UPS: research watch; right relative shape, insufficient absolute confirmation.
- FTNT: rejected under the post-invalidation quarantine.
- FFIV: rejected; six straight down sessions and still below the old invalidation.
- XOM/CVX: no meaningful divergence despite the oil shock.
- LHX: still excluded because the fresh CEO/governance event contaminates the peer comparison.
Bottom line: there is no clean U.S. catch-up buy today.
Candidate summary
| Name | Genuine relationship | What the screen sees | Status | What I need next |
|---|---|---|---|---|
| UPS | Direct parcel/logistics peer FedEx (FDX) | Short windows have turned positive while 10/20-session lag remains large | Best research watch | Absolute price reclaim above roughly $106–$108, then a second session holding while still beating FDX |
| Fortinet (FTNT) | Direct cyber peers PANW + CRWD | Short-window relative numbers are positive, but FTNT itself keeps falling and already broke its old invalidation | Rejected / quarantined | First repair $154–$155, then two-session hold; only later does the old mid-$160s trigger matter |
| F5 (FFIV) | App/API security vs. PANW + CRWD | Large 10/20-session gap, but six straight losing sessions and broken absolute structure | Rejected | Rebuild $400+, hold for two sessions, and restore neutral/positive 5-session relative performance |
| Exxon (XOM) | Chevron / integrated-oil cycle | Oil is elevated, but XOM and CVX have moved almost together | No divergence | Nothing to do until the relationship actually stretches |
These are observation conditions, not opening-bell instructions.
Relative performance through Thursday's close
For FTNT and FFIV, the benchmark is the equal-weight average return of Palo Alto Networks and CrowdStrike. For UPS, I use FedEx, the cleanest directly comparable U.S. parcel/logistics peer. For Exxon, I use Chevron.
The figures are candidate return minus benchmark return, in percentage points. Negative means the candidate lagged.
| Candidate vs benchmark | 1 session | 3 sessions | 5 sessions | 10 sessions | 20 sessions |
|---|---|---|---|---|---|
| UPS vs FDX | +0.44 pp | +2.19 pp | +2.67 pp | -2.24 pp | -11.96 pp |
| FTNT vs PANW + CRWD | +2.87 pp | +5.75 pp | +4.81 pp | -0.33 pp | -6.06 pp |
| FFIV vs PANW + CRWD | +2.95 pp | +4.51 pp | +4.48 pp | -3.05 pp | -8.21 pp |
| XOM vs CVX | +0.83 pp | +0.51 pp | +0.44 pp | -1.69 pp | -0.17 pp |
The table is why today's screen needs judgment rather than a simple ranking formula.
UPS has the cleanest shape: positive 1/3/5-session relative performance, but still a meaningful unresolved 10/20-session deficit.
FTNT and FFIV look even stronger on the short windows — but that is misleading. Their own five-session returns are roughly -8.9% and -9.2%, respectively. The peer basket fell even more. A catch-up trade should eventually show the laggard getting stronger, not merely falling more slowly than a collapsing benchmark.
The seven-session walk-forward adds a peer-shock contamination filter
I reconstructed the screen over the prior seven completed U.S. sessions — August 12, 13, 14, 17, 18, 19 and 20 — using only information that would have been known at each cutoff.
The existing rules did something important: they kept us out of the obvious failed turns.
- FTNT's August 13 relative bounce failed immediately on August 14.
- FTNT's short-window relative improvement appeared again on August 18, but the stock never repaired the absolute structure and then closed through the predefined $154–$155 invalidation.
- FFIV broke its roughly $400 invalidation and has now fallen for six straight sessions.
- UPS never completed the old reclaim-and-hold sequence around the upper-$100s.
- LHX was removed once the CEO-conduct investigation introduced a fresh company-specific reason for the divergence.
So I am not loosening the system after a difficult week.
But Thursday exposed another failure mode: benchmark contamination.
CrowdStrike fell roughly 5.6% Thursday, versus about 2.8% for Palo Alto Networks. If one major comparison stock is being hit by its own company-specific headline, a candidate can look like it is “catching up” simply because the benchmark has become artificially weak.
New peer-shock contamination rule
From today:
If a major comparison peer has a fresh company-specific shock, I do not count relative outperformance against that peer as catch-up confirmation. I either remove the contaminated peer for that session or require confirmation from an uncontaminated sector ETF / median peer basket. And if the candidate itself is still falling, relative outperformance alone is insufficient.
I reapplied that filter to the seven-session window. It does not rescue FTNT or FFIV. Their recent relative improvement remains a false positive because their own absolute structures are still deteriorating.
This is an important distinction. Relative-value screens are supposed to remove broad market noise. They should not accidentally turn a peer's bad news into somebody else's buy signal.
1. UPS: finally the right relative shape, but not the right absolute move
Why the relationship is genuine
UPS and FedEx are exposed to the same parcel and logistics cycle: U.S. business shipping, e-commerce, international air freight, fuel, labor, network utilization and global trade.
This is one of the cleanest relationships in the market for a catch-up screen.
The fundamental story at UPS also improved in Q2. The company reported $22.8 billion of consolidated revenue, adjusted EPS of $1.76, and raised its full-year 2026 outlook to roughly $91.2 billion of revenue, $8.65 billion of adjusted operating profit and $7.22 of adjusted EPS.
Management also said the planned Amazon volume glide-down and related network reconfiguration had been completed.
Those are legitimate reasons to investigate why UPS is still so far behind FedEx over 20 sessions.
Why I still won't call it a buy
The relative shape is improving, but the absolute price is not.
Using adjusted closes to avoid the dividend-date distortion, UPS is only about 1% higher over 10 sessions and still down roughly 8.7% over 20 sessions. It is outperforming FedEx recently mainly because FDX has softened, not because UPS has started a convincing breakout.
There are also real reasons for the discount. Reuters noted investor skepticism after Q2 even though UPS raised its outlook. The company expects a demanding second-half ramp, while its restructuring program carries significant costs. UPS recorded roughly $891 million of after-tax transformation charges in Q2, largely tied to workforce-reduction initiatives.
That means the lag is not a pure anomaly. Part of it may be the market demanding proof.
| Item | UPS plan |
|---|---|
| Observation area | Roughly $102–$105 on the current adjusted-price structure |
| First useful repair | Reclaim roughly $106–$108 during the regular session |
| Confirmation | Hold the repair into a second session while maintaining positive relative performance versus FDX |
| Invalidation | A fresh close below the recent ~$101–$102 area would make the setup less attractive |
| Likely catalyst | Evidence the post-Amazon network reset is improving domestic margin and volume quality; upward estimate revisions |
| Main risks | Weak U.S. parcel volumes, restructuring costs, international margin pressure, fuel, competition from FedEx/Amazon |
Analyst, news and community read
The professional reaction to Q2 was basically: better numbers, but prove the second half. UPS raised the headline guidance, yet the shares fell after the report because investors questioned the pace of the recovery.
The public community discussion is similarly split. Some UPS employee/investor threads speculate that Amazon volume could return during peak season, while other discussions focus on weak volume, transformation charges and margin pressure. That is useful operational color, not a trade signal.
Verdict: UPS is today's best research candidate because the 1/3/5-session relationship has turned while the 10/20-session deficit remains. But I want UPS itself to start rising before I call that a catch-up trade.
2. Fortinet: the spreadsheet says “turn”; the chart says “still broken”
FTNT is the best example of why relative strength needs an absolute-price gate.
Fortinet closed Thursday at $150.79, down another 1.35%. The company remains fundamentally strong: Q2 revenue rose 26% to $2.05 billion, product revenue rose 52% to $773 million, billings rose 33% to $2.37 billion, and management raised 2026 revenue guidance.
Fortinet has also continued adding AI-security capabilities, including the recently announced Virtue AI acquisition.
So the company is not being rejected because the business suddenly collapsed.
It is being rejected because the trade setup failed.
The old screen explicitly said a sustained close below roughly $154–$155 invalidated the setup. Wednesday broke it. Thursday pushed further below it. I am not going to lower that line because FTNT now “outperforms” peers that are falling faster.
| Item | FTNT reset plan |
|---|---|
| Current status | Post-invalidation quarantine |
| First repair | Reclaim roughly $154–$155 |
| Better repair | Reclaim roughly $158–$160 |
| Old catch-up trigger | Mid-$160s only becomes relevant after the lower structure is rebuilt |
| Required confirmation | Two regular sessions holding the repaired level with uncontaminated peer/ETF relative strength |
| Main risks | Valuation compression, preference for other cyber platforms, broad growth-stock pressure, further support failure |
Analyst views remain constructive on the operating story — recent target increases followed the strong quarter — but the market is currently arguing about valuation and timing rather than the quality of Q2.
Verdict: no entry. The stock has to rebuild support first.
3. F5: good quarter, six losing sessions
FFIV is even cleaner as a lesson in not buying fundamentals before price confirms them.
F5's latest quarter was genuinely strong: $865 million of revenue, up 11%, product revenue up 19%, and management raised fiscal-2026 revenue-growth guidance to roughly 9%–10% from 7%–8%.
Analyst expectations are not universally bearish either. S&P Global's recent consensus snapshot showed a Hold rating with an average target around $436 and a wide $300–$515 range — exactly the kind of disagreement that makes relative mispricing possible.
But Thursday was FFIV's sixth consecutive losing session, taking the stock to $377.79. It is now well below the roughly $400 invalidation used in the earlier screen.
The peer-relative numbers have improved only because the benchmark has been crushed too.
| Item | FFIV reset plan |
|---|---|
| Current status | Rejected / quarantined |
| First rebuild | Reclaim $400 |
| Confirmation | Hold above $400 for two regular sessions and restore at least neutral five-session relative performance |
| Old trigger | The old low-$420s catch-up zone matters only after that repair |
| Main risks | Continued security-sector de-rating, product-cycle volatility, enterprise spending slowdown, further absolute downside |
Verdict: the fundamental mismatch is still interesting research. The trade is not.
4. Energy: the common driver is real, but the gap isn't
Oil remains elevated amid the U.S.-Iran standoff, making integrated oil an obvious place to search for a laggard.
But Exxon and Chevron are still moving too closely together. XOM's relative gaps versus CVX are less than two percentage points across every useful window, and the 20-session difference is almost zero.
That is exactly what a shared macro driver should look like when there is no relative-value anomaly.
No divergence, no trade.
Why L3Harris is still excluded
LHX fell again Thursday, but I am not putting it back into the catch-up table.
The CEO transition followed a board investigation into conduct inconsistent with company policy. The company said the issue was unrelated to financial reporting, operations or customer relationships and reaffirmed its 2026 outlook, but it is still a fresh stock-specific governance event.
That contaminates the peer relationship. Until the market has had time to price that event separately, comparing LHX mechanically with the defense basket risks mistaking a governance discount for temporary sector lag.
Friday's macro backdrop: a bad week for forcing mean reversion
U.S. futures are modestly higher Friday morning, but the broader setup is still difficult.
The S&P 500 and Nasdaq are on course for weekly losses of roughly 2%, and the 30-year Treasury yield hit its highest level since 2007 this week. Long-end yields remain near recent highs as investors wrestle with fiscal concerns, inflation and the Federal Reserve's reaction function.
That matters for this strategy because a sharp macro de-rating can make many fundamentally sound stocks look “cheap versus peers” for several days before the relationship stabilizes.
Next week also brings two major regime tests: Nvidia reports on August 26, and the Fed's Jackson Hole symposium runs August 27–29. Those events can move growth, AI, rates and market breadth enough to change several relative relationships at once.
So Friday premarket green does not override Thursday's closing evidence.
Bottom line
Today's best signal is a distinction:
relative improvement is not always catch-up. Sometimes it is just the benchmark breaking faster.
- UPS has the only genuinely attractive transition shape today: positive short-window relative performance with a large unresolved 20-session gap versus FedEx. But UPS itself has not broken out, and its discount has real operating explanations. It stays a watch.
- FTNT looks relatively stronger because cyber peers sold off harder, but it is below its predefined invalidation and remains quarantined.
- FFIV has strong operating fundamentals but six consecutive down sessions. It stays out until $400 is rebuilt and held.
- XOM/CVX show no useful divergence.
- LHX remains excluded because company-specific governance news overwhelms a clean defense-peer thesis.
There is no clean U.S. catch-up buy today.
The new rule is worth keeping: if a peer-specific shock makes the benchmark unusually weak, remove that peer or require an uncontaminated benchmark before declaring victory. A catch-up trade should come from the candidate getting stronger — not from somebody else having a worse day.
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
Friday market / macro
- Reuters — U.S. stock futures rise after Thursday's losses, Aug. 21, 2026: https://www.reuters.com/business/us-stock-futures-rise-after-sharp-losses-prior-session-2026-08-21/
- Reuters — Morning Bid: Big, bad bond market, Aug. 21, 2026: https://www.reuters.com/commentary/reuters-open-interest/global-markets-view-usa-2026-08-21/
- Reuters — Nvidia earnings and Jackson Hole to test the rally: https://www.reuters.com/business/wall-st-week-ahead-nvidia-earnings-jackson-hole-test-pillars-stock-rally-2026-08-21/
Price history / relative-performance calculations
- Fortinet: https://stockanalysis.com/stocks/ftnt/history/
- Palo Alto Networks: https://stockanalysis.com/stocks/panw/history/
- CrowdStrike: https://stockanalysis.com/stocks/crwd/history/
- F5: https://stockanalysis.com/stocks/ffiv/history/
- UPS: https://stockanalysis.com/stocks/ups/history/
- FedEx: https://stockanalysis.com/stocks/fdx/history/
- Exxon Mobil: https://stockanalysis.com/stocks/xom/history/
- Chevron: https://stockanalysis.com/stocks/cvx/history/
UPS
- UPS Investor Relations — Q2 2026 results: https://investors.ups.com/news-events/press-releases/detail/2164/ups-releases-2q-2026-earnings
- Reuters — UPS raises forecasts after Amazon delivery reductions: https://www.reuters.com/business/ups-raises-full-year-revenue-forecast-2026-07-28/
- Reddit / UPSers — recent Amazon-volume discussion: https://www.reddit.com/r/UPSers/comments/1vi6lqt/amazon_coming/
- Reddit / UPSers — post-Q2 operating concerns discussion: https://www.reddit.com/r/UPSers/comments/1v928ml/why_the_stock_price_fell/
Fortinet / cybersecurity
- Fortinet Investor Relations — Q2 2026 results: https://investor.fortinet.com/press-releases
- Investor's Business Daily — Fortinet's AI acquisition strategy / Virtue AI: https://www.investors.com/research/ibd-stock-of-the-day/fortinet-stock-artificial-intelligence-acquisition/
- MarketWatch — FTNT Aug. 20 session: https://www.marketwatch.com/data-news/fortinet-inc-stock-underperforms-thursday-when-compared-to-competitors-3a502e5d-94dace113e13
- Reddit / ValueInvesting — Fortinet vs. CrowdStrike valuation discussion: https://www.reddit.com/r/ValueInvesting/comments/1upsubn/fortinet_vs_crowdstrike_cybersecurity_valuations/
F5
- F5 — Q3 FY2026 results and raised outlook: https://www.f5.com/company/news/press-releases/earnings-q3-fy26
- MarketWatch — FFIV Aug. 20 session / sixth consecutive decline: https://www.marketwatch.com/data-news/f5-inc-stock-underperforms-thursday-when-compared-to-competitors-d3251f08-3f08deb1c0c9
- StockAnalysis / S&P Global — F5 analyst consensus: https://stockanalysis.com/stocks/ffiv/forecast/
Community links are anecdotal and included only for sentiment / operating color, not as validation of a trade.